Prosecution Insights
Last updated: November 18, 2026
Application No. 18/430,107

COMPUTING SYSTEM AND METHOD FOR BUILDING AND EXECUTING AN ENSEMBLE MODEL FOR FORECASTING TIME-SERIES DATA

Non-Final OA §103
Filed
Feb 01, 2024
Priority
Dec 29, 2023 — continuation of PCTCN2023143405
Examiner
GOLAN, MATTHEW BRYCE
Art Unit
2123
Tech Center
2100 — Computer Architecture & Software
Assignee
Capital One Financial Corporation
OA Round
1 (Non-Final)
0%
Grant Probability
At Risk
1-2
OA Rounds
0%
With Interview

Examiner Intelligence

Grants only 0% of cases
0%
Career Allowance Rate
0 granted / 9 resolved
-55.0% vs TC avg
Minimal +0% lift
Without
With
+0.0%
Interview Lift
resolved cases with interview
Typical timeline
Avg Prosecution
27 currently pending
Career history
54
Total Applications
across all art units

Office Action

§103
DETAILED ACTION This communication is in response to Application No. 18/430,107 filed on February 1st, 2024 in which claims 1-20 are presented for examination. Notice of Pre-AIA or AIA Status The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . Information Disclosure Statement The information disclosure statement submitted on 04/29/2024 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement was considered by the examiner. Specification The contents of the specification are sufficient for examination purposes. Claim Objections Claims 4-6 are objected to because of the following informalities: “relative to other relative to other” (Claim 4, Ln. 32) should be “relative to other” (objection applies equally to dependent claims 5-6). Appropriate correction is required. Claim Rejections - 35 USC § 103 In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status. The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action: A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made. Claims 1-2, 8-14, 16-18, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Morgan et al. (hereinafter Morgan) (Pat. No. US 11,170,391 B2) in view of Ouyang et al. (hereinafter Ouyang) (“Multi-step Time Series Forecasting with an Ensemble of Varied Length Mixture Models”). Regarding Claim 1, Morgan teaches a computing platform comprising: at least one network interface; at least one processor; at least one non-transitory computer-readable medium; and program instructions stored on the at least one non-transitory computer-readable medium that, when executed by the at least one processor, cause the computing platform to (Fig. 2 and Col. 3-4, Ln. 46-48, “Referring now to FIG. 2 . . . the computing system 106 includes at least one central processing unit (“CPU”) 202 . . . The computing system 106 further includes a mass storage device 214. The mass storage device 214 is able to store software instructions and data . . . The mass storage device 214 and its associated computer-readable storage media provide non-volatile, non-transitory data storage for the computing system 106 . . . The computing system 106 may connect to the network 222 through a network interface unit 404 . . . The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”, where the computing platform, “the computing system 106”, comprises at least one network interface, “a network interface unit 404”; at least one processor, “at least one central processing unit (“CPU”) 202”; at least one non-transitory computer-readable medium, “The mass storage device 214 and its associated computer-readable storage media provide non-volatile, non-transitory data storage”; and program instructions stored on the at least one non-transitory computer-readable medium, “The mass storage device 214 is able to store software instructions”, that, when executed by the at least one processor, cause the computing platform to, “The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”): generate a . . . set of time-series models that are configured to predict forecast values of a target time-series variable for a first target timeframe, wherein the . . . set of time-series models comprises time-series models of at least two different model types (Fig. 4 and Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”, where a set of time-series models, “a demand forecasting model is built . . . from combining two or more component models”, comprises time-series models of at least two different model types, “selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”; see also Abstract, “Forecast models are built and used to score incoming sales data to predict future demand for items” and Col. 3, Ln. 23-28, “The systems and methods are scalable and customizable to different applications of the forecast data. For example, a forecast may be generated on a weekly basis at the chain level for a group of retail stores. However, the forecast can be broken down to an individual store or an individual date”, where the set of time-series models, “Forecast models are built”, are configured to predict forecast values of a target time-series variable, “used to score incoming sales data to predict future demand for items”, for a first target timeframe, “a forecast may be generated on a weekly basis at the chain level”); generate a . . . set of time-series models that are configured to predict forecast values of a target time-series variable for a second target timeframe that differs from the first target timeframe, wherein the . . . set of time-series models comprises time-series models of at least two different model types (Fig. 4 and Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”, where a set of time-series models, “a demand forecasting model is built . . . from combining two or more component models”, comprises time-series models of at least two different model types, “selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”; see also Abstract, “Forecast models are built and used to score incoming sales data to predict future demand for items” and Col. 3, Ln. 23-28, “The systems and methods are scalable and customizable to different applications of the forecast data. For example, a forecast may be generated on a weekly basis at the chain level for a group of retail stores. However, the forecast can be broken down to an individual store or an individual date”, where the set of time-series models, “Forecast models are built”, are configured to predict forecast values of a target time-series variable, “used to score incoming sales data to predict future demand for items”, for a second target timeframe, “individual store or an individual date”, that differs from the first target timeframe, “a forecast may be generated on a weekly basis at the chain level”; see also Col. 14, Ln. 34-38, “In instances when a client requests a demand forecast for a particular day or a particular store, the disaggregation service 336 computes the demand forecast based on the aggregate demand forecasts stored in the forecasts data store 312”); cause a client device associated with a user to present a graphical user interface (GUI) that enables a user to configure an ensemble model for forecasting values of the target time-series variable (Abstract, “Forecast models are built and used to score incoming sales data to predict future demand for items. Forecast models are validated by evaluating actual demand against predicted demand and using that information to inform how future ensemble forecast will be generated”; Fig. 1; and Col. 3, Ln. 29-45, “FIG. 1 illustrates a schematic diagram of an example computing network 100 . . . The administrator computing device 106 communicates with the demand forecasting system 108 through the network 110 to request demand forecasts, evaluate forecasts generated by the demand forecasting system 108 and visualize those results. The requesting, evaluating, and visualizing may occur with the assistance of a user interface 112 displayed on the administrator computing device 106”, where client device, “The administrator computing device 106”, is associated with a user, “administrator”, to present a graphical user interface (GUI), “user interface 112”, to perform a request, “requesting”, for model forecasting values of the target time-series variable, “Forecast models are built and used . . . to predict future demand for items”; see also Col. 19, Ln. 8-16, “receive a client request, the client request including parameters for a demand forecast . . . and based on the parameters, build an ensemble model, the ensemble model comprising a weighted combination of two or more forecasting models from the plurality of forecasting models” and Col. 5, Ln. 14-15, “The component models are generally selected from time series forecasting models”, where the GUI enables a user, the user associated with the “client request”, to configure an ensemble model, “build an ensemble model”, for forecasting values of the target time-series variable, “the ensemble model comprising a weighted combination of two or more forecasting models”); receive, from the client device over a network-based communication path, configuration data for a given ensemble model that identifies a user-selected group of time-series models to be included in the given ensemble model, wherein the user-selected group of time-series models includes at least one time-series model from the . . . set of time-series models and at least one time-series model from the . . . set of time-series models (Fig. 1; Col. 3, Ln. 38-41, “The administrator computing device 106 communicates with the demand forecasting system 108 through the network 110 to request demand forecasts”; Fig. 3; and Col. 6, Ln. 27-36, “A query is received from a client at the API . . . Optionally, the API can also receive a selection of a particular model or model collection with which to forecast demand”, where configuration data, “a selection of a . . . model collection”, is “receive[d]” from the client device, “The administrator computing device 106 communicates”, over a network-based communication path, “through the network 110”; see also Col. 19, Ln. 8-16, “receive a client request, the client request including parameters for a demand forecast . . . and based on the parameters, build an ensemble model, the ensemble model comprising a weighted combination of two or more forecasting models from the plurality of forecasting models” and Col. 5, Ln. 14-15, “The component models are generally selected from time series forecasting models”, where the configuration information, “the client request including parameters”, is for a given ensemble model, “build an ensemble model”, that identifies a user-selected group of time-series models to be included in the given ensemble model , “the ensemble model comprising a weighted combination of two or more forecasting models”, from the set of time-series models, “from the plurality of forecasting models”); based on the received configuration data, construct the given ensemble model from the user-selected group of time-series models (Fig. 3-4; Col. 6, Ln. 27-36, “A query is received from a client at the API . . . Optionally, the API can also receive a selection of a particular model or model collection with which to forecast demand”; and Col. 5, Ln. 5-9, “At operation 404, a demand forecasting model is built . . . an enterprise “consensus” model is built from combining two or more component models”, where based on the received configuration data, “a selection of a particular model or model collection”, the given ensemble model is constructed, “a demand forecasting model is built”, from the user-selected group of time-series models, “the API can also receive a selection of a particular . . . model collection with which to forecast demand” and “an enterprise “consensus” model is built from combining two or more component models”; see also Col. 19, Ln. 8-16, “receive a client request, the client request including parameters for a demand forecast . . . and based on the parameters, build an ensemble model, the ensemble model comprising a weighted combination of two or more forecasting models from the plurality of forecasting models”), wherein the given ensemble model is configured to blend forecast values of the target time-series variable that are predicted by the user-selected group of time-series models (Fig. 6; Col. 3, Ln. 11-18, “An ensemble forecast is compiled using component models that are combined and weighted to produce a consensus forecast for predicting item demand . . . Models having superior performance for predicting item demand are weighted more heavily in the overall consensus forecast”; Col. 8, Ln. 54-58, “The ensemble model utilizes a weighted combination of two or more component models producing a more accurate representation of demand than any individual component model. The weighted combinations of models are adjusted over time, increasing accuracy as more data is analyzed”; and Col. 6, Ln. 27-36, “A query is received from a client at the API . . . Optionally, the API can also receive a selection of a particular model or model collection with which to forecast demand”, wherein the given ensemble model, “ensemble model”, is configured to blend forecast values, “compiled using component models that are combined and weighted to produce a consensus forecast”, of the target time-series variable, “for predicting item demand”, that are predicted by the user-selected group of time-series models, “A query is received from a client . . . [with a] selection of a particular model or model collection”; see also Col. 10, Ln. 2-9, “The ensemble forecast, ŷth, is an affine function of the model forecasts: [EQUATION 1]”); and after constructing the given ensemble model, utilize the given ensemble model to predict a given sequence of forecast values for the target time-series variable (Fig. 4 and Col. 5, Ln. 39-43, “At operation 408, an aggregate demand forecast is generated. The enterprise forecast engine 304 generates demand forecasts in batches by default. In one embodiment, forecasts are generated for each item, across all stores in a retail chain, every week”, where after construction, the model is utilized to predict a given sequence of forecast values for the target time-series variable, “an aggregate demand forecast is generated . . . for each item, across all stores in a retail chain, every week”; see also Col. 19, Ln. 8-16, “receive a client request, the client request including parameters for a demand forecast . . . and based on the parameters, build an ensemble model, the ensemble model comprising a weighted combination of two or more forecasting models from the plurality of forecasting models”, where the model is the “ensemble model”). Morgan does not explicitly disclose . . . first . . . second . . . first . . . second . . . (where only a single set of time-series models are specifically described in regard to predicting the forecast values of the target time-series variable for first and second target timeframes and for inclusion within the ensemble model; redundant recitations of first and second omitted). However, Ouyang teaches . . . [a] first [time-series model set that is configured to predict forecast values for a first target timeframe and] . . . [a] second [time-series model set that is configured to predict forecast values for a second target timeframe that differs from the first target timeframe] . . . (Pg. 1, Abstract, “A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance” and Pg. 2, Col. 2, Para. 1, “an adaptive combination of VLM models of various orders, short-term (e.g. daily) and long-term (e.g. weekly), together with an ensemble, is employed for multi-step forecasting of time series, resulting the ensemble VLM (eVLM) method”, where a first time-series model set that is configured to predict forecast values for a first target timeframe and, “VLM models of various orders, short-term (e.g. daily)” and a second time-series model set that is configured to predict forecast values for a second target timeframe that differs from the first target timeframe, “VLM models of various order . . . long-term (e.g. weekly)”, are disclosed; see also Pg. 8, Col. 1, Para. 2, “In FX forecasting, when daily data is used, short-time usually means a day, or a couple of days or few days ahead; while long terms could be weekly or fortnight, or even months ahead”) [wherein at least one model from each of the] first . . . [and] second . . .[sets of time-series models are included in an ensemble model] (Pg. 1, Abstract, “A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance” and Pg. 2, Col. 2, Para. 1, “an adaptive combination of VLM models of various orders, short-term (e.g. daily) and long-term (e.g. weekly), together with an ensemble, is employed for multi-step forecasting of time series, resulting the ensemble VLM (eVLM) method”). Before the effective filing date of the invention, it would have been obvious to one of ordinary skill in the art to combine the generating of a set of time-series models that are configured to predict forecast values of a target time-series variable for a first target timeframe and a second target timeframe that differs from the first target timeframe, wherein the set of time-series models comprises time-series models of at least two different model types and wherein the time-series models are selected for inclusion in an ensemble model of Morgan with the first time-series model set that is configured to predict forecast values for a first target timeframe and the second time-series model set that is configured to predict forecast values for a second target timeframe that differs from the first target timeframe, wherein at least one model from each of the first and second sets of time-series models are included in an ensemble model of Ouyang in order to improve performance of the ensemble forecasting model through increased accuracy from use of varied length models and reduced prediction errors from ensemble forecasting (Ouyang, Pg. 12, Col. 1-12, Para. 3-1, “varied length models retain as much dependency information as possible among the points within the prediction horizon. These dependencies between consecutive future points serve the key role in improving the prediction accuracy of these varied length models can further reduce the prediction error and enhance the performance”; see also Ouyang, Pg. 1, Abstract, “A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance” and Ouyang, Pg. 7, Col. 2, Para. 1, “As combining forecasts or ensemble predicting can lead to performance improvement over individual models by reducing prediction variance or errors, an adaptive combination of VLM models of various orders, short-term (e.g. daily) and long-term (e.g. weekly), together with an ensemble, is employed for multi-step forecasting of time series, resulting the ensemble VLM (eVLM) method”), while retaining prediction flexibility and accuracy optimization benefits for each target timeframe through use of multiple models for both the first and second sets of time-series models (Morgan, Col. 16, Ln. 49-57, “The present systems and methods provide a novel approach to solving the problem of accurately predicting demand for particular items within a retail context. The use of weighted component models to generate an ensemble demand forecasting model is novel and advantageous over prior art methods because it allows for flexibility throughout a long time period such as a year, to accommodate changes in demand that occur due to seasonality, holidays, and promotions”; see also Morgan, Col. 8, Ln. 54-57, “The ensemble model utilizes a weighted combination of two or more component models producing a more accurate representation of demand than any individual component model” and Morgan, Col. 8, Ln. 26-32, “In one embodiment, the enterprise forecast engine 304 utilizes one main model based on a recurring neural-network (RNN). In some cases, supplemental models such as an ARIMA or LOESS model can be used in conjunction with the RNN model to accommodate for changes in demand caused by seasonality, holidays, or other variations in demand”). Regarding Claim 2, Morgan in view of Ouyang teach the computing platform of claim 1, further comprising program instructions stored on the at least one non-transitory computer-readable medium that, when executed by the at least one processor, cause the computing platform to (Morgan, Fig. 2 and Morgan, Pg. 19, Col. 3-4, Ln. 46-48, “Referring now to FIG. 2 . . . The computing system 106 further includes a mass storage device 214. The mass storage device 214 is able to store software instructions and data . . . The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”): before generating the first and second sets of time-series models, obtain model setup parameters and source data for use in generating the first and second sets of time-series models (Morgan, Fig. 4; Morgan, Col. 4, Ln. 62-65, “At operation 402, the common data preparation engine 302 receives and prepares past sales data and past demand forecasts. The common data preparation engine 302 receives and prepares both past data and incoming current data”; Morgan, Fig. 6; and Morgan, Col. 8, Ln. 61-67, “At operation 602, parameters for a demand forecast are received at a forecast generator. Parameters can include one or more of a group of items, a time period, a location, and other attributes”, wherein, before generating the set of time-series models, where “Receive and prepare data 402” is before “Build forecasting model 404” and “Receive parameters for demand forecast 602” is the first step in “404”, model setup parameters, “parameters for a demand forecast”, and source data, “past sales data and past demand forecasts”, for use in the generating the models are obtained, see Morgan, Col. 7, Ln. 65-66, “Building data includes past data that is used for building demand forecasting models” and Morgan, Col. 9, Ln. 1-4, “At operation 604, component models are selected based on past performance. Add operation 606, the component models are weighted based on the past performance of the component models for the selected parameters”, which, in view of Ouyang are first and second sets of time-series models, “short and long VLM models” see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 above, and remain applicable here. Regarding Claim 8, Morgan in view of Ouyang teach the computing platform of claim 1, wherein the given ensemble model is configured to blend the forecast values of the target time-series variable that are output by the at least one time-series model from the first set of time-series models for the first target timeframe across time with the forecast values of the target time-series variable that are output by at least one time-series model from the second set of time-series models for the second target timeframe (Morgan, Fig. 6; Morgan, Col. 3, Ln. 11-18, “An ensemble forecast is compiled using component models that are combined and weighted to produce a consensus forecast for predicting item demand . . . Models having superior performance for predicting item demand are weighted more heavily in the overall consensus forecast”; Morgan, Col. 8, Ln. 54-58, “The ensemble model utilizes a weighted combination of two or more component models producing a more accurate representation of demand than any individual component model. The weighted combinations of models are adjusted over time, increasing accuracy as more data is analyzed”; and Morgan, Col. 6, Ln. 27-36, “A query is received from a client at the API . . . Optionally, the API can also receive a selection of a particular model or model collection with which to forecast demand”, wherein the given ensemble model, “ensemble model”, is configured to blend forecast values, “compiled using component models that are combined and weighted to produce a consensus forecast”, of the target time-series variable, “for predicting item demand”, that are output by the user-selected group of time-series models, “A query is received from a client . . . [with a] selection of a particular model or model collection”; see also Morgan, Col. 10, Ln. 2-9, “The ensemble forecast, ŷth, is an affine function of the model forecasts: [EQUATION 1]”; Morgan, Col. 19, Ln. 8-16, “receive a client request, the client request including parameters for a demand forecast . . . and based on the parameters, build an ensemble model, the ensemble model comprising a weighted combination of two or more forecasting models from the plurality of forecasting models”, which, in view of Ouyang, are by the at least one time-series model from the first set of time-series models for the first target timeframe across time, “combination of VLM models of various orders, short-term (e.g. daily)”, with the forecast values of the target time-series variable that are output by at least one time-series model from the second set of time-series models for the second target timeframe, “combination of VLM models of various orders . . . long-term (e.g. weekly)”, Ouyang, Pg. 2, Col. 2, Para. 1, “an adaptive combination of VLM models of various orders, short-term (e.g. daily) and long-term (e.g. weekly), together with an ensemble, is employed for multi-step forecasting of time series, resulting the ensemble VLM (eVLM) method” and Ouyang, Pg. 8, Col. 1, Para. 2, “In FX forecasting, when daily data is used, short-time usually means a day, or a couple of days or few days ahead; while long terms could be weekly or fortnight, or even months ahead”; see also Morgan, Fig. 15; Morgan, Col. 14, Ln. 26-31, “FIG. 15 provides a flow chart of a method 416 of disaggregating an aggregate demand forecast to fulfill a client request. The disaggregation service 336 operates to break down aggregate demand forecasts into demand forecasts for shorter time periods or a smaller number of locations”; and Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 above, and remain applicable here. Regarding Claim 9, Morgan in view of Ouyang teach the computing platform of claim 1, wherein the user-selected group of time-series models includes two or more time-series models from the first set of time-series models (Morgan, Col. 19, Ln. 8-16, “receive a client request, the client request including parameters for a demand forecast . . . and based on the parameters, build an ensemble model, the ensemble model comprising a weighted combination of two or more forecasting models from the plurality of forecasting models” and Morgan, Col. 5, Ln. 14-15, “The component models are generally selected from time series forecasting models”, where the configuration information, “the client request including parameters”, is for a given ensemble model, “build an ensemble model”, that identifies a user-selected group of time-series models to be included in the given ensemble model, “the ensemble model comprising a weighted combination of two or more forecasting models”, from the set of time-series models, “from the plurality of forecasting models”, which can include more than two models, “two or more forecasting models”, such that, in view of Ouyang, two must be from either the “short” or “long” models, either of which can be considered the first set of time-series models, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”), and wherein the configuration data comprises respective weight value for the two or more time-series models from the first set of time-series models (Morgan, Fig. 1; Morgan, Col. 3, Ln. 38-41, “The administrator computing device 106 communicates with the demand forecasting system 108 through the network 110 to request demand forecasts”; Morgan, Fig. 3; and Morgan, Col. 6, Ln. 27-36, “A query is received from a client at the API . . . Optionally, the API can also receive a selection of a particular model or model collection with which to forecast demand”, where configuration data, “a selection of a . . . model collection”, is “receive[d]” from the client device, “The administrator computing device 106 communicates”, and can include “weighted” values for the “combination” of models, see Morgan, Col. 8, Ln. 54-58, “The ensemble model utilizes a weighted combination of two or more component models producing a more accurate representation of demand than any individual component model. The weighted combinations of models are adjusted over time, increasing accuracy as more data is analyzed”, which, in view of Ouyang are for each of the first and second sets of time-series models, “short and long VLM models”, including the above discussed two-or more models from the first set of time-series models, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 above, and remain applicable here. Regarding Claim 10, Morgan in view of Ouyang teach the computing platform of claim 9, wherein the given ensemble model is configured to blend the forecast values of the target time-series variable that are output by the two or more time-series models from the first set of time-series models (Morgan, Fig. 6; Morgan, Col. 3, Ln. 11-18, “An ensemble forecast is compiled using component models that are combined and weighted to produce a consensus forecast for predicting item demand . . . Models having superior performance for predicting item demand are weighted more heavily in the overall consensus forecast”; Morgan, Col. 8, Ln. 54-58, “The ensemble model utilizes a weighted combination of two or more component models producing a more accurate representation of demand than any individual component model. The weighted combinations of models are adjusted over time, increasing accuracy as more data is analyzed”; and Morgan, Col. 6, Ln. 27-36, “A query is received from a client at the API . . . Optionally, the API can also receive a selection of a particular model or model collection with which to forecast demand”, wherein the given ensemble model, “ensemble model”, is configured to blend forecast values, “compiled using component models that are combined and weighted to produce a consensus forecast”, of the target time-series variable, “for predicting item demand”, that are output by the user-selected group of time-series models, “A query is received from a client . . . [with a] selection of a particular model or model collection”; see also Morgan, Col. 10, Ln. 2-9, “The ensemble forecast, ŷth, is an affine function of the model forecasts: [EQUATION 1]”; Morgan, Col. 19, Ln. 8-16, “receive a client request, the client request including parameters for a demand forecast . . . and based on the parameters, build an ensemble model, the ensemble model comprising a weighted combination of two or more forecasting models from the plurality of forecasting models” and Morgan, Col. 5, Ln. 14-15, “The component models are generally selected from time series forecasting models”, where the configuration information, “the client request including parameters”, is for a given ensemble model, “build an ensemble model”, that identifies a user-selected group of time-series models to be included in the given ensemble model, “the ensemble model comprising a weighted combination of two or more forecasting models”, from the set of time-series models, “from the plurality of forecasting models”, which can include more than two models, “two or more forecasting models”, such that, in view of Ouyang, two must be from either the “short” or “long” models, either of which can be considered the first set of time-series models, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”) in accordance with the respective weight values for the two or more time-series models from the first set of time-series models (Morgan, Col. 8, Ln. 54-58, “The ensemble model utilizes a weighted combination of two or more component models producing a more accurate representation of demand than any individual component model. The weighted combinations of models are adjusted over time, increasing accuracy as more data is analyzed”, where the blending is in accordance with the respective “weighted” values for the “combination” of models, which, in view of Ouyang are for each of the first and second sets of time-series models, “short and long VLM models”, including the above discussed two-or more models from the first set of time-series models, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 above, and remain applicable here. Regarding Claim 11, Morgan in view of Ouyang teach the computing platform of claim 1, wherein the time-series models in the first and second sets each have: a first input feature corresponding to the target time-series variable; and one or more additional input features corresponding to one or more influencing variables (Morgan, Abstract, “the demand forecasting system and methods described herein are useful for predicting demand of products in a retail context. Forecast models are built and used to score incoming sales data to predict future demand for items” and Morgan, Col. 16, Ln. 51-57, “The use of weighted component models to generate an ensemble demand forecasting model is novel and advantageous over prior art methods because it allows for flexibility throughout a long time period such as a year, to accommodate changes in demand that occur due to seasonality, holidays, and promotions”, where the models, as evidenced by the “ensemble forecast”, each have a first input feature corresponding to the target time-series variable, “predicting demand of products in a retail context” such as the input for “The value of feature j in week t”, and one or more additional input features corresponding to one or more influencing variables, “changes in demand that occur due to seasonality, holidays, and promotions” such as the input for “a promotion in period t”, see Morgan, Col. 10, Ln. 2-12, “The ensemble forecast, ŷth, is an affine function of the model forecasts: [EQUATION 1] . . . The value of feature j in week t, xjt, might represent a promotion in period t, for example”, which, in view of Ouyang are each of the first and second sets of time-series models, “short and long VLM models” see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 above, and remain applicable here. Regarding Claim 12, Morgan in view of Ouyang teach the computing platform of claim 1, further comprising program instructions stored on the at least one non-transitory computer-readable medium that, when executed by the at least one processor, cause the computing platform to (Morgan, Fig. 2 and Morgan, Pg. 19, Col. 3-4, Ln. 46-48, “Referring now to FIG. 2 . . . The computing system 106 further includes a mass storage device 214. The mass storage device 214 is able to store software instructions and data . . . The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”): cause the client device associated with the user to present the given sequence of forecast values for the target time-series variable (Morgan, Fig. 4 and Morgan, Col. 7, Ln. 15-19, “At operation 420, the demand forecast is output to the client in response to the client's request. In some embodiments, the client is an administrator user interface and the demand forecast is visualized for viewing, for example, by an administrator user”, where the client device, “an administrator user interface”, associated with the user, “by an administrator user”, presents the given sequence of forecast values for the target time-series variable, “the demand forecast is visualized”, see also Morgan, Abstract, “Forecast models are built and used to score incoming sales data to predict future demand for items. Forecast models are validated by evaluating actual demand against predicted demand and using that information to inform how future ensemble forecast will be generated. Forecasts may be broken down into smaller components to satisfy a variety of requests for data from client applications”; see generally Morgan, Fig. 7-15). Regarding Claim 13, Morgan teaches a non-transitory computer-readable medium, wherein the non-transitory computer-readable medium is provisioned with program instructions that, when executed by at least one processor, cause a computing platform to: . . . (Fig. 2; Pg. 19, Col. 3-4, Ln. 63-7, “The mass storage device 214 and its associated computer-readable storage media provide non-volatile, non-transitory data storage for the computing system 106 . . . [the] computer-readable data storage media can include any available tangible, physical device or article of manufacture from which the CPU 402 can read data and/or instructions. In certain embodiments, the computer-readable storage media comprises entirely non-transitory media”; and Pg. 19, Col. 4, Ln. 41-45, “The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”). The remaining limitations are substantially the same as limitations of Claim 1, therefore it is rejected under the same rationale. Regarding Claim 14, the additional elements of the dependent claim are substantially the same as limitations of Claim 2, therefore it is rejected under the same rationale. Regarding Claim 16, the additional elements of the dependent claim are substantially the same as limitations of Claim 8, therefore it is rejected under the same rationale. Regarding Claim 17, Morgan teaches a method implemented by a computing platform, the method comprising: . . . (Pg. 1, Abstract, “Methods and systems for forecasting demand for a plurality of items are provided”; see also Fig. 4; Fig. 3; and Pg. 19, Col. 4, Ln. 59-61, “FIG. 4 illustrates an example flow chart of a method 400 of forecasting demand that may be implemented by the demand forecasting system 108 of FIG. 3”, where the “method . . . may be implemented by the . . . system 108 of FIG. 3”, which is a computing platform, see Pg. 18, Col. 2, Ln. 36-37, “FIG. 3 illustrates a more detailed schematic diagram of the demand forecasting system of FIG. 1”; Fig. 1; and Pg. 19, Col. 3, Ln. 29-45, “FIG. 1 illustrates a schematic diagram of an example computing network 100 . . . The administrator computing device 106 communicates with the demand forecasting system 108 through the network 110 to request demand forecasts, evaluate forecasts generated by the demand forecasting system 108 and visualize those results. The requesting, evaluating, and visualizing may occur with the assistance of a user interface 112 displayed on the administrator computing device 106”). The remaining limitations are substantially the same as limitations of Claim 1, therefore it is rejected under the same rationale. Regarding Claim 18, the additional elements of the dependent claim are substantially the same as limitations of Claim 2, therefore it is rejected under the same rationale. Regarding Claim 20, the additional elements of the dependent claim are substantially the same as limitations of Claim 8, therefore it is rejected under the same rationale. Claims 3, 15, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Morgan in view of Ouyang and Passalis et al. (hereinafter Passalis) (“Deep Adaptive Input Normalization for Time Series Forecasting”). Regarding Claim 3, Morgan in view of Ouyang teaches the computing platform of claim 1, further comprising program instructions stored on the at least one non-transitory computer-readable medium that, when executed by the at least one processor, cause the computing platform to (Morgan, Fig. 2 and Morgan, Pg. 19, Col. 3-4, Ln. 46-48, “Referring now to FIG. 2 . . . The computing system 106 further includes a mass storage device 214. The mass storage device 214 is able to store software instructions and data . . . The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”), before generating the first and second sets of time-series models: obtain historical data for the target time-series variable; obtain historical data for one or more . . . [other] variables (Morgan, Fig. 4 and Morgan, Col. 4, Ln. 62-65, “At operation 402, the common data preparation engine 302 receives and prepares past sales data and past demand forecasts. The common data preparation engine 302 receives and prepares both past data and incoming current data”, wherein, before generating the set of time-series models, “Receive and prepare data 402” is before “Build forecasting model 404”, historical data for the target time-series variable and historical data for one or more other variables is obtained, “receives and prepares past sales data and past demand forecasts”, which, in view of Ouyang, are first and second sets of time-series models, “short and long VLM models” see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”); and . . . [process] the historical data for the target time-series variable based on the historical data for one or more . . . [other] variables (Morgan, Col. 17, Ln. 23-25, “The present disclosure describes computational methods for analyzing both past sales data and currently occurring sales data”, where the historical data for the target time-series variable, “past sales data”, is processed, “analyz[ed]”, based on the historical data for one or more other variables, “Past forecasting performance”, to determine “which models are best used for particular sets of items”, see Morgan, Col. 3, Ln. 14-15, “Past forecasting performance is evaluated to determine which models are best used for particular sets of items”), wherein the first and second sets of time-series models are thereafter generated based on the . . . historical data for the target time-series variable (Morgan, Col. 7, Ln. 65-66, “Building data includes past data that is used for building demand forecasting models” and Morgan, Col. 9, Ln. 1-4, “At operation 604, component models are selected based on past performance. Add operation 606, the component models are weighted based on the past performance of the component models for the selected parameters”, where the “component models” are thereafter generated, “building demand forecasting models” and “selected”, based on the historic data, “Building data includes past data that is used”, which, in view of Ouyang are first and second sets of time-series models, “short and long VLM models” see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 above, and remain applicable here. Morgan in view of Ouyang do not explicitly disclose . . . offset . . . normalize . . . (where normalization of the data using an offset is not specifically described; redundant recitations of offset and normalize omitted). However, Passalis teaches . . . [use of an] offset [variable to] . . . normalize . . . [time-series data for model forecasting] (Pg. 1, Col. 1, Abstract, “Deep Learning (DL) models can be used to tackle time series analysis tasks with great success. However, the performance of DL models can degenerate rapidly if the data are not appropriately normalized. This issue is even more apparent when DL is used for financial time series forecasting tasks” and Pg. 2, Col. 2, Para. 2, “The goal of the proposed method is to learn how the measurements x(i) j should be normalized by appropriately shifting and scaling them: PNG media_image1.png 50 236 media_image1.png Greyscale ”, where “α” is an offset variable, see Pg. 3, Col. 1, Para. 2, “the shifting operator α(i) is defined using a linear transformation of the extracted summary representation as: α(i)”). Before the effective filing date of the invention, it would have been obvious to one of ordinary skill in the art to combine the obtaining of historical data for the target time-series variable historical data and for one or more other variables before generating the first and second sets of time-series models and processing the historical data for the target time-series variable based on the historical data for one or more other variables, wherein the first and second sets of time-series models are thereafter generated based on the historical data for the target time-series variable of Morgan in view of Ouyang with the use of an offset variable to normalize time-series data for model forecasting of Passalis in order to prevent performance degeneration of the time series forecasting models (Passalis, Pg. 1, Col. 1, Abstract, “Deep Learning (DL) models can be used to tackle time series analysis tasks with great success. However, the performance of DL models can degenerate rapidly if the data are not appropriately normalized. This issue is even more apparent when DL is used for financial time series forecasting tasks”), using an easy to implement approach with superior performance to other normalization techniques (Passalis, Pg. 6, Col. 1, Para. 2, “The proposed method is easy to implement, while at the same allows for directly using the raw time series data. The ability of the proposed method to improve the forecasting performance was evaluated using three different deep learning models and two time series forecasting datasets. The proposed method consistently outperformed all the other evaluated normalization approaches”). Regarding Claim 15, the additional elements of the dependent claim are substantially the same as limitations of Claim 3, therefore it is rejected under the same rationale. Regarding Claim 19, the additional elements of the dependent claim are substantially the same as limitations of Claim 3, therefore it is rejected under the same rationale. Claims 4-6 are rejected under 35 U.S.C. 103 as being unpatentable over Morgan in view of Ouyang and Tran (“Grid Search of Convolutional Neural Network model in the case of load forecasting”). Regarding Claim 4, Morgan in view of Ouyang teach the computing platform of claim 1, wherein: the program instructions that, when executed by the at least one processor, cause the computing platform to (Morgan, Fig. 2 and Morgan, Pg. 19, Col. 3-4, Ln. 46-48, “Referring now to FIG. 2 . . . The computing system 106 further includes a mass storage device 214. The mass storage device 214 is able to store software instructions and data . . . The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”) generate the first set of time-series models (Morgan, Abstract, “Forecast models are built and used to score incoming sales data to predict future demand for items”, where the “models” are generated, “built”, which, in view of Ouyang are the first and set of time-series models, “short . . . VLM models”, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”) comprise program instructions stored on the at least one non-transitory computer-readable medium that, when executed by the at least one processor, cause the computing platform to (Morgan, Fig. 2 and Morgan, Pg. 19, Col. 3-4, Ln. 46-48, “Referring now to FIG. 2 . . . The computing system 106 further includes a mass storage device 214. The mass storage device 214 is able to store software instructions and data . . . The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”), for each given model type of the at least two different model types: train . . . the given model type that are configured to predict forecast values of the target time-series variable for the first target timeframe [and] . . . select . . . the given model type to include in the first set of time-series models (Morgan, Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)” and Morgan, Abstract, “Forecast models are built and used to score incoming sales data to predict future demand for items”, where, for each given model type of the at least two different model types, “an enterprise “consensus” model is built from combining two or more component models”, the “models” are selected, “built”, “selected from time series forecasting models”, and “used”, which, in view of Ouyang are the first set of time-series models, “short . . . VLM models”, that are trained, “the VLM models are trained”, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance” and Ouyang, Pg. 1, Abstract, “Training data are segmented to various lengths corresponding to various forecasting horizons, and the VLM models are trained in a self-organizing fashion on these segments to capture these dependencies in its component autoregressive models of various predicting horizons”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”); and the program instructions that, when executed by the at least one processor, cause the computing platform to (Morgan, Fig. 2 and Morgan, Pg. 19, Col. 3-4, Ln. 46-48, “Referring now to FIG. 2 . . . The computing system 106 further includes a mass storage device 214. The mass storage device 214 is able to store software instructions and data . . . The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”) generate the second set of time-series models (Morgan, Abstract, “Forecast models are built and used to score incoming sales data to predict future demand for items”, where the “models” are generated, “built”, which, in view of Ouyang are the second set of time-series models, “long VLM models”, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”) comprise program instructions stored on the at least one non-transitory computer-readable medium that, when executed by the at least one processor, cause the computing platform to (Morgan, Fig. 2 and Morgan, Pg. 19, Col. 3-4, Ln. 46-48, “Referring now to FIG. 2 . . . The computing system 106 further includes a mass storage device 214. The mass storage device 214 is able to store software instructions and data . . . The mass storage device 414 and/or the RAM 410 also store software instructions, that when executed by the CPU 402, cause the computing system 106 to provide the functionality discussed in this document”), for each given model type of the at least two different model types: train . . . the given model type that are configured to predict forecast values of the target time-series variable for the second target timeframe . . . [and] select . . . the given model type to include in the second set of time-series models (Morgan, Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)” and Morgan, Abstract, “Forecast models are built and used to score incoming sales data to predict future demand for items”, where, for each given model type of the at least two different model types, “an enterprise “consensus” model is built from combining two or more component models”, the “models” are selected, “built”, “selected from time series forecasting models”, and “used”, which, in view of Ouyang are the second set of time-series models, “long VLM models”, that are trained, “the VLM models are trained”, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance” and Ouyang, Pg. 1, Abstract, “Training data are segmented to various lengths corresponding to various forecasting horizons, and the VLM models are trained in a self-organizing fashion on these segments to capture these dependencies in its component autoregressive models of various predicting horizons”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 above, and remain applicable here. Morgan in view of Ouyang do not explicitly disclose . . . a given batch of candidate time-series models of . . . wherein the candidate time-series models in the given batch have different hyperparameter combinations . . . based on an evaluation of the given batch of candidate time-series models of the given model type, determine a respective measure of performance for each of the different hyperparameter combinations; identify a hyperparameter combination that has a best measure of performance relative to other hyperparameter combinations; and . . . a candidate time-series model having the identified hyperparameter combination as a time-series model of . . . (where training, determining, identifying, and selecting candidate time-series models with different hyperparameter combinations are not specifically discussed in regard to each model type of each target timeframe; redundant recitations omitted). However, Tran teaches . . . [for a model type, train] a given batch of candidate time-series models of [of the model type] . . . wherein the candidate time-series models in the given batch have different hyperparameter combinations . . . (Pg. 25, Abstract, “This paper proposes a framework for Grid Search hyperparameters of the CNN model. In a training process, the optimal models will specify conditions that satisfy requirement for minimum of accuracy”; Pg. 25, Para. 2, “Depending on the period of prediction time, load forecast problems can be divided into 4 groups: very short-term, short-term, medium-term and long-term load forecasting”; Pg. 30-31, Para. 4-1, “Combining all tuning hyperparameters gives 384 cases corresponding to 384 possible models of 1D CNN methodology”; and Pg. 29, Fig. 4, “The 1D CNN pseudocode: training stage (a)”, where for each model type, “the CNN model” of the “load forecast problems . . . divided into 4 groups”, a given batch of candidate time-series models of the model type, “384 possible models of 1D CNN”, see Pg. 27, Para. 2, “a 1D CNN for time series prediction”, is trained, “a training process” as depicted in Fig. 4, “Input . . . combination of tuning hyperparameters . . . Training model”, and wherein the candidate time-series models in the given batch have different hyperparameter combinations, “Combining all tuning hyperparameters gives 384 cases”) based on an evaluation of the given batch of candidate time-series models of the given model type, determine a respective measure of performance for each of the different hyperparameter combinations (Pg. 25, Abstract, “This paper proposes a framework for Grid Search hyperparameters of the CNN model. In a training process, the optimal models will specify conditions that satisfy requirement for minimum of accuracy scores”; Pg. 26, Para. 1, “In the present paper, theCNNGrid Search methodology is proposed with the consideration of not only hyperparameters that determine the network structure and training, but also the pre-processing input data as the number of input and differencing data . . . The number of accuracy scores . . .were also used. In the training process, the optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”; Pg. 30-31, Para. 4-1, “Combining all tuning hyperparameters gives 384 cases corresponding to 384 possible models of 1D CNN methodology”; and Pg. 29, Fig. 4, “prediction and evaluation stage (b)”, where based on an evaluation of the given batch of candidate time-series models of the given model type, “384 possible models of 1D CNN”, see Pg. 27, Para. 2, “a 1D CNN for time series prediction”, determine a respective measure of performance, “the optimal models will specify conditions that satisfy requirement for minimum of accuracy scores”, for each of the different hyperparameter combinations, “theCNNGrid Search methodology is proposed with the consideration of . . . hyperparameters that determine the network structure and training”, as shown in Fig. 4, “Input . . . model from training stage . . . Calculate accuracy scores”; see also Pg. 30, Para. 2, “In the testing process, these optimal models will be compared to all other models according to their accuracy scores in order to evaluate reliability of the Grid Search model of 1D CNN methodology); identify a hyperparameter combination that has a best measure of performance relative to other hyperparameter combinations; and . . . [select] a candidate time-series model having the identified hyperparameter combination as a time-series model of . . . [the model type] (Pg. 26, Para. 1, “In the present paper, theCNNGrid Search methodology is proposed with the consideration of not only hyperparameters that determine the network structure and training, but also the pre-processing input data as the number of input and differencing data . . . The number of accuracy scores . . .were also used. In the training process, the optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores” and Pg. 30, Para. 2, “In the testing process, these optimal models will be compared to all other models according to their accuracy scores in order to evaluate reliability of the Grid Search model of 1D CNN methodology”, where a hyperparameter combination that has a best measure of performance relative to other hyperparameter combinations, see also “these optimal models will be compared to all other models”, is included in the “the optimal models with certain hyperparameters”, which must be identified and selected to be “obtained to meet the minimum requirement of the accuracy scores”; see also Pg. 25, Abstract, “This paper proposes a framework for Grid Search hyperparameters of the CNN model”). Before the effective filing date of the invention, it would have been obvious to one of ordinary skill in the art to combine the, for each given model type of the at least two different model types of the first and second sets of time series models: training of the given model type that are configured to predict forecast values of the target time-series variable for the respective timeframe and the selecting of the given model type to include in the respective set of time-series models of Morgan in view of Ouyang with the, for a model type, training of a given batch of candidate time-series models of the model type, wherein the candidate time-series models in the given batch have different hyperparameter combinations; determining a respective measure of performance for each of the different hyperparameter combinations based on an evaluation of the given batch of candidate time-series models of the given model type; identifying a hyperparameter combination that has a best measure of performance relative to other hyperparameter combinations; and selecting a candidate time-series model having the identified hyperparameter combination as a time-series model of the model type of Tran in order to select the optimal hyperparameter combinations for each model type of each set of time series models (Tran, Pg. 26, Para. 1, “theCNNGrid Search methodology is proposed with the consideration of not only hyperparameters that determine the network structure and training, but also the pre-processing input data as the number of input and differencing data . . . The number of accuracy scores such as Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and Mean Absolute Error (MAE) were also used. In the training process, the optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”; see also Tran, Pg. 31, Para. 2, “the optimal model was selected according the minimum values of accuracy scores RMSE, MAPE and MAE”) using an accurate and reliable framework (Tran, Pg. 34, Para. 2, “These results clearly show that the optimal model received during the training by applying 1D CNN Grid Search model will give good values in the testing process”; Tran, Pg. 26, Para. 1, “Load demand data of Queensland (Australia) and Ho Chi Minh City (Vietnam) were analyzed through training and testing processes to verify the accuracy and reliability of the Grid Search framework”) with broad applicability across multiple model architectures, use cases, and target timeframes (Tran, Pg. 25, Para. 2, “Depending on the period of prediction time, load forecast problems can be divided into 4 groups: very short-term, short-term, medium-term and long-term load forecasting [1–4]. Recently, many techniques and methodologies have been applied to forecast electricity load. These forecasting techniques are mainly classified into two classes: artificial intelligence methods (Support Vector Machine, Artificial Neural Networks, etc.) and statistical methods (Multiple Regression, Exponential Smoothing, ARIMA and Seasonal ARIMA, etc.)”; see also Tran, Pg. 34, Para. 2, “These results clearly show that the optimal model received during the training by applying 1D CNN Grid Search model will give good values in the testing process. In this regard, it is promising for applying the Grid Search model of 1D CNN methodology in any load demand times series”), which will lead to effective time-series forecasting (Tran, Pg. 34, Para. 3, “The results indicated the existence of an optimal model that satisfies the minimum requirement for accuracy scores. During the testing process, accuracy scores of the optimal model gave good values close to the minimum and much lower than the average. The positive results obtained in this study show an effective way to forecast load demand”). Regarding Claim 5, Morgan in view of Ouyang and Tran teach the computing platform of claim 4, wherein: in the given batch of candidate time-series models for each given model type of the at least two different model types of the first set of time-series models, the different hyperparameter combinations are selected using grid search (Tran, Pg. 25, Abstract, “This paper proposes a framework for Grid Search hyperparameters of the CNN model”; Tran, Pg. 26, Para. 1, “In the training process, the optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”; Tran, Pg. 25, Para. 2, “Depending on the period of prediction time, load forecast problems can be divided into 4 groups: very short-term, short-term, medium term and long-term load forecasting”; and Tran, Pg. 30-31, Para. 4-1, “Combining all tuning hyperparameters gives 384 cases corresponding to 384 possible models of 1D CNN methodology”, where in the given batch of candidate time-series models, “384 possible models of 1D CNN methodology”, for a given set of time-series models, “4 groups: very short-term, short-term, medium term and long-term load forecasting”, the different hyperparameter combinations are selected using grid search, “a framework for Grid Search hyperparameters of the CNN model” where “the optimal models with certain hyperparameters were obtained”, and where, as applied to Morgan, the search is conducted for each given model type of the at least two different model types for the set, see Morgan, Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”, which, in view of Ouyang, are of the first set of time-series models, “short . . . VLM models”, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”); and in the given batch of candidate time-series models for each given model type of the at least two different model types of the second set of time-series models, the different hyperparameter combinations are selected using grid search (Tran, Pg. 25, Abstract, “This paper proposes a framework for Grid Search hyperparameters of the CNN model”; Tran, Pg. 26, Para. 1, “In the training process, the optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”; Tran, Pg. 25, Para. 2, “Depending on the period of prediction time, load forecast problems can be divided into 4 groups: very short-term, short-term, medium term and long-term load forecasting”; and Tran, Pg. 30-31, Para. 4-1, “Combining all tuning hyperparameters gives 384 cases corresponding to 384 possible models of 1D CNN methodology”, where in the given batch of candidate time-series models, “384 possible models of 1D CNN methodology”, for a given set of time-series models, “4 groups: very short-term, short-term, medium term and long-term load forecasting”, the different hyperparameter combinations are selected using grid search, “a framework for Grid Search hyperparameters of the CNN model” where “the optimal models with certain hyperparameters were obtained”, and where, as applied to Morgan, the search is conducted for each given model type of the at least two different model types for the set, see Morgan, Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”, which, in view of Ouyang, are of the second set of time-series models, “long VLM models”, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 and the rejection of Claim 4 above, and remain applicable here. Regarding Claim 6, Morgan in view of Ouyang and Tran teach the computing platform of claim 4, wherein: based on the evaluation of the given batch of candidate time-series models for each given model type of the at least two different model types of the first set of time-series models (Tran, Pg. 25, Abstract, “This paper proposes a framework for Grid Search hyperparameters of the CNN model. In a training process, the optimal models will specify conditions that satisfy requirement for minimum of accuracy scores”; Tran, Pg. 26, Para. 1, “In the present paper, theCNNGrid Search methodology is proposed with the consideration of not only hyperparameters that determine the network structure and training, but also the pre-processing input data as the number of input and differencing data . . . The number of accuracy scores . . .were also used. In the training process, the optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”; Tran, Pg. 30-31, Para. 4-1, “Combining all tuning hyperparameters gives 384 cases corresponding to 384 possible models of 1D CNN methodology”; and Tran, Pg. 29, Fig. 4, “prediction and evaluation stage (b)”, where based on an evaluation of the given batch of candidate time-series models of the given model type, “384 possible models of 1D CNN”, see Tran, Pg. 27, Para. 2, “a 1D CNN for time series prediction”, determine a respective measure of performance, “the optimal models will specify conditions that satisfy requirement for minimum of accuracy scores”, for each of the different hyperparameter combinations, “theCNNGrid Search methodology is proposed with the consideration of . . . hyperparameters that determine the network structure and training”, as shown in Fig. 4, “Input . . . model from training stage . . . Calculate accuracy scores”; see also Tran, Pg. 30, Para. 2, “In the testing process, these optimal models will be compared to all other models according to their accuracy scores in order to evaluate reliability of the Grid Search model of 1D CNN methodology”, and where, as applied to Morgan, the search is conducted for each given model type of the at least two different model types for the set, see Morgan, Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”, which, in view of Ouyang, are of the first set of time-series models, “short . . . VLM models”, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”), the respective measure of performance that is determined for each of the different hyperparameter combinations comprises a respective mean absolute percentage error (MAPE) value (Tran, Pg. 30-31, Para. 4-1, “Combining all tuning hyperparameters gives 384 cases corresponding to 384 possible models of 1D CNN methodology”; Tran, Pg. 30, Para. 2, “In the testing process, these optimal models will be compared to all other models according to their accuracy scores in order to evaluate reliability of the Grid Search model of 1D CNN methodology”; and Tran, Pg. 26, Para. 1, “theCNNGrid Search methodology is proposed with the consideration of not only hyperparameters that determine the network structure and training, but also the pre-processing input data as the number of input and differencing data . . . The number of accuracy scores such as Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and Mean Absolute Error (MAE) were also used. In the training process, the optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”, where the respective measure of performance, “optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”, that is determined for each of the different hyperparameter combinations, “Combining all tuning hyperparameters gives 384 cases” “these optimal models will be compared to all other models”, comprises a respective mean absolute percentage error (MAPE) value, “The number of accuracy scores such as . . . Mean Absolute Percentage Error (MAPE)”); and based on the evaluation of the given batch of candidate time-series models for each given model type of the at least two different model types of the second set of time-series models (Tran, Pg. 25, Abstract, “This paper proposes a framework for Grid Search hyperparameters of the CNN model. In a training process, the optimal models will specify conditions that satisfy requirement for minimum of accuracy scores”; Tran, Pg. 26, Para. 1, “In the present paper, theCNNGrid Search methodology is proposed with the consideration of not only hyperparameters that determine the network structure and training, but also the pre-processing input data as the number of input and differencing data . . . The number of accuracy scores . . .were also used. In the training process, the optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”; Tran, Pg. 30-31, Para. 4-1, “Combining all tuning hyperparameters gives 384 cases corresponding to 384 possible models of 1D CNN methodology”; and Tran, Pg. 29, Fig. 4, “prediction and evaluation stage (b)”, where based on an evaluation of the given batch of candidate time-series models of the given model type, “384 possible models of 1D CNN”, see Tran, Pg. 27, Para. 2, “a 1D CNN for time series prediction”, determine a respective measure of performance, “the optimal models will specify conditions that satisfy requirement for minimum of accuracy scores”, for each of the different hyperparameter combinations, “theCNNGrid Search methodology is proposed with the consideration of . . . hyperparameters that determine the network structure and training”, as shown in Fig. 4, “Input . . . model from training stage . . . Calculate accuracy scores”; see also Tran, Pg. 30, Para. 2, “In the testing process, these optimal models will be compared to all other models according to their accuracy scores in order to evaluate reliability of the Grid Search model of 1D CNN methodology”, and where, as applied to Morgan, the search is conducted for each given model type of the at least two different model types for the set, see Morgan, Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”, which, in view of Ouyang, are of the second set of time-series models, “long VLM models”, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”; see also Morgan, Col. 5, Ln. 12-18, “the forecasting models are typically built using python or R software programming languages. The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”), the respective measure of performance that is determined for each of the different hyperparameter combinations comprises a respective MAPE value (Tran, Pg. 30-31, Para. 4-1, “Combining all tuning hyperparameters gives 384 cases corresponding to 384 possible models of 1D CNN methodology”; Tran, Pg. 30, Para. 2, “In the testing process, these optimal models will be compared to all other models according to their accuracy scores in order to evaluate reliability of the Grid Search model of 1D CNN methodology”; and Tran, Pg. 26, Para. 1, “theCNNGrid Search methodology is proposed with the consideration of not only hyperparameters that determine the network structure and training, but also the pre-processing input data as the number of input and differencing data . . . The number of accuracy scores such as Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and Mean Absolute Error (MAE) were also used. In the training process, the optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”, where the respective measure of performance, “optimal models with certain hyperparameters were obtained to meet the minimum requirement of the accuracy scores”, that is determined for each of the different hyperparameter combinations, “Combining all tuning hyperparameters gives 384 cases” “these optimal models will be compared to all other models”, comprises a respective mean absolute percentage error (MAPE) value, “The number of accuracy scores such as . . . Mean Absolute Percentage Error (MAPE)”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 and the rejection of Claim 4 above, and remain applicable here. Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Morgan in view of Ouyang and Freeman et al. (hereinafter Freeman) (Pat. Pub. No. US 2021/0042382 A1). Regarding Claim 7, Morgan in view of Ouyang teach the computing platform of claim 1, wherein the at least two different model types of the first set of time-series models and the at least two different model types of the second set of time-series models each comprises (Morgan, Fig. 4 and Morgan, Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”, where a set of time-series models, “a demand forecasting model is built . . . from combining two or more component models”, comprises time-series models of at least two different model types, “selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”; see also Morgan, Abstract, “Forecast models are built and used to score incoming sales data to predict future demand for items” and Morgan, Col. 3, Ln. 23-28, “The systems and methods are scalable and customizable to different applications of the forecast data. For example, a forecast may be generated on a weekly basis at the chain level for a group of retail stores. However, the forecast can be broken down to an individual store or an individual date”, where the set of time-series models, “Forecast models are built”, are configured to predict forecast values of a target time-series variable, “used to score incoming sales data to predict future demand for items”, for a second target timeframe, “individual store or an individual date”, that differs from the first target timeframe, “a forecast may be generated on a weekly basis at the chain level”; see also Morgan, Col. 14, Ln. 34-38, “In instances when a client requests a demand forecast for a particular day or a particular store, the disaggregation service 336 computes the demand forecast based on the aggregate demand forecasts stored in the forecasts data store 312”, which, in view of Ouyang each of the first and second sets of time-series models, “short and long VLM models”, include the above discussed at least two different model types, see Ouyang, Pg. 1, Abstract, “The VLM models form a probabilistic mixture of these varied length models. A combination of short and long VLM models and an ensemble of them are proposed to further enhance the prediction performance”) at least two of [the models] . . . (Morgan, Fig. 4 and Morgan, Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models”). The reasons for obviousness were discussed in regard to the rejection of Claim 1 above, and remain applicable here. Morgan in view of Ouyang do not explicitly disclose . . . (i) a Seasonal Autoregressive Integrated Moving-Average with eXogenous regressors (SARIMAX) type of time-series model, (ii) an Unobserved Components type of time-series model, (iii) an Exponential Smoothing type of time-series model, or (iv) a Prophet type of time-series model (but see Morgan, Col. 5, Ln. 5-18, “At operation 404, a demand forecasting model is built . . . In other embodiments, an enterprise “consensus” model is built from combining two or more component models . . . The component models are generally selected from time series forecasting models such as recurrent neural network (RNN) models or Autoregressive integrated moving average models (ARIMA), seasonal trend decomposition by LOESS (STL)”). However, Freeman teaches . . . [use of] (i) a Seasonal Autoregressive Integrated Moving-Average with eXogenous regressors (SARIMAX) type of time-series model, (ii) an Unobserved Components type of time-series model, (iii) an Exponential Smoothing type of time-series model, or (iv) a Prophet type of time-series model (Abstract, “A system and method for selecting an anomaly detection method from among a plurality of known anomaly detection methods includes selecting a set of anomaly detections methods based on characteristics of the time series, such as missing time steps, trend, drift, seasonality and concept drift” and Para. [0068], “for seasonality and trend, decomposition-based anomaly detection methods such as SARIMAX (seasonal auto-regressive integrated moving average with exogeneous variables) and Facebook Prophet perform the best. SARIMAX and Prophet have decomposition methods with components specifically built for seasonality and trend, which might explain their performance on this characteristic”). Before the effective filing date of the invention, it would have been obvious to one of ordinary skill in the art to combine the at least two different model types of the first set of time-series models and the at least two different model types of the second set of time-series models that each comprises at least two of the models of Morgan in view of Ouyang with the use of a Seasonal Autoregressive Integrated Moving-Average with eXogenous regressors (SARIMAX) type of time-series model, an Unobserved Components type of time-series model, an Exponential Smoothing type of time-series model, or a Prophet type of time-series model in order to select model types for inclusion in the ensemble seasonal forecasting model (Morgan, Col. 17, Ln. 6-14, “Because so many items within a retail context have changing demand based on seasonality, whether the items are on promotion, or whether the items are relevant to a particular holiday, it is important to be able to take into account seasonal effects. The ensemble model comprising weighted component models is advantageous in that the weighting of the various models can be modified to take into account changes for demand throughout the year based on seasonal effects”) that have components specifically built for time-series forecasting with seasonality and trend analysis, which will lead to optimal performance (Freeman, Para. [0068], “for seasonality and trend, decomposition-based anomaly detection methods such as SARIMAX (seasonal auto-regressive integrated moving average with exogeneous variables) and Facebook Prophet perform the best. SARIMAX and Prophet have decomposition methods with components specifically built for seasonality and trend, which might explain their performance on this characteristic”). Conclusion The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Reppucci et al. (“A Clustering-Base Approach For City Electricity Demand Forecasting”) discloses strengths and weaknesses of time-series prediction models, including SARIMAX and Prophet. Any inquiry concerning this communication or earlier communications from the examiner should be directed to MATTHEW BRYCE GOLAN whose telephone number is (571)272-5159. The examiner can normally be reached Monday through Friday, 8:00 AM to 5:00 PM ET. Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Alexey Shmatov can be reached at (571) 270-3428. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300. Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000. /MATTHEW BRYCE GOLAN/Examiner, Art Unit 2123 /ALEXEY SHMATOV/Supervisory Patent Examiner, Art Unit 2123
Read full office action

Prosecution Timeline

Feb 01, 2024
Application Filed
Aug 13, 2026
Non-Final Rejection mailed — §103
Oct 22, 2026
Interview Requested
Nov 04, 2026
Applicant Interview (Telephonic)
Nov 04, 2026
Examiner Interview Summary

Strategy Recommendation AI-generated — please review before filing

Get a prosecution strategy drawn from examiner precedents, rejection analysis, and claim mapping.
Typically takes 5-10 seconds — AI-generated, attorney review required before filing

Prosecution Projections

1-2
Expected OA Rounds
0%
Grant Probability
0%
With Interview (+0.0%)
Low
PTA Risk
Based on 9 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month