DETAILED ACTION
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 .
This action is responsive to the Application filed on March 31, 2023. Claims 1-18 are pending in the case. Claim 1 is the independent claim.
This action is non-final.
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 1-18 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea (mental steps) without significantly more. This judicial exception is not integrated into a practical application because any additional elements amount to implementing the abstract idea on a generic computer. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Regarding independent claim 1, and relying on the evaluation flowchart in MPEP 2106:
Step 1 (Is the claim to a process, machine, manufacture, or composition of matter?): Yes. Claim 1 is a method of fraud detection (process).
Step 2a Prong One (Does the claim recite an abstract idea?): Yes. Claim 1 recites:
applying a Hellinger decision tree to detect fraudulent transactions in the data set of financial transactions whereby using a Hellinger distance is part of the applying (a mental process involving performing a mathematical calculation, with or without the aid of pen and paper, such as a human mentally applying a decision tree to detect fraudulent transactions in the data set of financial transactions, including mentally performing calculations necessary to use/apply the Hellinger distance when applying the decision tree).
Under the broadest reasonable interpretation, these steps may be performed mentally, using mental observation and mental determination, including by a human using a physical aid such as pen and paper, including a human mentally performing observations and mentally performing mathematical calculations, and therefore correspond to the Mental Processes grouping.
Step 2a Prong Two (Does the claim recite additional elements that integrate the judicial exception into a practical application?): No. Claim 1 additionally recites:
receiving a data set of financial transactions (insignificant extra-solution activity of transmitting data over a network as discussed in MPEP 2106.05(g));
that the receiving is at a processor and that the applying of the Hellinger decision tree is using the processor (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Therefore, in view of the considerations set forth in MPEP 2106.04(d), 2106.05(a)-(c) and (e)-(h), the additional elements as disclosed above alone or in combination do not integrate the judicial exception into a practical application as they are merely implementing the abstract idea using generic computer components.
Step 2b (Does the claim recite additional elements that amount to siqnificantly more than the judicial exception): No. Relying on the same analysis as Step 2a Prong Two (see MPEP 2106.05.I.A: Limitations that the courts have found not to be enough to qualify as “significantly more” when recited in a claim with a judicial exception include:…Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, e.g., a limitation indicating that a particular function such as creating and maintaining electronic records is performed by a computer, as discussed in Alice Corp., 573 U.S. at 225-26, 110 USPQ2d at 1984 (see MPEP 2106.05(f));…Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception...; Adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g);…)), claim 1 does not recite any additional elements that amount to significantly more than the abstract idea. As discussed above, Claim 1 recites:
receiving a data set of financial transactions (insignificant extra-solution activity as discussed in MPEP 2106.05(g) which can be reevaluated as well-understood, routine, conventional activity such as transmitting data over a network as discussed in MPEP 2106.05(d))
that the receiving is at a processor and that the applying of the Hellinger decision tree is using the processor (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Moreover, to the extent that the claims recite that the method is for fraud detection, that the data set is of financial transactions, and that the Hellinger decision tree is applied to detect fraudulent transactions, these limitations also describe a field of use and technological environment as discussed in MPEP 2106.05(h).
The additional elements as discussed above, in combination with the abstract idea, are not sufficient to amount to significantly more than the judicial exception as they are well, understood, routine and conventional activity as disclosed in combination with generic computer functions and components used to implement the abstract idea.
Regarding dependent claim 2:
Step 2a Prong One: incorporates the rejection of claim 1.
Step 2a Prong Two: the claims additionally recite wherein the Hellinger decision tree is part of a machine learning algorithm (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Step 2b: the claims additionally recite wherein the Hellinger decision tree is part of a machine learning algorithm (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Regarding dependent claim 3:
Step 2a Prong One: incorporates the rejection of claim 1; the claim further recite wherein the Hellinger distance tree is configured to capture divergence between positive and negative class distribution without being dominated by class imbalance (a mental process involving performing a mathematical calculation, with or without the aid of pen and paper, such as a human mentally configuring the decision tree based on calculations to capture divergence in positive and negative class distribution without being dominated by class imbalance).
Step 2a Prong Two: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Step 2b: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Regarding dependent claim 4:
Step 2a Prong One: incorporates the rejection of claim 1; the claims further recite wherein the Hellinger decision tree is configured to use class prior to estimate counts of positives and negatives in each node (a mental process involving performing a mathematical calculation, with or without the aid of pen and paper, such as a human mentally configuring the decision tree based on calculations to use class prior to estimate counts of positives and negatives in each node).
Step 2a Prong Two: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Step 2b: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Regarding dependent claim 5:
Step 2a Prong One: incorporates the rejection of claim 1; the claim further recite limiting a size of the Hellinger decision tree after a tree node reaches a maximum height thereby avoiding overfitting (a mental process of evaluation, such as a human mentally determining to limit the size of the decision tree using a maximum height to avoid overfitting).
Step 2a Prong Two: the claim additionally recites using the processor (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Step 2b: the claim additionally recites using the processor (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Regarding dependent claim 6:
Step 2a Prong One: incorporates the rejection of claim 1; the claim further recite wherein the Hellinger decision tree is a positive and unbalanced Hellinger decision tree, and wherein the data set of fraudulent transactions is imbalanced positive and unlabeled data (a mental process of evaluation, such as a human mentally determining to configure the decision tree as a positive and unbalanced decision tree and to use a data set of positive and unlabeled data).
Step 2a Prong Two: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Step 2b: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Regarding dependent claim 7:
Step 2a Prong One: incorporates the rejection of claim 1; the claims further recite receiving an estimated fraud rate prior to the applying (a mental process of evaluation, such as a human mentally observing/determining an estimated fraud rate).
Step 2a Prong Two: the claims additionally recite that the receiving is at the processor (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f) and a field of use and technological environment as discussed in MPEP 2106.05(h)).
Step 2b: the claims additionally recite that the receiving is at the processor (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f) and a field of use and technological environment as discussed in MPEP 2106.05(h)).
Regarding dependent claim 8:
Step 2a Prong One: incorporates the rejection of claim 1.
Step 2a Prong Two: the claim additionally recite wherein the Hellinger decision tree is used as a base learner in a modified random forest (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f) and a field of use and technological environment as discussed in MPEP 2106.05(h)).
Step 2b: the claim additionally recite wherein the Hellinger decision tree is used as a base learner in a modified random forest (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f) and a field of use and technological environment as discussed in MPEP 2106.05(h)).
Regarding dependent claim 9:
Step 2a Prong One: incorporates the rejection of claims 1 and 8; the claim additionally recites wherein the Hellinger decision tree is a positive and unbalanced Hellinger decision tree, and wherein the data set of fraudulent transactions is imbalanced positive and unlabeled data (a mental process of evaluation, such as a human mentally determining to configure the decision tree as a positive and unbalanced decision tree and to use a data set of imbalanced positive and unlabeled data).
Step 2a Prong Two: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Step 2b: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Regarding dependent claim 10:
Step 2a Prong One: incorporates the rejection of claim 1, 8, and 9; the claims further recite wherein the Hellinger decision tree is configured to consider random feature selection when initializing a tree node (a mental process of evaluation, such as a human mentally determining to configure the decision tree to consider random feature selection when initializing a tree node).
Step 2a Prong Two: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Step 2b: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Regarding dependent claim 11:
Step 2a Prong One: incorporates the rejection of claim 1, 8, and 9; the claims further recite wherein the Hellinger decision tree is configured to use a size of a stratified bootstrap sample and a class prior (a mental process of evaluation, such as a human mentally determining to configure the decision tree to use a size of a stratified bootstrap sample and a class prior).
Step 2a Prong Two: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Step 2b: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Regarding dependent claim 12:
Step 2a Prong One: incorporates the rejection of claim 1; the claims further recite wherein the Hellinger distance is used to capture divergence between positive and negative class distribution (a mental process of evaluation, such as a human mentally using, including via mathematical calculation, the Hellinger distance to capture divergence between positive and negative class distribution).
Step 2a Prong Two: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Step 2b: the claims do not recite any other limitations in addition to the abstract idea discussed above.
Regarding dependent claim 13:
Step 2a Prong One: incorporates the rejection of claim 1; the claim further recite receiving a positive and unlabeled dataset (a mental process of evaluation, such as a human mentally observing a positive and unlabeled dataset).
Step 2a Prong Two: the claim further recites that the receiving is at the processor; and training the Hellinger decision tree with the positive and unlabeled dataset (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f) and a field of use and technological environment as discussed in MPEP 2106.05(h)).
Step 2b: the claims do not recite any other limitations in addition to the abstract idea discussed above. the claim further recites that the receiving is at the processor; and training the Hellinger decision tree with the positive and unlabeled dataset (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f) and a field of use and technological environment as discussed in MPEP 2106.05(h)).
Regarding dependent claim 14:
Step 2a Prong One: incorporates the rejection of claim 1.
Step 2a Prong Two: the claim further recites wherein the positive and unlabeled dataset is unbalanced (a field of use and technological environment as discussed in MPEP 2106.05(h)).
Step 2b: the claim further recites wherein the positive and unlabeled dataset is unbalanced (a field of use and technological environment as discussed in MPEP 2106.05(h)).
Regarding dependent claim 15:
Step 2a Prong One: incorporates the rejection of claim 1.
Step 2a Prong Two: the claim further recites wherein the financial transactions are credit card transactions (a field of use and technological environment as discussed in MPEP 2106.05(h)).
Step 2b: the claim further recites wherein the financial transactions are credit card transactions (a field of use and technological environment as discussed in MPEP 2106.05(h)).
Regarding dependent claim 16:
Step 2a Prong One: incorporates the rejection of claim 1.
Step 2a Prong Two: the claim further recites wherein the financial transactions are insurance transactions (a field of use and technological environment as discussed in MPEP 2106.05(h)).
Step 2b: the claim further recites wherein the financial transactions are insurance transactions (a field of use and technological environment as discussed in MPEP 2106.05(h)).
Regarding dependent claim 17:
Step 2a Prong One: incorporates the rejection of claim 1.
Step 2a Prong Two: the claim further recites wherein the system is a computer or a server (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Step 2b: the claim further recites wherein the system is a computer or a server (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Regarding dependent claim 18:
Step 2a Prong One: incorporates the rejection of claim 1.
Step 2a Prong Two: the claim further recites a non-transitory computer readable medium storing a program configured to instruct the processor to execute the method of claim 1 (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Step 2b: the claim further recites a non-transitory computer readable medium storing a program configured to instruct the processor to execute the method of claim 1 (mere instructions to apply the exception using generic computer components as discussed in MPEP 2106.05(f)).
Therefore, in view of the considerations set forth in MPEP 2106.04(d), 2106.05(a)-(c) and (e)-(h), the additional elements as recited in the dependent claims discussed above alone or in combination do not integrate the judicial exception into a practical application as they are mere insignificant extra solution activity, combined with implementing the abstract idea using generic computer components, and limitations describing a field of use or technological environment. The additional elements as discussed above, in combination with the abstract idea, are not sufficient to amount to significantly more than the judicial exception as they are well, understood, routine and conventional activity as disclosed in combination with generic computer functions and components used to implement the abstract idea, and limitations describing a field of use or technological environment.
Claim Rejections – 35 USC § 102
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 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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1-3, 7, 12, and 15 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by A. Dal Pozzolo, R. Johnson, O. Caelen, S. Waterschoot, N. V. Chawla and G. Bontempi, "Using HDDT to avoid instances propagation in unbalanced and evolving data streams," 2014 International Joint Conference on Neural Networks (IJCNN), Beijing, China, 2014, pp. 588-594, doi: 10.1109/IJCNN.2014.6889638. (hereinafter Pozzolo).
With respect to claim 1, Pozzolo teaches a method of fraud detection comprising:
receiving, at a processor, a data set of financial transactions (e.g. page 588, second column, first and second full paragraphs, highly unbalanced credit card fraud dataset with concept drift; page 591, second column, first paragraph, experimental setup includes using real-world credit card dataset which is highly unbalanced and whose frauds are changing in type and distribution); and
applying a Hellinger decision tree, using the processor, to detect fraudulent transactions in the data set of financial transactions whereby using a Hellinger distance is part of the applying (e.g. page 588, second column, first and second full paragraphs, batch ensemble model combination based on Hellinger Distance and Information Gain, tested with different types of datasets including unbalanced credit card fraud dataset; page 589, first column, first full paragraph, Hellinger distance as splitting criteria in decision trees to improve accuracy in unbalanced problems; page 589, first column, final paragraph, describing Hellinger distance decision trees; page 591, first column, describing experimental setup in section VI using Hellinger Distance Decision Tree (HDDT); framework implemented in Java, using Weka implementation of HDDT; page 591, second column, first paragraph, experimental setup includes using real-world credit card dataset which is highly unbalanced and whose frauds are changing in type and distribution).
With respect to claim 2, Pozzolo teaches all of the limitations of claim 1 as previously discussed, and further teaches wherein the Hellinger decision tree is part of a machine learning algorithm (e.g. page 588, first column, final paragraph, static learning setting; learning from non-stationary data streams; using HDDT as base learner; page 588, second column, first and second full paragraphs, batch ensemble model combination based on Hellinger Distance and Information Gain; page 591, first full paragraph, using decision tree as base learner (comparing C4.5 and HDDT)).
With respect to claim 3, Pozzolo teaches all of the limitations of claim 1 as previously discussed, and further teaches wherein the Hellinger decision tree is configured to capture divergence between positive and negative class distribution without being dominated by class imbalance (e.g. page 589, first column, first and second paragraphs, majority class is negative and minority class is positive; Hellinger distance quantifies similarity between two probability distributions; page 589, first column, seventh paragraph (titled “IV. Hellinger Distance Decision Trees”), for each feature f calculating distance between the classes over all of the feature’s partitions; Hellinger distance between positive class and negative class; page 589, second column, first paragraph, class imbalance ratio does not influence the distance calculation).
With respect to claim 7, Pozzolo teaches all of the limitations of claim 1 as previously discussed, and further teaches the method further comprising receiving an estimated fraud rate at the processor prior to the applying (e.g. page 590, second column, second paragraph, indicating that the number of frauds in each chunk of streaming data is usually less than 1% (i.e. where this indicates an estimated fraud rate); page 591, second column, first paragraph, indicating that the fraud rate of the credit card dataset is known to be 0.15% of the transactions, i.e., prior to the use of the Hellinger decision tree with the dataset).
With respect to claim 12, Pozzolo teaches all of the limitations of claim 1 as previously discussed, and further teaches wherein the Hellinger distance is used to capture divergence between positive and negative class distribution (e.g. page 589, first column, first and second paragraphs, majority class is negative and minority class is positive; Hellinger distance quantifies similarity between two probability distributions; page 589, first column, seventh paragraph (titled “IV. Hellinger Distance Decision Trees”), for each feature f calculating distance between the classes over all of the feature’s partitions; Hellinger distance between positive class and negative class).
With respect to claim 15, Pozzolo teaches all of the limitations of claim 1 as previously discussed, and further teaches wherein the financial transactions are credit card transactions (e.g. page 588, second column, first and second full paragraphs, highly unbalanced credit card fraud dataset with concept drift; page 591, second column, first paragraph, experimental setup includes using real-world credit card dataset which is highly unbalanced and whose frauds are changing in type and distribution).
Claim Rejections – 35 USC § 103
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.
The factual inquiries set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims under pre-AIA 35 U.S.C. 103(a), the examiner presumes that the subject matter of the various claims was commonly owned at the time any inventions covered therein were made absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and invention dates of each claim that was not commonly owned at the time a later invention was made in order for the examiner to consider the applicability of pre-AIA 35 U.S.C. 103(c) and potential pre-AIA 35 U.S.C. 102€, (f) or (g) prior art under pre-AIA 35 U.S.C. 103(a).
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Pozzolo in view of Gribelyuk et al. (US 10754946 B1).
With respect to claim 5, Pozzolo teaches all of the limitations of claim 1 as previously discussed. Pozzolo does not explicitly disclose limiting a size of the Hellinger decision tree after a tree node reaches a maximum height using the processor thereby avoiding overfitting. However, Gribelyuk teaches limiting a size of the Hellinger decision tree after a tree node reaches a maximum height using the processor thereby avoiding overfitting (e.g. col. 12 lines 19-46, discussing parameters for decision trees, including setting a maximum depth (i.e. height) in order to prevent overfitting).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo and Gribelyuk in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), to incorporate the teachings of Gribelyuk (directed to a machine learning approach to modeling entity behavior, including using decision trees) to include the capability to limit the size of the decision tree to a maximum height/depth in order to avoid overfitting. One of ordinary skill would have been motivated to perform such a modification in order to avoid overfitting as described in Gribelyuk (col. 12 lines 19-46).
Claims 8, 17, and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Pozzolo in view of Guo et al. (US 10754946 B1).
With respect to claim 8, Pozzolo teaches all of the limitations of claim 1 as previously discussed, and further teaches wherein the Hellinger decision tree is used as a base learner in a modified random forest (e.g. page 588, first column, final paragraph, static learning setting; learning from non-stationary data streams; using HDDT as base learner; page 588, second column, first and second full paragraphs, batch ensemble model combination based on Hellinger Distance and Information Gain; page 591, first full paragraph, using decision tree as base learner (comparing C4.5 and HDDT)).
Hellinger does not explicitly disclose wherein the decision tree is used in a modified random forest. However, Guo teaches wherein the decision tree is used in a modified random forest (e.g. paragraph 0043, generating decision trees in modified random forest model).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo and Guo in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), to incorporate the teachings of Guo (directed to facilitating automatic handling of incomplete data in a random forest model) to include the capability to implement the Hellinger decision tree (i.e. of Pozzolo) in a modified random forest. One of ordinary skill would have been motivated to perform such a modification in order to address challenges in handling missing data values in a dataset as described in Guo (paragraph 0023).
With respect to claim 17, Pozzolo teaches all of the limitations of claim 1 as previously discussed. Assuming arguendo that Pozzolo does not explicitly disclose a system configured to perform the method of claim 1, wherein the system is a computer or a server, Guo teaches a system configured to perform the method of claim 1, wherein the system is a computer or server (e.g. paragraph 0027, Fig. 1, system including a computing device 102, including modified random forest component that can facilitate automatically analyzing one or more datasets).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo and Guo in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), to incorporate the teachings of Guo (directed to facilitating automatic handling of incomplete data in a random forest model) to include the capability to implement the Hellinger decision tree (i.e. of Pozzolo) using a computing device. One of ordinary skill would have been motivated to perform such a modification in order to address challenges in handling missing data values in a dataset as described in Guo (paragraph 0023).
With respect to claim 18, Pozzolo teaches all of the limitations of claim 1 as previously discussed. Assuming arguendo that Pozzolo does not explicitly disclose a non-transitory computer readable medium storing a program configured to instruct the processor to execute the method of claim 1, Guo teaches a non-transitory computer readable medium storing a program configured to instruct the processor to execute the method of claim 1 (e.g. paragraph 0028, memory 108 storing computer executable components including modified random forest component 104 and associated components, along with processor 106 that executes the computer executable components stored in memory 108).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo and Guo in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), to incorporate the teachings of Guo (directed to facilitating automatic handling of incomplete data in a random forest model) to include the capability to implement the Hellinger decision tree (i.e. of Pozzolo) via a program stored in a computer memory. One of ordinary skill would have been motivated to perform such a modification in order to address challenges in handling missing data values in a dataset as described in Guo (paragraph 0023).
Claims 6, 13, 14, and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Pozzolo in view of C. Phua, V. Lee, K. Smith and R. Gayler, "A Comprehensive Survey of Data Mining-based Fraud Detection Research," Arxiv.org, arXiv:1009.6119. (hereinafter Phua).
With respect to claim 6, Pozzolo teaches all of the limitations of claim 1 as previously discussed, and further teaches wherein the Hellinger decision tree is a positive and unbalanced Hellinger decision tree, and wherein the data set of fraudulent transactions is imbalanced positive data (e.g. page 589, first and second lines, indicating that the majority class refers to negative and the minority class refers to positive data in the dataset; page 589, first column final paragraph, describing HDDT as being based on distances between positive and negative classes over all of a feature’s partitions; page 591, first column, second paragraph, indicating the HDDT is used as a base learner on the unbalanced data; page 591, second column first paragraph, describing real-world credit card dataset as being highly unbalanced but including 0.15% of transactions as being fraudulent (i.e. positive); page 593, first column, final paragraph indicating that the fraud dataset is extremely unbalanced and exhibiting concept drift within the minority class, where the HDDT performs very well on the dataset; i.e. the dataset is unbalanced and positive (includes positive data), and the HDDT is trained on this unbalanced and positive dataset).
Hellinger does not explicitly disclose that the dataset is unlabeled data. However, Phua teaches that the dataset is unlabeled data (e.g. page 5 second column, second paragraph, indicating that some research recommends use of unlabelled data in fraud detection use cases; page 8, first column, third and fourth paragraphs, describing unsupervised approaches with unabelled data, including in conjunction with Hellinger distance for comparing probability distributions and giving suspicions scores, and detecting statistical outliers using Hellinger distance).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo and Phua in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), to incorporate the teachings of Phua (directed to fraud detection research) to include the capability to utilize an unlabeled dataset. One of ordinary skill would have been motivated to perform such a modification in order to address criticisms associated with using labelled data to detect fraud as described in Phua (page 5, second column, second paragraph).
With respect to claim 13, Pozzolo teaches all of the limitations of claim 1 as previously discussed, and further teaches the method further comprising: receiving a positive dataset at the processor; and training the Hellinger decision tree with the positive dataset (e.g. page 589, first and second lines, indicating that the majority class refers to negative and the minority class refers to positive data in the dataset; page 589, first column final paragraph, describing HDDT as being based on distances between positive and negative classes over all of a feature’s partitions; page 591, first column, second paragraph, indicating the HDDT is used as a base learner on the unbalanced data; page 591, second column first paragraph, describing real-world credit card dataset as being highly unbalanced but including 0.15% of transactions as being fraudulent (i.e. positive); page 593, first column, final paragraph indicating that the fraud dataset is extremely unbalanced and exhibiting concept drift within the minority class, where the HDDT performs very well on the dataset; i.e. the dataset is unbalanced and positive (includes positive data), and the HDDT is trained on this unbalanced and positive dataset).
Pozzolo does not explicitly disclose that the data set is an unlabeled dataset. However, Phua teaches that the data set is an unlabeled dataset (e.g. page 5 second column, second paragraph, indicating that some research recommends use of unlabelled data in fraud detection use cases; page 8, first column, third and fourth paragraphs, describing unsupervised approaches with unabelled data, including in conjunction with Hellinger distance for comparing probability distributions and giving suspicions scores, and detecting statistical outliers using Hellinger distance).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo and Phua in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), to incorporate the teachings of Phua (directed to fraud detection research) to include the capability to utilize an unlabeled dataset. One of ordinary skill would have been motivated to perform such a modification in order to address criticisms associated with using labelled data to detect fraud as described in Phua (page 5, second column, second paragraph).
With respect to claim 14, Pozzolo teaches all of the limitations of claim 13 as previously discussed, and further teaches wherein the positive dataset is unbalanced (e.g. page 589, first and second lines, indicating that the majority class refers to negative and the minority class refers to positive data in the dataset; page 589, first column final paragraph, describing HDDT as being based on distances between positive and negative classes over all of a feature’s partitions; page 591, first column, second paragraph, indicating the HDDT is used as a base learner on the unbalanced data; page 591, second column first paragraph, describing real-world credit card dataset as being highly unbalanced but including 0.15% of transactions as being fraudulent (i.e. positive); page 593, first column, final paragraph indicating that the fraud dataset is extremely unbalanced and exhibiting concept drift within the minority class, where the HDDT performs very well on the dataset; i.e. the dataset is unbalanced and positive (includes positive data), and the HDDT is trained on this unbalanced and positive dataset).
Pozzolo does not explicitly disclose that the dataset is unlabeled. However, Phua teaches that the dataset is unlabeled (e.g. page 5 second column, second paragraph, indicating that some research recommends use of unlabelled data in fraud detection use cases; page 8, first column, third and fourth paragraphs, describing unsupervised approaches with unabelled data, including in conjunction with Hellinger distance for comparing probability distributions and giving suspicions scores, and detecting statistical outliers using Hellinger distance).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo and Phua in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), to incorporate the teachings of Phua (directed to fraud detection research) to include the capability to utilize an unlabeled dataset. One of ordinary skill would have been motivated to perform such a modification in order to address criticisms associated with using labelled data to detect fraud as described in Phua (page 5, second column, second paragraph).
With respect to claim 16 Pozzolo teaches all of the limitations of claim 1 as previously discussed. Pozzolo does not explicitly disclose wherein the financial transactions are insurance transactions. However, Phua teaches wherein the financial transactions are insurance transactions (e.g. page 8, first column, first paragraph, indicating fraud detection based on analyzing twelve months worth of insurance claims; page 8, first column, fourth paragraph, detecting statistical outliers using Hellinger distance on medical insurance data).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo and Phua in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), to incorporate the teachings of Phua (directed to fraud detection research) to include the capability to utilize as, the dataset of financial transactions, a dataset of insurance transactions. One of ordinary skill would have been motivated to perform such a modification in order to address criticisms associated with using labelled data to detect fraud as described in Phua (page 5, second column, second paragraph).
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Pozzolo in view of Guo, further in view of Phua.
With respect to claim 9, Pozzolo in view of Guo teaches all of the limitations of claim 8 as previously discussed, and further Pozzolo teaches wherein the Hellinger decision tree is a positive and unbalanced Hellinger decision tree, and wherein the data set of fraudulent transactions is imbalanced positive data (e.g. page 589, first and second lines, indicating that the majority class refers to negative and the minority class refers to positive data in the dataset; page 589, first column final paragraph, describing HDDT as being based on distances between positive and negative classes over all of a feature’s partitions; page 591, first column, second paragraph, indicating the HDDT is used as a base learner on the unbalanced data; page 591, second column first paragraph, describing real-world credit card dataset as being highly unbalanced but including 0.15% of transactions as being fraudulent (i.e. positive); page 593, first column, final paragraph indicating that the fraud dataset is extremely unbalanced and exhibiting concept drift within the minority class, where the HDDT performs very well on the dataset; i.e. the dataset is unbalanced and positive (includes positive data), and the HDDT is trained on this unbalanced and positive dataset).
Pozzolo does not explicitly disclose that the data set is unlabeled data. However, Phua teaches that the data set is unlabeled data (e.g. page 5 second column, second paragraph, indicating that some research recommends use of unlabelled data in fraud detection use cases; page 8, first column, third and fourth paragraphs, describing unsupervised approaches with unabelled data, including in conjunction with Hellinger distance for comparing probability distributions and giving suspicions scores, and detecting statistical outliers using Hellinger distance).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo, Guo, and Phua in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams) and Guo (directed to facilitating automatic handling of incomplete data in a random forest model), to incorporate the teachings of Phua (directed to fraud detection research) to include the capability to utilize an unlabeled dataset. One of ordinary skill would have been motivated to perform such a modification in order to address criticisms associated with using labelled data to detect fraud as described in Phua (page 5, second column, second paragraph).
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Pozzolo in view of Guo, further in view of Phua, further in view of Patel et al. (US 20190213605 A1).
With respect to claim 10, Pozzolo in view of Guo, further in view of Phua teaches all of the limitations of claim 9 as previously discussed. Pozzolo and Phua do not explicitly disclose wherein the Hellinger decision tree is configured to consider random feature selection when initializing a tree node.
However, Patel teaches wherein the Hellinger decision tree is configured to consider random feature selection when initializing a tree node (e.g. paragraph 0124, building decision tree in random forest; each node splits on a feature selected from a random subset of the full feature set).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo, Guo, Phua, and Patel in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), Guo (directed to facilitating automatic handling of incomplete data in a random forest model), and Phua (directed to fraud detection research), to incorporate the teachings of Patel (directed to prediction of automotive warranty fraud) to include the capability to configure the decision tree to consider random feature selection when initializing nodes. One of ordinary skill would have been motivated to perform such a modification in order to implement the tree such that it can be trained quickly, while also being resistant to overfitting and providing a good estimate of generalization error as described in Patel (paragraph 0124).
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Pozzolo in view of Barrow et al. (US 20060167655 A1).
With respect to claim 4, Pozzolo teaches all of the limitations of claim 1 as previously discussed. Pozzolo does not explicitly disclose wherein the Hellinger decision tree is configured to use class prior to estimate counts of positives and negatives in each node. However, Barrow teaches wherein the Hellinger decision tree is configured to use class prior to estimate counts of positives and negatives in each node (e.g. paragraph 0105, every probability estimate of class membership for every possible combination of probability range bines at each node of the tree is the prior probability of class membership; throughout probability range bin combinations at each node of tree, count of total number of instances is set to 1 and the count of the number of positive outcome instances is set to the prior probability estimate).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo and Barrow in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), to incorporate the teachings of Barrow (directed to classification using probability estimate resampling) to include the capability to use, in the decision tree, class prior to estimate counts of positives and negatives in each node. One of ordinary skill would have been motivated to perform such a modification in order to ensure the model will output the prior probability for any new instances to be classified which happen to fall in probability range bin combinations with no observed instances in the training data, as described in Barrow (paragraph 0105).
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Pozzolo in view of Guo, further in view of Phua, further in view of Barrow, further in view of Miranda et al. (US 20180357299 A1).
With respect to claim 11, Pozzolo in view of Guo, further in view of Phua teaches all of the limitations of claim 9 as previously discussed. Pozzolo does not explicitly disclose wherein the Hellinger decision tree is configured to use a class prior. However, Barrow teaches wherein the Hellinger decision tree is configured to use a class prior (e.g. paragraph 0105, every probability estimate of class membership for every possible combination of probability range bines at each node of the tree is the prior probability of class membership; throughout probability range bin combinations at each node of tree, count of total number of instances is set to 1 and the count of the number of positive outcome instances is set to the prior probability estimate).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo, Guo, Phua, and Barrow in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), Guo (directed to facilitating automatic handling of incomplete data in a random forest model), and Phua (directed to fraud detection research), to incorporate the teachings of Barrow (directed to classification using probability estimate resampling) to include the capability to use, in the decision tree, a class prior. One of ordinary skill would have been motivated to perform such a modification in order to ensure the model will output the prior probability for any new instances to be classified which happen to fall in probability range bin combinations with no observed instances in the training data, as described in Barrow (paragraph 0105).
Pozzolo and Barrow do not explicitly disclose wherein the Hellinger decision tree is configured to use a size of a stratified bootstrap sample. However, Miranda teaches wherein the Hellinger decision tree is configured to use a size of a stratified bootstrap sample (e.g. paragraph 0068, performing statistical sampling; choice based stratified sampling of log entries; paragraphs 0087-0088, assignment of category identifiers to log entries based on groupings determined by classifier which may be a random forest classifier based on a decision tree; fitting decision trees on sub-samples of the dataset; sub-sample size may be the same as the original input sample size; samples drawn with replacement if bootstrap=True; i.e. the system uses a bootstrap sampling method of stratified samples, based on a size of the samples).
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention having the teachings of Pozzolo, Guo, Phua, Barrow, and Miranda in front of him to have modified the teachings of Pozzolo (directed to using Hellinger distance decision trees to avoid instances propagation in unbalanced and evolving data streams), Guo (directed to facilitating automatic handling of incomplete data in a random forest model), Phua (directed to fraud detection research), and Barrow (directed to classification using probability estimate resampling), to incorporate the teachings of Miranda (directed to assignment of category identifiers to log entries based on decision trees) to include the capability to use a size of a stratified bootstrap sample. One of ordinary skill would have been motivated to perform such a modification in order to achieve accuracy of learning algorithms available while also allowing processing of high dimensional space using a large number of training examples as described in Miranda (paragraph 0088).
It is noted that any citation to specific pages, columns, lines, or figures in the prior art references and any interpretation of the references should not be considered to be limiting in any way. “The use of patents as references is not limited to what the patentees describe as their own inventions or to the problems with which they are concerned. They are part of the literature of the art, relevant for all they contain,” In re Heck, 699 F.2d 1331, 1332-33, 216 USPQ 1038, 1039 (Fed. Cir. 1983) (quoting in re Lemelson, 397 F.2d 1006, 1009, 158 USPQ 275, 277 (GCPA 1968)). Further, a reference may be relied upon for all that it would have reasonably suggested to one having ordinary skill the art, including nonpreferred embodiments. Merck & Co, v. Biocraft Laboratories, 874 F.2d 804, 10 USPQ2d 1843 (Fed. Cir.), cert, denied, 493 U.S. 975 (1989). See also Upsher-Smith Labs. v. Pamlab, LLC, 412 F,3d 1319, 1323, 75 USPQ2d 1213, 1215 (Fed. Cir, 2005): Celeritas Technologies Ltd. v. Rockwell International Corp., 150 F.3d 1354, 1361, 47 USPQ2d 1516, 1522-23 (Fed. Cir. 1998).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant’s disclosure.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JEREMY L STANLEY whose telephone number is (469)295-9105. The examiner can normally be reached on Monday-Friday from 9:00 AM to 5:00 PM CST.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Abdullah Al Kawsar, can be reached at telephone number (571) 270-3169. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from Patent Center and the Private Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from Patent Center or Private PAIR. Status information for unpublished applications is available through Patent Center and Private PAIR for authorized users only. Should you have questions about access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
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) Form at https://www.uspto.gov/patents/uspto-automated- interview-request-air-form.
/JEREMY L STANLEY/
Primary Examiner, Art Unit 2127