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 .
Response to Amendment
The amendment filed on 08/31/2026 has been entered. Claims 1, 6, 9, 12, 16, 18, & 20 are amended. Claim 9 is canceled. Claims 1-7 & 9-20 are pending in the application.
Claim Objections
Claims 18 is objected to because of the following informalities:
Claim 18 recites ‘classification ode’; however, it should recite - - classification code - -.
Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 10 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
The following claims lack antecedent basis:
Claim 10, line 1, the phrase “the comparison”;
Claim 10, line 3, the phrase “the centroid”.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-7 & 11-20 are rejected under 35 U.S.C. 103 as being unpatentable over Bothwell et al. (US 10109017 B2, published 10/23/2018), hereafter Bothwell, in view of Komissarchik et al. (US 8965877 B2, published 02/24/2015), hereafter Komissarchik, and further in view of Azari et al. (US 9361377 B1, published 06/07/2016), hereafter Azari.
Komissarchik was cited in the IDS submitted 10/20/2023.
Regarding independent claim 1, Bothwell teaches a system comprising:
a classification data store that contains electronic records associated with entity classifications ([Col. 11, Lines 8-20] discusses a database that stores information related to the industrial classifications of entities);
and the back-end application computer server, coupled to the classification data store, including: a computer processor; and a memory coupled to the computer processor and storing instructions that, when executed by the computer processor, cause the back-end application computer server to ([Col. 10, Lines 15-17] discusses a memory storing instructions and processor);
receive a name and an address associated with the name ([Col. 16, Lines 30-34] discusses obtaining a name for the entity and obtaining a web address);
receive third-party data associated with the name ([Fig. 4, 406] discusses obtaining third-party data associated with the name);
execute a supervised classifier of a trained hybrid machine learning (ML) model included in a classification tool and output a pre-configured number of top predicted internal classification codes for the received name ([Col. 14, Lines 42-58, Col. 19, Lines 17-19, & Col. 27, Lines 7-9, 46-48] discusses using a predictive model (rotating forest technique or K-nearest neighbors) to classify the name; this predictive model falls under supervised learning; the predictive model outputs a pre-configured number of classifications for the received name);
execute an additional predictive model ([Col. 23, Lines 43-53] discusses the classification may rely on more than one predictive model or classifier applied to the same data);
classify the name as an enterprise type using the output and the third-party data ([Col. 14, Lines 42-45] discusses classifying the name as an enterprise type based on the predictions made by the classification method; [Col. 17, Lines 29-48] discusses using the third-party data to output a predicted classification);
a communication port coupled to the back-end application computer server to facilitate an exchange of data with a remote device via a distributed communication network to support interactive user interface displays that include information about the classification ([Col. 27, Lines 46-61] discusses the web server delivering results to a user device encompassing information about the classification).
Bothwell does not explicitly teach to compare, via execution of an unsupervised labelling classifier of the trained hybrid ML model, the received name to a plurality of class centroids for each internal classification code, and output at least one additional predicted internal classification code as unsupervised classification output for the name, wherein each class centroid is generated from a plurality of a top pre-configured number of most common words for each internal classification code; combine, via the classification tool, the pre-configured number of top predicted internal classification codes output from the executed supervised classifier with the unsupervised classification output, and output one or more predicted classifications; classify the name as an enterprise type using the output one or more predicted classifications.
However, Komissarchik teaches a system for generating classification predictions wherein an unsupervised labelling classifier is executed ([Col. 5, Lines 10-20] discusses executing an unsupervised labelling classifier to generate a centroid of words for a class, using every node available; [Col. 5, Lines 45-58] discusses comparing received names to the constructed centroid for a classification code and outputting at least one additional predicted internal classification code); combining the supervised and unsupervised output ([Col. 5-6, Lines 48-4] discusses a trained, supervised matching procedure that produces a first list of matches; then discusses combining the results of both lists and outputting one or more of the best matches as predicted classifications); and classifying the name as an enterprise type based on the outputted one or more predicted classifications ([Col. 5-6, Lines 59-4] discusses using the predicted classifications to assign and classify the name as an enterprise type).
Because Bothwell teaches a classification data store, memory, processor, receiving a name, third-party information, and address, executing a supervised classifier to output a pre-configured number of classification codes, classifying the name as an enterprise type based on the prediction and third party data, and a communication port to facilitate the sending of results to a user interface; and Komissarchik teaches executing an unsupervised labelling classifier to generate a centroid of words for a class, comparing received names to the constructed centroid for a classification code and outputting at least one additional predicted internal classification code, combining the supervised and unsupervised output, and classifying the name as an enterprise type based on the outputted one or more predicted classifications, accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate executing an unsupervised labelling classifier to generate a centroid of words for a class, comparing received names to the constructed centroid for a classification code and outputting at least one additional predicted internal classification code, combining the supervised and unsupervised output, and classifying the name as an enterprise type based on the outputted one or more predicted classifications as taught by Komissarchik into Bothwell’s computer-implemented system, with a reasonable expectation of success, to teach a classification data store that contains electronic records associated with entity classifications; and the back-end application computer server, coupled to the classification data store, including a computer processor; and a memory coupled to the computer processor and storing instructions that, when executed by the computer processor, cause the back-end application computer server to receive a name and an address associated with the name; receive third-party data associated with the name; execute a supervised classifier of a trained hybrid machine learning (ML) model included in a classification tool and output a pre-configured number of top predicted internal classification codes for the received name; compare, via execution of an unsupervised labelling classifier of the trained hybrid ML model, the received name to a plurality of class centroids for each internal classification code, and output at least one additional predicted internal classification code as unsupervised classification output for the name; combine, via the classification tool, the pre-configured number of top predicted internal classification codes output from the executed supervised classifier with the unsupervised classification output, and output one or more predicted classifications; classify the name as an enterprise type using the output one or more predicted classifications and the third-party data; a communication port coupled to the back-end application computer server to facilitate an exchange of data with a remote device via a distributed communication network to support interactive user interface displays that include information about the classification. This combination would have been motivated by the desire to mine data presented on an Internet web page and properly attributes information relevant to the company of interest while rejecting extraneous information (Komissarchik [Col. 2, Lines 43-45]).
The combination of Bothwell and Komissarchik does not explicitly teach wherein each class centroid is generated from a plurality of a top pre-configured number of most common words for each internal classification code.
However, Azari teaches a system for classification wherein a category list is built from a pre-configured number of words ([Col. 16, Lines 10-19] discusses building a category or list, which represents a class centroid, using a pre-configured number of the highest scoring words).
Because the combination of Bothwell and Komissarchik teaches a classification data store, memory, processor, receiving a name, third-party information, and address, executing a supervised classifier to output a pre-configured number of classification codes, classifying the name as an enterprise type based on the prediction and third party data, a communication port to facilitate the sending of results to a user interface, executing an unsupervised labelling classifier to generate a centroid of words for a class, comparing received names to the constructed centroid for a classification code and outputting at least one additional predicted internal classification code, combining the supervised and unsupervised output, and classifying the name as an enterprise type based on the outputted one or more predicted classifications; and Azari teaches generating a centroid from a pre-configured number of top words, accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate generating a centroid from a pre-configured number of top words as taught by Azari into the combination of Bothwell and Komissarchik’s computer-implemented system, with a reasonable expectation of success, to teach a classification data store that contains electronic records associated with entity classifications; and the back-end application computer server, coupled to the classification data store, including a computer processor; and a memory coupled to the computer processor and storing instructions that, when executed by the computer processor, cause the back-end application computer server to receive a name and an address associated with the name; receive third-party data associated with the name; execute a supervised classifier of a trained hybrid machine learning (ML) model included in a classification tool and output a pre-configured number of top predicted internal classification codes for the received name; compare, via execution of an unsupervised labelling classifier of the trained hybrid ML model, the received name to a plurality of class centroids for each internal classification code, and output at least one additional predicted internal classification code as unsupervised classification output for the name, wherein each class centroid is generated from a plurality of a top pre-configured number of most common words for each internal classification code; combine, via the classification tool, the pre-configured number of top predicted internal classification codes output from the executed supervised classifier with the unsupervised classification output, and output one or more predicted classifications; classify the name as an enterprise type using the output one or more predicted classifications and the third-party data; a communication port coupled to the back-end application computer server to facilitate an exchange of data with a remote device via a distributed communication network to support interactive user interface displays that include information about the classification. This combination would have been motivated by the desire to tune the pre-configured number of words until a desired accuracy for the category is met (Azari [Col. 16]) and thus, making the predicted classification more accurate.
Regarding dependent claim 2, the combination of Bothwell, Komissarchik, and Azari teaches the claimed invention as claimed in claim 1, including wherein the third-party data is received using the name and the address (Bothwell [Col. 13, Lines 58-66] discusses the system searching for third-party data using information related to the entity and thus, the third-party data is “received” using the name, address, and other information).
Regarding dependent claim 3, the combination of Bothwell, Komissarchik, and Azari teaches the claimed invention as claimed in claim 2, including wherein the third-party data includes a plurality of Standard Industry Classification (SIC) codes and a plurality of North American Industry Classification System (NAICS) codes (Bothwell [Col 12-13, Lines 66-67, 1-9] discusses the third-party data also including codes of SIC or NAICS).
Regarding dependent claim 4, the combination of Bothwell, Komissarchik, and Azari teaches the claimed invention as claimed in claim 3, including to determine the name maps to at least one of the SIC code and the NAICS code; identify an internal classification code mapped to the at least one SIC code and NAICS code mapped to the name; and assign the internal classification code to the name (Bothwell [Col 14, Lines 42-50] discusses the predictive model determining a classification mapped to either SIC, NAICS, or ICB; and assigning the classification code associated to the name).
Regarding dependent claim 5, the combination of Bothwell, Komissarchik, and Azari teaches the claimed invention as claimed in claim 3, including to determine the name does not map to at least one of the SIC code and the NAICS code; and execute the trained hybrid ML model using the name as input (Bothwell [Col. 1, Lines 57-63] discusses that when a company does not have an industrial classification provided by a third party, the burden falls onto itself or the agent; Bothwell [Col. 14, Lines 54-58] discusses calculating an estimation error and confidence score in how well the initial mapping of SIC and NAICS codes was and thus, when a name does not initially map to an SIC or NAICS code, the ML model will be executed to assign a classification code).
Regarding dependent claim 6, the combination of Bothwell, Komissarchik, and Azari teaches the claimed invention as claimed in claim 5, including to:
execute a K-Nearest Neighbors (KNN) classifier using the name, wherein execution of the KNN classifier outputs at least one internal classification code as KNN classification output for the name (Bothwell [Col 8, Lines 24-29] discusses examples of predictive models including K-Nearest Neighbor models. In this specific case, a rotating forest technique is used to output at least one internal classification code to the name; however, it is noted that a K-Nearest Neighbor model would yield predictable and similar results in this system);
select at least one KNN classifier output and at least one unsupervised labelling classifier output (Komissarchik [Col. 5-6, Lines 59-4] discusses selecting one KNN classifier output and one unsupervised labelling classifier output and then comparing the two; if there are matches, the pair is assigned a weight and if that weight is above a threshold, that classification is assigned to an entity);
and determine a confidence score for each selected KNN classifier output and unsupervised labelling classifier output (Bothwell [Col. 14, Lines 54-58] discusses generating a confidence score that represents how well a particular classification describes an entity).
Regarding dependent claim 7, the combination of Bothwell, Komissarchik, and Azari teaches the claimed invention as claimed in claim 6, including wherein execution of the KNN classifier outputs three internal classification codes for the name (Bothwell [Col. 30, Lines 30-36] discusses outputting 2 or more classification codes to the user interface).
Regarding dependent claim 11, the combination of Bothwell, Komissarchik, and Azari teaches the claimed invention as claimed in claim 6, including wherein the confidence score indicates the confidence the trained hybrid ML model has that the output internal classification code is accurate (Bothwell [Col. 14, Lines 54-58] discusses generating a confidence score that represents how well a particular classification describes an entity).
Regarding claims 12-16, claims 12-16 are method claims that are substantially the same as the system of claims 1 & 3-6. Therefore, claims 12-16 are rejected for the same reasons as claims 1 & 3-6, respectively.
Regarding dependent claim 17, the combination of Bothwell, Komissarchik, and Azari teaches the claimed invention as claimed in claim 12, including pre-filling an entry field on a client system user interface with the classified enterprise type, wherein pre-filling is based on a confidence value for the classified enterprise type (Bothwell [Fig. 8] discusses the user interface pre-filling the classified enterprise type and also displays the confidence score associated with that enterprise classification code).
Regarding claims 18-20, claims 18-20 are non-transitory computer-readable storage medium claims that are substantially the same as the system of claims 1, 4, & 6. Therefore, claims 18-20 are rejected for the same reasons as claims 1, 4, & 6, respectively.
Claims 9-10 are rejected under 35 U.S.C. 103 as being unpatentable over Bothwell, in view of Komissarchik, in view of Azari, as applied in claim 1, and further in view of Hore et al. (US 11625535 B1, published 04/11/2023), hereafter Hore.
Regarding dependent claim 9, the combination of Bothwell, Komissarchik, and Azari teaches the claimed invention as claimed in claim 1, including generating class centroids (Komissarchik [Col. 5, Lines 10-20] discusses executing an unsupervised labelling classifier to generate a centroid of words for a class).
The combination of Bothwell, Komissarchik, and Azari does not explicitly teach to wherein each class centroid includes a plurality of key words representing the enterprise type.
However, in the same field of endeavor, Hore teaches an entity classification system wherein each class centroid includes a plurality of key words representing the enterprise type (Hore [Col. 4, Lines 9-15] discusses subcategories are comprised of a collection of related key words and the category prototypes that represent the different enterprises are composed of these subcategories).
Because the combination of Bothwell, Komissarchik, and Azari teaches generating class centroids; and Hore teaches an entity classification system wherein each class centroid includes a plurality of key words representing the enterprise type, accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate each class centroid including a plurality of key words representing the enterprise type as taught by Hore into the combination of Bothwell, Komissarchik, and Azari’s system, with a reasonable expectation of success, to teach a system wherein each class centroid includes a plurality of key words representing the enterprise type. This combination would have been motivated by the desire to produce a set of category profiles represented by word embeddings and accurate context (Hore [Col. 1]).
Regarding dependent claim 10, the combination of Bothwell, Komissarchik, Azari, and Hore teaches the claimed invention as claimed in claim 9, including to convert each key word of the centroid to a centroid vector; convert the name to a name vector; determine a proximity of the name vector to each centroid vector; and output at least one internal classification code for the name as the unsupervised classification output, wherein the output internal classification code is mapped to the smallest determined proximity (Hore [Col. 7, Lines 52-67] discusses generating vectors derived from the subcategories represented by word embeddings; producing an input vector from input 601, wherein the input is the entity name and data; the name vector’s proximity is measured to each centroid using a cosine similarity technique to find the nearest match; and the nearest vector is used to output a category and subcategory, e.g. a classification code, associated with the input).
Response to Arguments
Applicant’s amendments and remarks filed 08/31/2026, on pages 9-15, traversing the 35 U.S.C. 101 rejections set forth in the Office Action dated 06/02/2026, are persuasive and hereby withdrawn.
Applicant’s arguments filed 08/31/2026, on pages 15-16, with respect to claims 1-7 & 9-20 under 35 U.S.C. 103 have been considered but are moot because of the new ground of rejection (see rejection above).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Wood et al. ("Automated Industry Classification with Deep Learning", 12th IEEE International Conference on Semantic Computing, IEEE) (Year: 2018) ([Abstract] We examine the capacity of our model to predict six-digit NAICS codes, as well as the ability of our model architecture to adapt to other industry segmentation schemas. Additionally, we investigate the ability of our model to generalize despite the presence of noise in the labels in our training set. Finally, we explore the possibility of increasing predictive precision by thresholding based on the confidence scores that our model outputs along with its predictions. We find that our approach yields six-digit NAICS code predictions that surpass the precision of gold-standard databases).
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RILEY S ACOSTA whose telephone number is (571)272-8714. The examiner can normally be reached Monday-Thursday 6am-4pm.
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, Jennifer N Welch can be reached at (571)272-7212. 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.
/RILEY S ACOSTA/Examiner, Art Unit 2143
/JENNIFER N WELCH/Supervisory Patent Examiner, Art Unit 2143
Prosecution Insights