DETAILED ACTION
1. This communication is in response to the request for continued examination and corresponding amendments filed on April 13, 2026 for Application No. 18/120,895 in which Claims 1-3, 5-6, 9-11, 13-14, 17, and 19 are presented for examination.
Notice of Pre-AIA or AIA Status
2. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Continued Examination Under 37 CFR 1.114
3. A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 04/13/2026 has been entered.
Response to Arguments
4. The amendments filed on April 13, 2026 have been considered. Claims 1, 9, and 17 have been amended. Thus, Claims 1-3, 5-6, 9-11, 13-14, 17, and 19 are pending and presented for examination.
5. Applicant’s arguments filed April 13, 2026 with respect to the 35 U.S.C. 112(b) rejection have been fully considered and are persuasive. Thus, the 35 U.S.C. 112(b) rejection has been withdrawn.
6. Applicant’s arguments filed April 13, 2026 with respect to the 35 U.S.C. 101 rejection have been fully considered but they are not persuasive.
Applicant’s Arguments on Pgs. 11-12 of Arguments/Remarks state:
“The claims are not directed to a judicial exception under Prong Two of Step 2A because they integrate the asserted abstract idea into a practical application
Even assuming arguendo that the claims recite a judicial exception, the claims as amended are not "directed to" that exception under Step 2A, Prong Two because they integrate any alleged exception into a practical application. As set forth below, when the claims are evaluated as a whole, the additional elements reflect a specific technological improvement to automated machine learning model selection for time series forecasting, rather than a mere recitation of an abstract idea.
The amended claims recite specific additional elements that go beyond any alleged judicial exception and integrate the claims into a practical application. In particular, claim 1 as amended recites "training, by the one or more processors, the classifier model configured to classify a training feature vector to match a vector label distribution based on the class labels, wherein the vector label distribution represents the plurality of available machine learning models ordered based on performance, thereby optimizing the classifier model to predict a time series forecasting accuracy of each of the plurality of available machine learning models." This limitation specifies how the classifier model is trained to classify feature vectors to match a vector label distribution that represents the available machine learning models ordered by performance, thereby optimizing the classifier model. This is not a generic recitation of "training" but rather a specific technical implementation that configures the classifier model to achieve a particular technical result.
Additionally, claim 1 as amended recites "selecting, by the one or more processors based on the plurality of numerical performance scores, a machine learning model from the plurality of available machine learning models to analyze the set of time series data, wherein the trained classifier model enables the selecting without requiring each of the plurality of available machine learning models to be individually trained on the set of time series data." This limitation explicitly recites the practical improvement achieved by the claimed invention: automated model selection that circumvents time constraints by eliminating the need to individually train each available model on new time series data.
The specification describes the technological problem addressed by the claimed invention. As disclosed in paragraph [0019], "[a] current approach to ensure an adequate- performing forecast is to try to run data in a range of different models, with the hopes that one of them might perform well enough to be used in an application. However, this solution might not be viable for time-constrained projects." The specification further explains that "the amount of available computational resources may limit the ability to search through multiple different models. Therefore, a more efficient method to better assess which model should be used for forecasting is necessary" (see Id.).
The specification also describes the specific improvements provided by the claimed invention. As disclosed in paragraph [0023], "[t]hese systems and method improve on existing technologies because they circumvent time constraints and use intrinsic characteristics of time series datasets to assess the best possible machine learning model to use for assessing the time series datasets. Further, the systems and methods result in greater accuracy as well as greatly reduce the amount of time needed, from training to deployment, in a real-world scenario. Additionally, the training and use of the machine learning model(s) enables the systems and methods to process large datasets that conventional systems are unable to analyze as a whole. This results in improved processing time by the systems and methods."”
Examiner respectfully disagrees. First, the newly amended limitation of the Independent claims which recites "training, by the one or more processors, the classifier model configured to classify a training feature vector to match a vector label distribution based on the class labels, wherein the vector label distribution represents the plurality of available machine learning models ordered based on performance, thereby optimizing the classifier model to predict a time series forecasting accuracy of each of the plurality of available machine learning models.” does not recite a “specific technical implementation that configures the classifier model to achieve a particular technical result” – instead, the training, as currently drafted, is again merely “applied”. Examiner notes that the “training” limitation does not provide any specific method/process/strategy, but instead only states the solution of training a classifier model which is configured to classify a training feature vector to match a vector label distribution based on the class labels without significantly more. At this level of detail, Examiner asserts that this exemplifies a mere statement of what the classifier model is to accomplish, rather than how the classifier model is actually trained and/or how it functions. Thus, this amounts to adding the words “apply it” with the judicial exception, or mere instructions to implement an abstract idea on a computer at Step 2A Prong 2 and Step 2B.
Second, the newly amended limitation of the Independent claims which recites "selecting, by the one or more processors based on the plurality of numerical performance scores, a machine learning model from the plurality of available machine learning models to analyze the set of time series data” is considered to be abstract idea/mental process at Step 2A Prong 2. For example, a user may observe/analyze the plurality of numerical performance scores associated with a plurality of available machine learning models and accordingly use judgement/evaluation to select a machine learning model from the plurality, based on the numerical performance score. For example, the user may select a machine learning model with the highest numerical performance score, as this model may perform better and/or more accurately on the time series data. Further, the remaining limitation “[…] wherein the trained classifier model enables the selecting without requiring each of the plurality of available machine learning models to be individually trained on the set of time series data." again simply amounts to adding the words “apply it” or mere instructions to implement an abstract idea on a computer without significantly more at Step 2A Prong 2 and Step 2B. Again, this limitation merely recites what the classifier may accomplish without further expanding on the training configuration and/or training details.
Applicant further points to the instant specification which seemingly describes the technological problem addressed by the claimed invention. But again, as also stated by Examiner above, this technological improvement is not evident and/or reflected into the currently drafted claim language, as the claim language merely states these assertions without providing further technical details that support such configurations/improvements.
Applicant’s Arguments on Pg. 13 of Arguments/Remarks state:
“The amended claims are analogous to the claims found eligible in Desjardins. In particular, the amended claims recite specific training details, namely, classifying a training feature vector to match a vector label distribution that represents the plurality of available machine learning models ordered based on performance, thereby optimizing the classifier model. This reflects an improvement to how the classifier model operates, not merely a generic recitation of training.
The claims impose meaningful limits on any alleged exception because they do not merely recite instructions to apply an abstract idea on a generic computer. Rather, the claims recite a specific technical implementation: training a classifier model to classify feature vectors to match a vector label distribution representing models ordered by performance, and using that trained classifier to select a model without requiring each model to be individually trained on new data. This is a concrete technical solution to the problem of time-constrained model selection, not an abstract concept.
The ordered combination of elements works together to achieve a technical result. The claims recite a sequence of steps that begins with training and testing a plurality of available machine learning models across multiple time intervals, assessing their performance using WMAPE scores that are normalized and used to create class labels, training a classifier model to classify feature vectors to match a vector label distribution ordered by performance, and then using that trained classifier to select a machine learning model for new time series data without requiring each model to be individually trained on that new data. This ordered combination achieves automated model selection that circumvents time constraints, as described in the specification.
For at least the foregoing reasons, the claims integrate any alleged judicial exception into a practical application. The claims therefore satisfy Step 2A, Prong Two and are patent- eligible under 35 U.S.C. §101.”
Examiner respectfully disagrees for substantially the same reasons as stated above. Furthermore, Applicant states that “the amended claims are analogous to the claims found eligible in Desjardins” but again, this is not the case, as the “training” is generically recited and does not actually describe the processes/methods/strategy for actually configuring and technically training the models. Instead, the training is merely applied without further detail. Although Applicant states that the claims recite an ordered combination of elements which works together to achieve a technical result, the claims still recite an abstract idea and the judicial exception is not integrated into a practical application. Thus, the claims do not satisfy Step 2A Prong 2.
Applicant’s Arguments on Pgs. 14-15 of Arguments/Remarks state:
“C. The claims recite additional elements that amount to significantly more than any judicial exception
Even if the claims were considered to be directed to a judicial exception under Step 2A, the claims nonetheless recite additional elements that amount to significantly more than the alleged exception under Step 2B. Applicant submits that the claims include additional elements that transform the nature of the claims into patent-eligible subject matter.
The additional elements recited in the claims are not merely generic computer components performing generic functions. Claim 1 as amended recites "training, by the one or more processors, the classifier model configured to classify a training feature vector to match a vector label distribution based on the class labels, wherein the vector label distribution represents the plurality of available machine learning models ordered based on performance, thereby optimizing the classifier model to predict a time series forecasting accuracy of each of the plurality of available machine learning models." This limitation specifies a particular training configuration in which the classifier model is trained to classify feature vectors to match a vector label distribution that orders the available machine learning models by performance. This is not a generic recitation of training a model with data, but rather a specific technical implementation that configures the classifier model to achieve a particular result.
Additionally, claim 1 as amended recites "selecting, by the one or more processors based on the plurality of numerical performance scores, a machine learning model from the plurality of available machine learning models to analyze the set of time series data, wherein the trained classifier model enables the selecting without requiring each of the plurality of available machine learning models to be individually trained on the set of time series data." This limitation recites a non-conventional result achieved by the ordered combination of elements: the ability to select a machine learning model for new time series data without requiring each available model to be individually trained on that new data.
The ordered combination of elements provides a technical solution that is not conventional. Even if individual elements such as training machine learning models, generating WMAPE scores, or normalizing scores may be known, the specific combination recited in the claims is non-conventional and yields a technical benefit. The claims recite a sequence of steps that begins with training and testing a plurality of available machine learning models across multiple time intervals, assessing their performance using WMAPE scores that are normalized and used to create class labels, training a classifier model to classify feature vectors to match a vector label distribution ordered by performance, and then using that trained classifier to select a machine learning model for new time series data without requiring each model to be individually trained on that new data. This ordered combination achieves automated model selection that circumvents time constraints, as described in paragraph [0023] of the specification: "These systems and method improve on existing technologies because they circumvent time constraints and use intrinsic characteristics of time series datasets to assess the best possible machine learning model to use for assessing the time series datasets."
For at least the foregoing reasons, the claims recite significantly more than any alleged judicial exception and therefore are patent-eligible under Step 2B.
Accordingly, Applicant respectfully submits that claims 1-3, 5-6, 9-11, 13-14, 17, and 19 are patent eligible. Therefore, Applicant respectfully submits that the rejection under 35 U.S.C. § 101 is overcome and requests that the rejection be withdrawn.”
Examiner respectfully disagrees for substantially the same reasons as stated above. Again, the newly amended “training” limitation merely states what the classifier model is configured to and/or “trained” to do, without specifically outlining the processes/methods/strategy for training the model itself. While the limitation “specifies a particular configuration”, the training of the model is still not described and does not provide the specific technical implementation that would actually configure the model to achieve said results.
Further, regarding the newly added limitation “selecting […] wherein the trained classifier model enables the selecting without requiring each of the plurality of available machine learning models to be individually trained on the set of time series data” – the “wherein” clause again merely recites how the trained classifier model operates without significantly more. The fact that the classifier model enables the selecting without requiring individual training is a statement – it is not readily apparent how such a generic “training” of a classifier model may achieve this result. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Applicant is encouraged to amend the claims to clarify the “training” of the classifier model – rather than describing what the model is “trained” or “configured” to do, Applicant should focus on describing how the model is “trained” or “configured” to perform the specific operation of “classifying a training feature vector to match a vector label distribution”.
Thus, the 35 U.S.C. 101 rejection is maintained for all pending claims.
Claim Rejections - 35 USC § 101
7. 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.
8. Claims 1-3, 5-6, 9-11, 13-14, 17, and 19 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Claim 1:
Step 1: Claim 1 is a method type claim. Therefore, Claims 1-3 and 5-6 are directed to either a process, machine, manufacture, or composition of matter.
2A Prong 1: If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the “Mental Processes” grouping of abstract ideas. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation by mathematical calculation but for the recitation of generic computer components, then it falls within the “Mathematical Concepts” grouping of abstract ideas.
based on testing each of the plurality of available machine learning models, assessing a time series forecasting accuracy metric of each of the plurality of available machine learning models, for each of the multiple time intervals […] (mental process – assessing a time series forecasting accuracy metric of each of the plurality of models may be performed manually by a user observing/analyzing the results of testing each of the plurality of models and accordingly using judgement/evaluation to assess a time series forecasting accuracy metric for each of the models, based on the analysis of the testing results for each model)
(i) generating a weighted mean absolute percentage error (WMAPE) score for each of the plurality of available machine learning models for each of the multiple time intervals (mathematical process – generating a weighted mean absolute percentage error (WMAPE) score for each of a plurality of machine learning models for each of multiple time intervals may be performed by mathematical process, utilizing a formula/equation for calculating a WMAPE score and calculating the WMAPE score for each of the plurality of models for each of the time intervals – this is similarly supported by Applicant’s specification Par. [0063])
(ii) normalizing each WMAPE score according to a highest WMAPE score for a corresponding time series input to create a normalized score between zero and one (mathematical process – normalizing each WMAPE score according to a highest WMAPE score for a corresponding time series input may be performed by mathematical process, utilizing a formula/equation for normalization and normalizing each score according to an identified highest WMAPE score for a corresponding time series input, in order to create a normalized score between zero and one – this is similarly supported by Applicant’s specification Par. [0063])
(iii) selecting a lowest normalized score as a class label for training a classifier model (mental process – selecting a lowest normalized score as a class label may be performed manually by a user observing/analyzing the plurality of normalized scores and accordingly using judgement/evaluation to determine and select the lowest normalized score to be used as a class label that will be used for training a classifier model)
preparing, by the one or more processors, a set of time series data (mental process – other than reciting “by the one or more processors”, preparing a set of time series data may be performed manually by a user observing/analyzing the set of time series data and accordingly using judgement/evaluation to prepare said time series data. This is similarly supported by Applicant’s instant dependent claim 2, which recites that the “preparing” may be performed by an outlier removal technique and value imputation technique – those of which may be performed manually by a user observing/analyzing the set of time series data and accordingly using judgement/evaluation to remove outlier values and/or replace missing or null values)
extracting, by the one or more processors, a plurality of features from the set of time series data that was prepared (mental process – other than reciting “by the one or more processors”, extracting a plurality of features may be performed manually by a user observing/analyzing the prepared time series data and accordingly using judgement/evaluation to extract features from the prepared data based on insights formed from the analysis of the prepared time series data)
generating, by the one or more processors, a feature vector based on the plurality of features that were extracted (mental/mathematical process – other than reciting “by the one or more processors”, generating a feature vector may be performed manually by a user observing/analyzing the plurality of features and accordingly using judgement/evaluation to generate a feature vector based on the extracted features. Alternatively, generating a feature vector may also be performed by mathematical process, utilizing an algorithm/formula for feature vector generation)
selecting, by the one or more processors based on the plurality of numerical performance scores, a machine learning model from the plurality of available machine learning models to analyze the set of time series data […] (mental process – other than reciting “by the one or more processors”, selecting a machine learning model based on the plurality of numerical performance scores may be performed manually by a user observing/analyzing the plurality of numerical performance scores and accordingly using judgement/evaluation to select a machine learning model from the plurality to analyze the set of time series data, based on said analysis of the numerical performance score. For example, a user may select the machine learning model with the highest according numerical performance score)
2A Prong 2: This judicial exception is not integrated into a practical application.
Additional elements:
[…] by one or more processors […] (recited at a high-level of generality (i.e., as a generic one or more processors configured to perform the specific operations of the claim language without significantly more) such that it amounts to no more than mere instructions to apply the exception using generic computer components)
accessing, by one or more processors, a set of time series training data and a set of time series testing data, wherein each of the set of time series training data and the set of time series testing data is segmented into multiple time intervals (Adding insignificant extra-solution activity to the judicial exception – see MPEP 2106.05(g))
training, by the one or more processors for each of the multiple time intervals, each of a plurality of available machine learning models using the set of time series training data (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of training a machine learning model with previously determined data without significantly more)
testing, by the one or more processors for each of the multiple time intervals, each of the plurality of available machine learning models that was trained using the set of time series testing data (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of testing a plurality of machine learning models without significantly more – there are no further details provided regarding the testing, thus it is merely “applied”)
wherein each of the plurality of available machine learning models that was trained and tested is configured to perform a time series data analysis on time series data (Field of Use – limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception does not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application; in this case specifying that the models are configured to perform a time series data analysis on time series data does not integrate the exception into a practical application nor amount to significantly more – See MPEP 2106.05(h))
training, by the one or more processors, the classifier model configured to classify a training feature vector to match a vector label distribution based on the class labels, wherein the vector label distribution represents the plurality of machine learning models ordered based on performance, thereby optimizing the classifier model to predict a time series forecasting accuracy of each of the plurality of available machine learning models, wherein the training feature vector is associated with the set of time series data (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of training a machine learning model with previously determined data without significantly more – training is merely stated/recited and “applied” without further explanation on the actual training steps and/or configuration involved)
inputting, by the one or more processors into the classifier model, the feature vector (Adding insignificant extra-solution activity to the judicial exception – see MPEP 2106.05(g))
wherein the classifier model outputs a numerical performance score for each of the plurality of available machine learning models (Adding insignificant extra-solution activity to the judicial exception – see MPEP 2106.05(g))
wherein the numerical performance score is based on the time series forecasting accuracy metric and indicates a predicted forecasting performance of the available machine learning model for the set of time series data (Field of Use – limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception does not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application; in this case specifying that the numerical performance score is based on the time series forecasting accuracy metric and indicates a predicted forecasting performance does not integrate the exception into a practical application nor amount to significantly more – See MPEP 2106.05(h))
[…] wherein the trained classifier model enables the selecting without requiring each of the plurality of available machine learning models to be individually trained on the set of time series data (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of applying an already trained/configured machine learning model without significantly more)
2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
[…] by one or more processors […] (mere instructions to apply the exception using generic computer components cannot provide an inventive concept)
accessing, by one or more processors, a set of time series training data and a set of time series testing data, wherein each of the set of time series training data and the set of time series testing data is segmented into multiple time intervals (MPEP 2106.05(d)(II) indicates that merely “Receiving or transmitting data over a network” is a well-understood, routine, conventional function when it is claimed in a merely generic manner (as it is in the present claim). Thereby, a conclusion that the claimed limitation is well-understood, routine, conventional activity is supported under Berkheimer)
training, by the one or more processors for each of the multiple time intervals, each of a plurality of available machine learning models using the set of time series training data (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of training a machine learning model with previously determined data without significantly more. This cannot provide an inventive concept)
testing, by the one or more processors for each of the multiple time intervals, each of the plurality of available machine learning models that was trained using the set of time series testing data (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of testing a plurality of machine learning models without significantly more – there are no further details provided regarding the testing, thus it is merely “applied”. This cannot provide an inventive concept)
wherein each of the plurality of available machine learning models that was trained and tested is configured to perform a time series data analysis on time series data (Field of Use – limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception does not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application; in this case specifying that the models are configured to perform a time series data analysis on time series data does not integrate the exception into a practical application nor amount to significantly more – See MPEP 2106.05(h))
training, by the one or more processors, the classifier model configured to classify a training feature vector to match a vector label distribution based on the class labels, wherein the vector label distribution represents the plurality of machine learning models ordered based on performance, thereby optimizing the classifier model to predict a time series forecasting accuracy of each of the plurality of available machine learning models, wherein the training feature vector is associated with the set of time series data (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of training a machine learning model with previously determined data without significantly more – training is merely stated/recited and “applied” without further explanation on the actual training steps and/or configuration involved. This cannot provide an inventive concept)
inputting, by the one or more processors into the classifier model, the feature vector (MPEP 2106.05(d)(II) indicates that merely “Receiving or transmitting data over a network” is a well-understood, routine, conventional function when it is claimed in a merely generic manner (as it is in the present claim). Thereby, a conclusion that the claimed limitation is well-understood, routine, conventional activity is supported under Berkheimer)
wherein the classifier model outputs a numerical performance score for each of the plurality of available machine learning models (MPEP 2106.05(d)(II) indicates that merely “Presenting offers and gathering statistics” is a well-understood, routine, conventional function when it is claimed in a merely generic manner (as it is in the present claim). Thereby, a conclusion that the claimed limitation is well-understood, routine, conventional activity is supported under Berkheimer)
wherein the numerical performance score is based on the time series forecasting accuracy metric and indicates a predicted forecasting performance of the available machine learning model for the set of time series data (Field of Use – limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception does not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application; in this case specifying that the numerical performance score is based on the time series forecasting accuracy metric and indicates a predicted forecasting performance does not integrate the exception into a practical application nor amount to significantly more – See MPEP 2106.05(h))
[…] wherein the trained classifier model enables the selecting without requiring each of the plurality of available machine learning models to be individually trained on the set of time series data (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of applying an already trained/configured machine learning model without significantly more. This cannot provide an inventive concept)
For the reasons above, Claim 1 is rejected as being directed to an abstract idea without significantly more. This rejection applies equally to dependent claims 2-3 and 5-6. The additional limitations of the dependent claims are addressed below.
Regarding Claim 2:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 2 depends on.
performing, on the set of time series data by the one or more processors, (i) an outlier removal technique, (ii) a signal smoothing technique, and (iii) a value imputation technique (mental/mathematical process – performing an outlier removal technique, a signal smoothing technique, and a value imputation technique may be performed manually by a user (e.g., a user may remove outliers and/or impute missing or null values) and/or by mathematical process utilizing a formula/equation for outlier removal, signal smoothing, and value imputation)
Step 2A Prong 2 & Step 2B:
Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 1.
Regarding Claim 3:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 3 depends on.
extracting, by the one or more processors from the set of time series data that was prepared, at least one of: entropy, linearity, trend strength, seasonality strength, instability, or lumpiness (mathematical process/mental process – other than reciting “by the one or more processors” extracting at least one of entropy, linearity, trend strength, seasonality strength, instability, or lumpiness may be performed by mathematical process utilizing a formula/equation for calculating said features based on the prepared time series data. Alternatively, extracting one or more features of the time series data may be performed manually by a user observing/analyzing the set of prepared time series data and accordingly extracting features such as seasonality strength and/or trend strength based on said analysis)
Step 2A Prong 2 & Step 2B:
Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 1.
Regarding Claim 5:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 5 depends on.
generating, by the one or more processors, a set of stacking input data using at least a portion of the sets of univariate forecast data and a set of additional covariate data (mental process – other than reciting “by the one or more processors”, generating a set of stacking input data may be performed manually by a user observing/analyzing a portion of the sets of univariate forecast data and set of additional covariate data and accordingly using judgement/evaluation to generate a set of stacking input data based on said analysis)
analyzing, by a stacking machine learning model, the set of stacking input data to output a set of final forecast data associated with the set of time series data (mental process – other than reciting “by a stacking machine learning model”, analyzing the set of stacking input data may be performed manually by a user observing/analyzing the set of stacking input data and accordingly using judgement/evaluation to determine a set of final forecast data associated with the set of time series data)
Step 2A Prong 2 & Step 2B:
wherein each of the plurality of available machine learning models has associated a set of univariate forecast data associated with the set of time series data (Field of Use – limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception does not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application; in this case specifying that each of the models has associated a set of univariate data associated with the set of time series data does not integrate the exception into a practical application nor amount to significantly more – See MPEP 2106.05(h))
[…] stacking machine learning model […] (mere instructions to apply the exception using generic computer components cannot provide an inventive concept)
Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 1.
Regarding Claim 6:
Step 2A Prong 1: See the rejection of Claim 5 above, which Claim 6 depends on.
generating, by the one or more processors, a set of stacking training data using at least a portion of the sets of training univariate forecast data and a set of additional training covariate data (mental process – other than reciting “by the one or more processors”, generating a set of stacking training data may be performed manually by a user observing/analyzing a portion of the sets of univariate forecast data and set of additional covariate data and accordingly using judgement/evaluation to generate a set of stacking training data based on said analysis)
Step 2A Prong 2 & Step 2B:
wherein each of the plurality of available machine learning models has associated a set of training univariate forecast data (Field of Use – limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception does not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application; in this case specifying that each of the models has associated a set of univariate data associated with the set of time series data does not integrate the exception into a practical application nor amount to significantly more – See MPEP 2106.05(h))
training, by the one or more processors, the stacking machine learning model using the set of stacking training data and a set of historical data indicating known time series results (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner' s note: high level recitation of training a machine learning model with previously determined data without significantly more. This cannot provide an inventive concept)
Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 1.
Independent Claim 9 recites substantially the same limitations as Claim 1, in the form of a system, including generic computer components. The claim is also directed to performing mental processes/mathematical calculations without significantly more, therefore it is rejected under the same rationale.
For the reasons above, Claim 9 is rejected as being directed to an abstract idea without significantly more. This rejection applies equally to dependent claims 10-11 and 13-14. The additional limitations of the dependent claims are addressed below.
Claim 10 recites substantially the same limitations as Claim 2, in the form of a system, including generic computer components. The claim is also directed to performing mental processes/mathematical calculations without significantly more, therefore it is rejected under the same rationale.
Claim 11 recites substantially the same limitations as Claim 3, in the form of a system, including generic computer components. The claim is also directed to performing mental processes/mathematical calculations without significantly more, therefore it is rejected under the same rationale.
Claim 13 recites substantially the same limitations as Claim 5, in the form of a system, including generic computer components. The claim is also directed to performing mental processes/mathematical calculations without significantly more, therefore it is rejected under the same rationale.
Claim 14 recites substantially the same limitations as Claim 6, in the form of a system, including generic computer components. The claim is also directed to performing mental processes/mathematical calculations without significantly more, therefore it is rejected under the same rationale.
Independent Claim 17 recites substantially the same limitations as Claim 1, in the form of a non-transitory computer-readable storage medium, including generic computer components. The claim is also directed to performing mental processes/mathematical calculations without significantly more, therefore it is rejected under the same rationale.
For the reasons above, Claim 17 is rejected as being directed to an abstract idea without significantly more. This rejection applies equally to dependent claim 19. The additional limitations of the dependent claims are addressed below.
Claim 19 recites substantially the same limitations as Claim 5, in the form of a non-transitory computer-readable storage medium, including generic computer components. The claim is also directed to performing mental processes/mathematical calculations without significantly more, therefore it is rejected under the same rationale.
Allowable Subject Matter
9. No prior art rejection is made for Claims 1-3, 5-6, 9-11, 13-14, 17, and 19. However, these claims are still rejected under 35 U.S.C. 112(b) rejection and 35 U.S.C. 101 – abstract idea.
10. Examiner has disclosed Amiri et al. (US PG-PUB 20230022401) and Hyndman et al. (“Meta-learning how to forecast time series”), which are the closest prior art as compared to the instant application. In particular, Amiri teaches time series forecasting comprising determining one or more forecasters (machine learning models) to be used based on a type of time series data and previous training, with the best forecaster for the time series being selected based on the output of a time series classifier used to determine a level of confidence in the forecaster. Hyndman teaches a general framework, labelled FFORMS (Feature-based FORecast Model Selection), which selects forecast models based on features calculated from each time series – more specifically, Hyndman teaches computing a standardized symmetric mean absolute percentage error (sMAPE) across forecast models and selecting the model with the lowest average value of a mean absolute scaled error (MASE) and scaled sMAPE as the output class label. However, Amiri and Hyndman seemingly do not explicitly disclose the newly added limitations “based on testing each of the plurality of available machine learning models, assessing a time series forecasting accuracy metric of each of the plurality of available machine learning models, for each of the multiple time intervals, including: (i) generating a weighted mean absolute percentage error (WMAPE) score for each of the plurality of available machine learning models for each of the multiple time intervals,(ii) normalizing each WMAPE score according to a highest WMAPE score for a corresponding time series input to create a normalized score between zero and one, and(iii) selecting a lowest normalized score as a class label for training a classifier model” and “training, by one the one or more processors, the classifier model configured to classify a training feature vector to match a vector label distribution based on the class labels, wherein the vector label distribution represents the plurality of machine learning models ordered based on performance, thereby optimizing the classifier model to predict a time series forecasting accuracy of each of the plurality of available machine learning models, wherein the training feature vector is associated with the set of time series training data” included in Independent Claim 1 (and Independent Claims 9 and 17 which recite substantially the same limitations), in combination with the remaining limitations of the Independent claims.
Conclusion
11. Any inquiry concerning this communication or earlier communications from the examiner should be directed to Devika S Maharaj whose telephone number is (571)272-0829. The examiner can normally be reached Monday - Thursday 8:30am - 5:30pm.
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 on (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.
/DEVIKA S MAHARAJ/Examiner, Art Unit 2123
Prosecution Insights