# US 2021/0342537 A1 · Data processing device, data processing method, and data processing program

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0342537 A1
- Document ID: US20210342537A1
- Status: Published application
- Kind: A1
- Application number: 17/284,225
- Filed: 2019-09-19
- Published: 2021-11-04
- Earliest claimed: 2018-10-17
- Assignees: Nippon Telegraph and Telephone Corporation
- Applicants: Nippon Telegraph and Telephone Corporation
- Inventors: Satoshi Sunaga; Koji Hoshino; Kazuhiro Kikuma
- Web page: https://patentgrep.com/patent/US20210342537A1

## CPC

- G06F 40/279: Recognition of textual entities
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06F 16/35: Clustering; Classification
- G06F 40/211: Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
- G06F 40/216: using statistical methods
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A data processing apparatus (10) includes a word selection unit (143) configured to select words corresponding to a part-of-speech type targeted for selection from words disassembled from document data, and a generation unit (144) configured to find, for each of the words selected by the word selection unit (143), a value obtained by weighting the number of occurrences of each of the words or frequency of occurrences of each of the words according to the part-of-speech type of each of the words to generate a vector with a plurality of the values that are found as values of elements.

## Claims

### Claim 1

A data processing apparatus comprising:
a selection unit, including one or more processors, configured to select words corresponding to a part-of-speech type targeted for selection from words disassembled from document data; and
a generation unit, including one or more processors, configured to find, for each of the words selected by the selection unit, a value obtained by weighting the number of occurrences of each of the words or frequency of occurrences of each of the words according to the part-of-speech type of each of the words to generate a vector with a plurality values that are found as values of elements.

### Claim 2 (depends on 1)

The data processing apparatus according to claim 1, wherein the selection unit selects words corresponding to a plurality of part-of-speech types.

### Claim 3 (depends on 1)

The data processing apparatus according to claim 1, wherein a noun alone and a noun compound word that combines a plurality of nouns are taken as different part-of-speech types.

### Claim 4

A data processing apparatus comprising:
a learning unit, including one or more processors, configured to:
select first words corresponding to a part-of-speech type targeted for selection from words disassembled from training data,
find, for each of the first words that are selected, a first value obtained by weighting the number of occurrences of each of the first words or frequency of occurrences of each of the first words according to the part-of-speech type of each of the first words to generate a first vector with a plurality of first values that are found as values of elements, and
learn the first vector that is generated and information relating to a necessity of verification associated with the training data; and a prediction unit, including one or more processors, configured to:
select second words corresponding to the part-of-speech type targeted for selection from words disassembled from data targeted for prediction,
find, for each of the second words that are selected, a second value obtained by weighting the number of occurrences of each of the second words or frequency of occurrences of each of the second words according to the part-of-speech type of each of the second words to generate a second vector with a plurality of second values that are found as values of elements, and
predict a necessity of verification on the data targeted for prediction based on the second vector that is generated and a learning result from the learning unit.

### Claim 5

(canceled)

### Claim 6

A non-transitory computer readable medium storing one or more instructions causing a processor to execute:
selecting words corresponding to a part-of-speech type targeted for selection from words disassembled from document data, and
finding, for each of the words selected by the selecting step, a value obtained by weighting the number of occurrences of each of the words or frequency of occurrences of each of the words according to the part-of-speech type of each of the words to generate a vector with a plurality of values that are found as values of elements.

### Claim 7 (depends on 6)

The non-transitory computer readable medium according to claim 6, wherein the selection unit selects words corresponding to a plurality of part-of-speech types.

### Claim 8 (depends on 6)

The non-transitory computer readable medium according to claim 6, wherein a noun alone and a noun compound word that combines a plurality of nouns are taken as different part-of-speech types.
