Related data generating apparatus, related data generating method, and program

Patent metadata

Field	Value
Publication number	US-10157348-B2
Application number	US-201514637773-A
Country	US
Kind code	B2
Filing date	Mar 4, 2015
Priority date	Mar 7, 2014
Publication date	Dec 18, 2018
Grant date	Dec 18, 2018

Abstract

Provided are a co-occurrence word data generating unit configured to generate co-occurrence word data including a co-occurrence word that is a vocabulary used along with a predetermined keyword in posted data in all periods and an appearance frequency of the co-occurrence word, from among pieces of posted data posed in a plurality of different periods; and a related data generating unit configured to generate related data including the co-occurrence word as a usual related word when a temporal variation in the appearance frequency of the co-occurrence word is smaller than a first threshold value and the appearance frequency is higher than a second threshold value.

First claim

What is claimed is:

1. A related data generating apparatus, comprising:
a memory;
a network interface that is communicatively coupled to a user terminal and social network service; and
a processor communicatively coupled to the memory and the network interface;
wherein the processor is configured to execute instructions stored in the memory, that when executed by the processor cause the processor to:
receive, using the network interface, a request from the user terminal that includes a predetermined keyword,
retrieve, using the network interface, a vocabulary of word data that is used along with the predetermined keyword from the social network, wherein the vocabulary of word data is included in posted data from a plurality of different time periods,
generate co-occurrence word data and an appearance frequency of the co-occurrence word from the vocabulary of word data retrieved from the social network,
wherein the processor generates the co-occurrence word data including the co-occurrence word which is a second vocabulary used along with at least two different usual related words and an appearance frequency of the co-occurrence word in the posted data posted during a predetermined period, and stores, in the memory, the co-occurrence word in the related data as a related word similar name when a similarity between the co-occurrence word and the predetermined keyword is larger than a fifth threshold value,

calculates a score having a larger value as a similarity between the co-occurrence word and the predetermined keyword increases and stores in the memory, the score as a score of the usual related word similar name corresponding to the co-occurrence word,
determines whether a frequency of the co-occurrence word is high by:
determining distribution of frequencies of the keyword over the plurality of different periods,
calculating an average of the distribution of frequencies of the keyword over the plurality of different periods, and
determining that the frequency of the co-occurrence word is high when the average of the distribution of frequencies exceeds a frequency threshold;
on a condition that the frequency of the co-occurrence word is high, generates related data, the related data including the co-occurrence word as a usual related word, and
transmits, using the network interface, the related data to the user terminal.

2. The related data generating apparatus according to claim 1 , wherein the processor further:
determines whether a temporal variation in the co-occurrence word is small by:
calculating a ratio of the distribution of frequencies of the keyword over the plurality of different periods and the average of the distribution of frequencies of the keyword over the plurality of different periods, and
determines that the temporal variation in the co-occurrence word is small when the ratio is less than a temporal threshold; and

wherein the processor generates that related data on the condition that the frequency of the co-occurrence word is high and the temporal variation in the co-occurrence word is small.

3. The related data generating apparatus according to claim 1 , wherein the processor further:
calculates a score having a larger value as the appearance frequency of the co-occurrence word increases, and
stores, in the memory, the score as a score of the usual related word corresponding to the co-occurrence word.

4. The related data generating apparatus according to claim 3 , wherein the memory stores the co-occurrence word of which a score is greater than the score of the usual related word stored in the related data in the related data instead of the usual related word.

5. The related data generating apparatus according to claim 3 , wherein the processor further: calculates a score of the usual related word based on an appearance frequency of the usual related word in a particular period when the usual related word identical to the co-occurrence word is registered in the related data.

6. The related data generating apparatus according to claim 1 , further comprising a display that is communicatively coupled to the processor;
wherein the processor further: causes the display to display the usual related word along with a target indicated by the predetermined keyword.

7. The related data generating apparatus according to claim 1 , wherein the processor further:
calculates a degree at which the predetermined keyword or the related word similar name to which the predetermined keyword is mapped is identical to a vocabulary included in the posted data posted during the predetermined period as a first coincidence degree,
calculates a degree at which the related word similar name to which the predetermined keyword is mapped is identical to a vocabulary included in the posted data posted during the predetermined period as a second coincidence degree,
calculates a value obtained by summing up the first coincidence degree and the second coincidence degree calculated for all pieces of posted data within the predetermined period, as a trend level of a target indicated by the keyword, and
generates a trend level data in which the trend level is stored to be mapped to the predetermined keyword.

8. The related data generating apparatus according to claim 7 , wherein the processor further:
sets a predetermined value as the first coincidence degree when the predetermined keyword is identical to a vocabulary included in the posted data,
determines whether the related word similar name mapped to the keyword is identical to the vocabulary included in the posted data when the predetermined keyword is not identical to the vocabulary included in the posted data, and
sets the first coincidence degree, which is a value smaller than the predetermined value, by using a score of the related word similar name identical to the vocabulary when the related word similar name is identical to the vocabulary included in the posted data.

9. The related data generating apparatus according to claim 7 , further comprising a display that is communicatively coupled to the processor,
wherein the processor causes the display to display information indicating the trend level.

10. A related data generating and retrieving method comprising:
receiving from a user terminal a request that includes a predetermined keyword;
retrieving a vocabulary of word data that is used along with the predetermined keyword from a social network service, wherein the vocabulary of word data is included in posted data from a plurality of different time periods;
generating co-occurrence word data and an appearance frequency of the co-occurrence word, from the vocabulary of word data retrieved from the social network service;
wherein the generating the co-occurrence word data includes the co-occurrence word which is a second vocabulary used along with at least two different usual related words and an appearance frequency of the co-occurrence word in the posted data posted during a predetermined period, and stores, in the memory, the co-occurrence word in the related data as a related word similar name when a similarity between the co-occurrence word and the predetermined keyword is larger than a fifth threshold value,

calculating a score having a larger value as a similarity between the co-occurrence word and the predetermined keyword increases and stores in the memory, the score as a score of the usual related word similar name corresponding to the co-occurrence word,
determining whether a frequency of the co-occurrence word is high by:
determining a ratio of distribution of frequencies of the keyword over the plurality of different periods,
calculating an average of the distribution of frequencies of the keyword over the p

Assignees

• Clarion Co Ltd

Inventors

• Ishiguro Masao
• Hiroi Kazushige
• Hayashi Akio
• Sawajiri Haruhiko
• Horibe Yasuki

Classifications

• G06N99/005 Primary

  Physics · mapped topic

• G06F17/30861

  Physics · mapped topic

• G06F17/30663

  Physics · mapped topic

• G06F16/3334 Primary

  Selection or weighting of terms from queries, including natural language queries · CPC title

• G06N20/00

  Machine learning · CPC title