# US 12,118,308 B2 · Document classification device and trained model > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,118,308 B2 - Document ID: US12118308B2 - Status: Granted - Kind: B2 - Application number: 17/041,299 - Filed: 2019-05-29 - Published: 2021-01-28 - Granted: 2024-10-15 - Earliest claimed: 2018-07-24 - Assignees: Ntt Docomo, Inc. - Applicants: Ntt Docomo, Inc. - Inventors: Taishi Ikeda - Web page: https://patentgrep.com/patent/US12118308B2 ## CPC - G06F 40/284: Lexical analysis, e.g. tokenisation or collocates - G06F 16/285: {Clustering or classification} - G06F 16/93: Document management systems - G06F 40/289: Phrasal analysis, e.g. finite state techniques or chunking - G06F 40/30: Semantic analysis - G06N 3/0499: Feedforward networks - G06N 3/084: Backpropagation, e.g. using gradient descent - G06N 3/09: Supervised learning - G06V 30/40: Document-oriented image-based pattern recognition - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract A document classification device is a device that generates a document classification model which outputs identification information for identifying a result of classification on the basis of an input document by machine learning, and includes: an acquisition unit configured to acquire learning data including a document and the identification information correlated with the document; a feature extracting unit configured to extract words included in the document and character information which is a character string including one character of characters constituting the words or a plurality of characters consecutive in the words and which is one or more pieces of information capable of being extracted from the words as features; and a model generating unit configured to perform machine learning on the basis of the feature extracted from the document and the identification information correlated with the document and to generate the document classification model. ## Claims ### Claim 1 A document classification device that generates a document classification model which is a model for classifying a document and which outputs identification information for identifying a result of classification on the basis of an input document by machine learning, the document classification device comprising circuitry configured to: acquire learning data including a document and the identification information correlated with the document; extract words included in the document and character information which is a character string including one character of characters constituting the words or a plurality of characters consecutive in the words and which is one or more pieces of information capable of being extracted from the words as features; and perform machine learning on the basis of the feature extracted from the document and the identification information correlated with the document and to generate the document classification model, wherein the circuitry is configured to extract the character information from a word satisfying a predetermined condition out of the words included in the document. ### Claim 2 (depends on 1) The document classification device according to claim 1, wherein the circuitry is configured to exclude a word corresponding to a second part of speech out of the words included in the document from the words from which the character information is extracted. ### Claim 3 (depends on 2) The document classification device according to claim 2, wherein the circuitry is configured to extract the character information which is included in a predetermined frequency or less in the learning data acquired by the circuitry as features. ### Claim 4 (depends on 2) The document classification device according to claim 2, wherein the circuitry is configured to remove repeated character information from the character information extracted from the words included in the document. ### Claim 5 (depends on 1) The document classification device according to claim 1, wherein the circuitry is configured to extract the character information which is included in a predetermined frequency or less in the learning data acquired by the circuitry as features. ### Claim 6 (depends on 5) The document classification device according to claim 5, wherein the circuitry is configured to remove repeated character information from the character information extracted from the words included in the document. ### Claim 7 (depends on 1) The document classification device according to claim 1, wherein the circuitry is configured to remove repeated character information from the character information extracted from the words included in the document. ### Claim 8 (depends on 1) The document classification device according to claim 1, wherein the circuitry is configured to extract the character information which is included in a predetermined frequency or less in the learning data acquired by the circuitry as features. ### Claim 9 (depends on 1) The document classification device according to claim 1, wherein the circuitry is configured to remove repeated character information from the character information extracted from the words included in the document. ### Claim 10 A method, implemented by circuitry of a document classification device that generates a document classification model which is a model for classifying a document and which outputs identification information for identifying a result of classification on the basis of an input document by machine learning, the method comprising: acquiring learning data including a document and the identification information correlated with the document; extracting words included in the document and character information which is a character string including one character of characters constituting the words or a plurality of characters consecutive in the words and which is one or more pieces of information capable of being extracted from the words as features; and performing machine learning on the basis of the feature extracted from the document and the identification information correlated with the document and to generate the document classification model, wherein the method includes extracting the character information from a word satisfying a predetermined condition out of the words included in the document, wherein the circuitry extracts the character information from a word corresponding to a first part of speech out of the words included in the document.