Information analysis system, information analysis method, and information analysis program

• US 9,940,319 B2
• Filed: 05/25/2015
• Issued: 04/10/2018
• Est. Priority Date: 05/29/2014
• Status: Active Grant

• Current Assignee: Nippon Telegraph and Telephone Corporation
• Original Assignee: Nippon Telegraph and Telephone Corporation
• Inventors: Orihara, Shingo, Kumagai, Atsutoshi, Abe, Tetsuya
• Primary Examiner(s): VO, HUYEN X
• Application Number:
  US15/313,749
• Publication Number:
  US 20170185578A1
• Time in Patent Office: 1,051 Days
• Field of Search: 704 1- 10, 704250, 704251, 704255, 704257, 704270
• CPC Class Codes: G06F 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 40/10 Text processing natural language analysis G06F40/20; semantic analysis G06F40/30; processing or translation of natural language G06F40/40
• CPC Class Codes: G06F 40/237 Lexical tools
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

An information analysis system includes a remark analysis unit, a thread analysis unit, and a storing unit. The remark analysis unit analyzes importance of a remark included in a thread serving as a group of remarks posted on a network, based on remark data serving as data relating to the remark, for each of the remarks. The thread analysis unit analyzes which of a plurality of preset categories the thread belongs to, based on thread data serving as data relating to the thread. The storing unit stores the remark, the importance of the remark, and a category of the thread including the remark in association with each other for each of the remarks, in a predetermined storage unit.

Claims

1. An information analysis system comprising:

circuitry configured to

select a plurality of threads from a bulletin board service or a social networking service,

receive a classification of the plurality of threads as important and unimportant,

train a machine learning engine with the threads classified as important and the threads classified as unimportant,

analyze, using the trained machine learning engine, importance of a remark included in a thread serving as a group of remarks posted on a network, based on remark data serving as data relating to the remark, for each of the remarks,

analyze, using the trained machine learning engine, which of a plurality of preset categories the thread belongs to, based on thread data serving as data relating to the thread,

analyze, using the trained machine learning engine, importance of the thread based on the thread data,

calculate total importance of the remark based on the importance of the remark and the importance of the thread,

store the remark, the total importance of the remark, and a category of the thread including the remark in association with each other for each of the remarks, in a predetermined memory, and

retrieve, from the predetermined memory in response to an input, one or more of useful remarks to be presented to a user.

2. The information analysis system according to claim 1, wherein the circuitry is configured to

analyze a remark category indicating which of a plurality of preset categories contents of the remark belong to, for each of the remarks, based on the remark data, and

store the remark category of each of the remarks in association with the remark.

3. An information analysis method executed by an information analysis system including circuitry, the information analysis method comprising:

selecting, using the circuitry, a plurality of threads from a bulletin board service or a social networking service;

receiving, using the circuitry, a classification of the plurality of threads as important and unimportant;

training, using the circuitry, a machine learning engine with the threads classified as important and the threads classified as unimportant;

analyzing, using the circuitry and the trained machine learning engine, importance of a remark included in a thread serving as a group of remarks posted on a network, based on remark data serving as data relating to the remark, for each of the remarks;

analyzing, using the circuitry and the trained machine learning engine, which of a plurality of preset categories the thread belongs to, based on thread data serving as data relating to the thread;

analyzing, using the circuitry and the trained machine learning engine, importance of the thread based on the thread data;

calculating, using the circuitry, total importance of the remark based on the importance of the remark and the importance of the thread;

storing, using the circuitry, the total importance of the remark, and a category of the thread including the remark in association with each other for each of the remarks, in a predetermined memory; and

retrieving, using the circuitry, from the predetermined memory in response to an input, one or more of useful remarks to be presented to a user.

4. A non-transitory computer-readable recording medium having stored an information analysis program causing a computer to execute a process comprising:

selecting a plurality of threads from a bulletin board service or a social networking service;

receiving a classification of the plurality of threads as important and unimportant;

training a machine learning engine with the threads classified as important and the threads classified as unimportant;

analyzing, using the trained machine learning engine, importance of a remark included in a thread serving as a group of remarks posted on a network, based on remark data serving as data relating to the remark, for each of the remarks;

analyzing, using the trained machine learning engine, which of a plurality of preset categories the thread belongs to, based on thread data serving as data relating to the thread;

analyzing, using the trained machine learning engine, importance of the thread based on the thread data;

calculating total importance of the remark based on the importance of the remark and the importance of the thread;

storing the importance of the remark, and a category of the thread including the remark with each other for each of the remarks, in a predetermined memory; and

retrieving, from the predetermined memory in response to an input, one or more of useful remarks to be presented to a user.