DETAILED ACTION
Claim 1, 2, 4, 6, 7 and 9 are amended
Claim 3, 5, 8 and 10 are Originals
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 .
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-10 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception without significantly more.
Step 1 (The Statutory Categories): Is the claim to a process, machine, manufacture, or composition of matter? MPEP 2106.03.
Per Step 1, claim 1-5 is to a system (i.e., a machine), claim 6-10 to a method (i.e., a process). Thus, the claims are directed to statutory categories of invention. However, the claims are rejected under 35 U.S.C. 101 because they are directed to an abstract idea, a judicial exception, without reciting additional elements that integrate the judicial exception into a practical application.
The analysis proceeds to Step 2A Prong One.
Step 2A Prong One: Does the claim recite an abstract idea, law of nature, or natural phenomenon? MPEP 2106.04.
The abstract idea of independent claims 1 and 6 are (claim 1 being representative):
A spare-part quantity forecasting system for server repair based on a multiple-time-series model, comprising: a non-transitory computer-readable storage medium storing a plurality of instructions; and a hardware processor communicatively coupled to the non-transitory computer-readable storage medium, wherein the hardware processor executes the plurality of instructions to:
execute a data exploration with history experience data on an original feature of history time-series data to generate a feature differentiation rule, and split the original feature to form a plurality of differentiated features through the feature differentiation rule;
verify validity of the plurality of differentiated features through at least one of a statistical method and a machine learning model, and evaluate a response of the plurality of differentiated features for a change of a target variable to generate an evaluation result;
transmit the generated evaluation result to the data exploration and repeat executing the feature differentiation and the verifying validity to adjust the feature differentiation rule to re-form the plurality of differentiated features until the evaluation result meets a preset result;
split the target variable into N target sub- variables to correspond to the plurality of differentiated features, respectively, wherein N is a positive integer and equal to a quantity of the plurality of differentiated features formed finally;
store a multiple-time-series model having a plurality of time-series models, and use the N target sub-variables and the corresponding plurality of differentiated features as training data, and input the training data to the plurality of time- series models of the multiple-time-series model for training until the multiple-time-series model is trained completely; and
receive[[s]] current time-series data while a spare-part quantity is forecasted, input[[s]] the current time-series data to the multiple-time-series model which is pre-trained, make[[s]] the plurality of time-series models of the multiple-time-series model output forecasting results, make[[s]] the multiple-time-series model integrate the forecasting results as a spare-part quantity forecasting result, and store[[s]] the current time-series data as the history time-series data.
The abstract idea steps italicized above are those which could be performed mentally, including with pen and paper. The steps describe, at a high level, of obtaining and evaluation of data to determine a result. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, including Executing, verifying, splitting, transmitting, training, receiving time series data and forecasting the data, falls within the Mental Processes – Concepts Performed in the Human Mind grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
Additionally, and alternatively, the abstract idea steps italicized above relate to mathematical calculations (e.g., verify validity of the differentiated features through
PNG
media_image1.png
6
5
media_image1.png
Greyscale
at least one of a statistical method), which constitutes a process that, under its broadest reasonable interpretation, covers mathematical concepts. If a claim limitation, under its broadest reasonable interpretation, covers mathematical concepts, including mathematical relationships, mathematical formulas or equations, mathematical calculations, then it falls within the Mathematical Concepts grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two: Does the claim recite additional elements that integrate the judicial exception into a practical application? MPEP 2106.04.
This judicial exception is not integrated into a practical application because the additional elements are merely instructions to apply the abstract idea to a computer, as described in MPEP 2106.05(f).
Claim 1and 6 recites the following additional elements: A spare-part quantity forecasting system for server repair based on a multiple-time-series model, a non-transitory computer-readable storage medium, hardware processor, history experience data, history time-series data, a machine learning model, data exploration, training data, current time-series data.
Claim 2 and 7 recites the following additional elements: an autoregressive integrated moving average (ARIMA) model, a vector autoregression model, a vector error correction model, a long short-term memory model, a deep learning model, or a generalized additive model.
These elements are merely instructions to apply the abstract idea to a computer, per MPEP 2106.05(f). Applicant has only described generic computing elements and or machine learning in their specification, as seen in [0029] of applicant’s specification as filed for example.
Claim 4 and 9 recites the following additional elements: spare-part quantity forecasting system for server, history time-series data, hardware processor.
These elements are merely instructions to apply the abstract idea to a computer, per MPEP 2106.05(f). Applicant has only described generic computing elements and or machine learning in their specification, as seen in [0029] of applicant’s specification as filed for example.
Further, the combination of these elements is nothing more than a generic computing system applied to the tasks of the abstract idea. Because the additional elements are merely instructions to apply the abstract idea to a generic computing system, they do not integrate the abstract idea into a practical application, when viewed in combination. See MPEP 2106.05(f).
Therefore, per Step 2A Prong Two, the additional elements, alone and in combination, do not integrate the judicial exception into a practical application. The claim is directed to an abstract idea.
Step 2B (The Inventive Concept): Does the claim recite additional elements that amount to significantly more than the judicial exception? MPEP 2106.05.
Step 2B involves evaluating the additional elements to determine whether they amount to significantly more than the judicial exception itself.
The examination process involves carrying over identification of the additional element(s) in the claim from Step 2A Prong Two and carrying over conclusions from Step 2A Prong Two pertaining to MPEP 2106.05(f).
The additional elements and their analysis are therefore carried over: applicant has merely recited elements that facilitate the tasks of the abstract idea, as described in MPEP 2106.05(f).
Further, the combination of these elements is nothing more than a generic computing system with machine learning models. When the claim elements above are considered, alone and in combination, they do not amount to significantly more. See Specification [0007], [0009], and [0027-0029].
Therefore, per Step 2B, the additional elements, alone and in combination, are not significantly more. The claims are not patent eligible.
The analysis takes into consideration all dependent claims as well:
No prior Art Rejection Applied to claims 1-10
There is no prior art rejection to claims 1 and 6
Crowe et al (US20100257133A1), Jati et al (US20240045926A1), Singh et al (US 20240005177A1) and Jordan et al (US20200387832A1) are considered the closet references.
Crowe potentially teaches (Claims 1 being representative):
A spare-part quantity forecasting system for server repair based on a multiple-time-series model, comprising: a non-transitory computer-readable storage medium storing a plurality of instructions; and a hardware processor communicatively coupled to the non-transitory computer-readable storage medium, wherein the hardware processor executes the plurality of instructions to: execute a data exploration with history experience data on an original feature of history time-series data to generate a feature differentiation rule, and split the original feature to form a plurality of differentiated features through the feature differentiation rule;
{forecasting system; history experience data [0033]: A dynamic model 70 receives historical data 60 and future data 62 in order to perform forecasting 80. Examples of historical data 60 include the number of observations in a time series, demand series information, exogenous input series data, and historical seasonality data. Examples of future data 62 include a forecast horizon (e.g., number of time periods into the future which are forecasted; forecasts for next year are said to have a one-year forecast horizon), future exogenous inputs, and decision variables.
history time-series data to generate, suggested in [0003] [0136]: Given a time series 608 and a model specification 604, a fitted time series model 614 can be estimated. Given a fitted model 614 and a time series 608, a forecast function can be generated via a statistical forecasting engine 606 that efficiently encapsulates all information needed to provide a forecast 612 of the series when future values of price and advertising are provided and stored in concrete form in a forecast score file 610. The engine 606 can also generate a forecast evaluation 616. The forecast evaluation 616 can be useful to the statistical analyst to determine how well the forecast performed. Based upon the evaluation 616, the statistical analyst 602 can determine whether another iteration (e.g., using a different model) is needed to improve performance of the forecasting.
Multiple forecasting models is suggested in paragraph [0045]; Examples of models include: local level models, local trend models, local seasonal models, local models, ARIMA models, causal models, transformed models, intermittent demand models, external and user-defined models, etc.}
Crowe discloses {an evaluating module, configured to verify validity of the differentiated features suggested in [0109-0110]; plurality of statistical time-series models as suggested in [0042] [0138]; Evaluate a response to changes in variables as suggested in [0039]; Evaluation of generated result is suggested in [0134-0136]}.
verify validity of the plurality of differentiated features through at least one of a statistical method and a machine learning model, and evaluate a response of the plurality of differentiated features for a change of a target variable to generate an evaluation result;
{[0134] As can be seen by the above steps, other than the decision process 508 most of the computational effort takes place in the forecast scoring engine 504, which evaluates the forecast function based on the future values 510 of the causal factors. This computational effort is very low when compared to the process that generated the forecast (function) score file 502, which makes the iterative nature of the decision-making process 508 more tractable.
[0136] The engine 606 can also generate a forecast evaluation 616. The forecast evaluation 616 can be useful to the statistical analyst to determine how well the forecast performed. Based upon the evaluation 616, the statistical analyst 602 can determine whether another iteration (e.g., using a different model) is needed to improve performance of the forecasting.
[0138] The user 626 specifies an objective function 700 to the optimizing software 702. The optimizer 702 algorithmically varies the future causal factors 628, and the forecast scoring engine 622 provides the forecasts 624 to be evaluated by the objective function 700.}
Crowe discloses {receives current time-series data [0033]; integrate and store the forecasting results and historical time series data is disclosed in [0172-0173]; updating the time data is disclosed in [0166]}
receive[[s]] current time-series data while a spare-part quantity is forecasted, input[[s]] the current time-series data to the multiple-time-series model which is pre-trained, make[[s]] the plurality of time-series models of the multiple-time-series model output forecasting results, make[[s]] the multiple-time-series model integrate the forecasting results as a spare-part quantity forecasting result, and store[[s]] the current time-series data as the history time-series data.
{receives current time-series data, suggested in [0033] A dynamic model 70 receives historical data 60 and future data 62 in order to perform forecasting 80. Examples of historical data 60 include the number of observations in a time series, demand series information, exogenous input series data, and historical seasonality data. Examples of future data 62 include a forecast horizon (e.g., number of time periods into the future which are forecasted; forecasts for next year are said to have a one-year forecast horizon), future exogenous inputs, and decision variables.
[0172] A forecast results repository stores information about the forecasts, forecast evaluations, and forecast performance for each time series. The forecast results repository consists of several data sets. Since each time series has forecasts and statistics of fit associated with these forecasts, the forecast results repository will often be large. There is one forecast results repository for each time-stamped data set.}
Crowe does not explicitly disclose the following, however; Singh, considered the closet reference, potentially teaches {generated evaluation and analyzing result to [0017] [0034-0035]}:
transmit the generated evaluation result to the data exploration and repeat executing the feature differentiation and the verifying validity to adjust the feature differentiation rule to re-form the plurality of differentiated features until the evaluation result meets a preset result;
{generated evaluation and analyzing result are disclosed in [0017] [0034-0035];
[0034] Performance metric generation 320 may then obtain new time series data 321. For example, the notification of data set arrival The new time series data 321 may be obtained as well as one or more prior forecast(s) 322. Performance metric generation 320 may perform the various comparisons to generate the performance metrics and update model performance 315.
transmitting, analyzing and comparing the evaluated results against the stated criteria are disclosed in [0046-0050].}
The combination of Crowe and Singh does not explicitly disclose the following, however; Jordan, considered the closest reference, potentially teaches {splitting; configured to split, variable}
split the target variable into N target sub- variables to correspond to the plurality of differentiated features, respectively, wherein N is a positive integer and equal to a quantity of the plurality of differentiated features formed finally;
{splitting; configured to split, variable [0034-0035] [0075] and [0115-0116]:
[0035] A splitting rule 122 can be used to divide a subset of the data samples 116 (i.e., response variable values) based on the corresponding data samples 112 (i.e., independent variable values). For instance, a splitting rule 122 may divide response variable values into two partitions based on whether the corresponding independent variable values are greater than or less than a threshold independent variable value.
[0115-0116] In block 1002, the process 1000 involves determining a splitting rule for partitioning data samples in a decision tree. For example, the machine-learning model module 210 can access one or more independent variables xj and one or more threshold values θj. In some aspects, the machine-learning model module 210 selects a given independent variable xj and a corresponding threshold value θj such that an objective function is maximized.
[0116] In block 1004, the process 1000 involves partitioning, based on the splitting rule, data samples into a first tree region and a second tree region. For example, the machine-learning model module 210 can access data samples 112, which include values of various independent variables 114, from a data structure stored in the network-attached data stores 110 (or other memory device). The machine-learning model module 210 can identify a first subset of data samples 112 for which the independent variable xj is less than or equal to a threshold value θj. The machine-learning model module 210 can partition the data samples 112 into a left tree region, RL, having a boundary corresponding to xj≤θ1, and a right tree region, RR, having a boundary corresponding to xj>θ1.}
The combination of Crowe, Singh and Jordan does not explicitly disclose the following, however; Jati, considered the closet reference, potentially teaches {Storing a multiple-time-series model is disclosed in [0044]; training data is disclosed in [0067-0069]; training completely is disclosed in [0070] [0075].}
store a multiple-time-series model having a plurality of time-series models, and use the N target sub-variables and the corresponding plurality of differentiated features as training data, and input the training data to the plurality of time- series models of the multiple-time-series model for training until the multiple-time-series model is trained completely; and
{[0070] The training process may involve iteratively executing the respective models until a desired model is reached. For example, training may be performed until a predetermined level of accuracy is reached or a predetermined amount of data has been input to the model. The result is an initially trained student model 421, and initially-trained teacher models 422, 423, and 424, respectively.}
However, the combination of Crowe, Jati, Singh and Jordan does not teach the following:
A spare-part quantity {…} for server repair based on a multiple-time-series model, comprising:
a feature differentiation rule {…} differentiated features through the feature differentiation rule;
{…} the target {…} into N target sub- variables to correspond the differentiated features, respectively, wherein N is a positive integer and equal to a quantity of the differentiated features formed finally;
{…}, and use the N target sub-variables and the corresponding plurality of differentiated features as {…}
{…} spare-part quantity, {…} a spare-part quantity forecasting result {…}.
Additionally, the references considered also teach:
Crowe (US 20100257133), which teaches:
{[0042] With reference to FIG. 5, a model 100 can be selected from a plurality of different models 140. Each of the different models 140 can be models from different families of models, (e.g., ARIMA, UCM, ESM, and other families of models). A model selection list 140 can be used to specify a list of candidate model specifications and how to choose which model specification is best suited to forecast a particular time series. Different techniques can be utilized in determining how to select a model. As an illustration, the model selection techniques discussed in the Forecasting Provisional Application can be used.
[0054] The Autoregressive Integrated Moving Average Models (ARIMA) are used to forecast time series whose level, trend, or seasonal properties vary with time. These models predict the future values of the time series by applying non-seasonal or seasonal polynomial filters to the disturbances. Using different types of polynomial filters permits the modeling of various properties of the time series. (Series)=Disturbance Filter (Error)}
[0107] Once the model parameters are estimated, forecasts (predictions, prediction standard errors, prediction errors, and confidence limits) are made using the model parameter estimates, the model residual variance, and the full-range of data. If a model transformation was used, the forecasts are inverse transformed on a mean or median basis.
[0108] When it comes to decision-making based on the forecasts, the analyst decides whether to base the decision on the predictions, lower confidence limits, upper confidence limits or the distribution (predictions and prediction standard errors). If there is a greater penalty for over predicting, the lower confidence limit could be used. If there is a greater penalty for under predicting, the upper confidence limit could be used. Often for inventory control decisions, the distribution (mean and variance) is important.
[0074] FIG. 6 depicts that event data 140 can be used along with time series data 112 and a fitted model 110 in order to generate a forecast function 120. The factoring in of event data 114 when generating a forecast function 120 can improve the overall forecasting operations because the effects of events on time-stamped data may be an important factor in creating an accurate computer model of the data, for example, to improve the predictive ability of the model. Examples of events that may affect time-stamped data include advertising campaigns, retail promotions, strikes, natural disasters, policy changes, a data recording error, etc. Different types of events, event processing, and event processing methods are discussed in the Forecasting Provisional Application.}
Jati et al (US20240045926), teaches:
{[0059] The cloud computing environment 50 can be used to significantly improve both a training process 372 of the machine learning model and a predictive process 374 based on a trained machine learning model. For example, in 372, rather than requiring a data scientist/engineer or another user to collect the data, historical data may be stored by the assets 378 themselves (or through an intermediary, not shown) on the cloud computing environment 50. This can significantly reduce the collection time needed by the host platform 376 when performing predictive model training. For example, data can be directly and reliably transferred straight from its place of origin to the cloud computing environment 50. By using the cloud computing environment 50 to ensure the security and ownership of the collected data, smart contracts may directly send the data from the assets to the individuals that use the data for building a machine learning model. This allows for sharing of data among the assets 378.}
Thus, while appearing to teach the forecasting system, the additional references stop well short of the specificity required by the claim.
Independent claims 1 and 6 and the dependent claims 2-5 and 7-10 have no prior art applied.
Response to Argument
In response to the argument filled on March 03, 2026, regarding the 101 rejections. Examiner withdraws the 112(f), in view of the comment made on March 03, 2026.
Applicant argues that the amended claims are Integrated into a Practical Application. The Applicant further argues that the amended claim recites a specific "spare-part quantity forecasting system for server repair" which is not a disembodied mathematical calculation but a specific technical tool used to solve a specific problem in the field of server maintenance.
Examiner Respectfully disagrees.
The Examiner notes that the aspect of a spare-part quantity forecasting system for server repair, generating a feature differentiation rule, splitting, and verifying using a statistical method, target variable to generate an evaluation result such as splitting the N target sub- variables, the Examiner views as Mathematical Concepts grouping of abstract ideas.
Applicant argues that the amended claims "Iterative Feedback Loop" cannot be performed mentally “transmit the generated evaluation result to the data exploration and repeat executing the feature differentiation and the verifying validity to adjust the feature differentiation rule... until the evaluation result meets a preset result;”. Applicant further argues that the limitation creates an automated, recursive feedback loop and that the system does not merely "calculate" a result; it dynamically modifies its own internal rules ("adjust the feature differentiation rule") based on real-time evaluation feedback.
Examiner Respectfully disagrees.
The Examiner notes that transmitting the generated evaluation result to the data exploration, repeating and executing the feature differentiation and verifying validity to adjust the feature differentiation rule... until the evaluation result meets a preset result; the Examiner views these steps of identified abstract idea in the Step 2A Prong 1 Analysis and data exploration as an additional element in the Step 2A Prong 2 Analysis. Therefore, the Examiner maintains the Mental Processes – Concepts Performed in the Human Mind grouping of abstract idea. See initial Applicant specification [0031] and [0007-0008] of the Applicants amended specification.
Applicant argues that the amended claims specific data structure configuration (The “N-to-N” correlation) is not a generic mathematical split rather a specific data configuration technique where the target variables are structurally forced to match the differentiated features.
Examiner Respectfully disagrees.
The Examiner notes that the N-to-N" structure, which is a data configuration technique is merely a technical improvement to the abstract idea using generic technology. Applicant specification [0007-0009].
Applicant argues that the amended claims have a combination of (1) Iteratively adjusting the feature differentiation rule based on validity verification; (2) Structurally splitting the target variable into N sub-variables to strictly correspond to the N differentiated features; and (3) Integrating these into a multiple-time-series model for server repair; constitutes an ordered combination of elements that is not conventional or routine. This specific arrangement improves the technical field of spare-part forecasting by enabling the system to handle complex, multi-stage failure data that generic computers cannot handle effectively.
Examiner Respectfully disagrees.
The Examiner notes that these specific arrangements that improve spare-part forecasting by enabling the system to handle complex, multi-stage failure data are merely generic technology with no technical improvement rather an improvement to the abstract idea using generic technology.
Applicant argues that the independent claim 1 and 6 integrates the collaboration of mathematical algorithms and hardware resources into a specific practical application to solve the technical challenges of spare-part quantity forecasting for server repair.
Examiner Respectfully disagrees.
The Examiner notes that the integration of mathematical algorithms such as the N target sub-variable, differentiation rule and other statistical methods are directed merely to a Mathematical Concepts grouping of abstract ideas.
The Examiner maintains these claims recite an abstract idea.
Therefore, lacking any further argument, claims 1-10 are maintaining the 35 USC 101 rejection, as considered above in light of the amended claim limitation above.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Crowe et al (US20100257133), which teaches: Computer-implemented systems and methods for processing time series data that is indicative of a data generation activity occurring over a period of time. A model specification hierarchical data structure is used for storing characteristics that define a time series model.
Jati et al (US20240045926), which teaches: Storing a hierarchical time-series data set in memory, initially training a first time-series forecasting model based on a lower level of time-series data in the hierarchical data set, training a second time-series teaching forecasting model based on an upper level of time-series data from the hierarchical data set.
Singh et al (US 20240005177), which teaches: Performing monitoring operation for time series prediction models. Generating new time series forecast, and determining previously generated time series forecast with the use of machine learning model.
Jordan et al (20200387832), which teaches: Training machine-learning models for computing predicted responses and generating explanatory data for the models. Generating splitting rules, which includes decision trees for determining relationships between independent variables and a predicted response associated with the response variable and also adjusting models.
Nakayama et al (JP2019121296 A), which discloses; a requirement forecasting device, a requirement forecasting method, and a program are provided that realize both suppression of surplus inventory and suppression of product sales opportunity loss. A total required quantity acquisition unit acquires a total required quantity based on information related to a production plan stored in a production plan storage unit. The trend acquisition unit identifies a past period in which the distribution of operating days is similar to the target period. The trend acquisition unit receives from the performance storage unit information indicating the history of the amount of use of the article in the specified past period.
H. Lee and J. Kim, "A Predictive Model for Forecasting Spare Parts Demand in Military Logistics," 2018 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), Bangkok, Thailand, 2018, pp. 1106-1110
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to VICTOR ESONU whose telephone number is (571)272-4883. The examiner can normally be reached Monday - Friday 9:00 am - 5pm.
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, Monfeldt Sarah can be reached on (571) 270-1833. 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, vis it: 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.
/VICTOR ESONU/
Examiner, Art Unit 3629
/SARAH M MONFELDT/Supervisory Patent Examiner, Art Unit 3629
Prosecution Insights