Information extracting device, information extracting method, and information extracting program

Patent metadata

Field	Value
Publication number	US-11995115-B2
Application number	US-201917618820-A
Country	US
Kind code	B2
Filing date	Jun 14, 2019
Priority date	Jun 14, 2019
Publication date	May 28, 2024
Grant date	May 28, 2024

Abstract

An information extracting device includes an acquiring unit that acquires, with regard to each information source, a data group made up of data including a content relating to an object, a location where the content was recorded, and a date and time at which the content was recorded, a generating unit that generates a word vector of which the date and time, the location, content using a weight based on the date and time, and a type of the content, are each components, for each of the data of the data group, a distance calculating unit that calculates a distance among the word vectors, a classifying unit that classifies each of data of the data group on the basis of the distance among the word vectors, and an extracting unit that calculates a reliability with regard to the classification, and extracts data from the data group on the basis of the reliability.

First claim

The invention claimed is:

1. An information extracting device comprising circuitry configured to execute operations comprising:
acquiring, using a device as an information source, a data group, wherein the device comprises a sensor capturing data of an object, and the data group comprises:
the data including a content describing the object,
a location where the content was recorded, and
a date and time at which the content was recorded;

generating a word vector of each of the data of the data group,
wherein the word vector is based at least on a textual expression of:
the date and time,
the location,
the content according to a first weight, and
a type of the content, and

the first weight is based on a difference between the date and time and a current date and time;

determining a distance among a plurality of word vectors of a plurality of data in the data group;
classifying each data of the data group on the basis of the distance among the plurality of word vectors to generate one or more clusters of the data according to classification;
determining, based at least on a frequency of appearance of the classified each data of the data in the classification and a second weight associated with each information source, reliability data of the classification;
extracting data from the data group on the basis of the reliability data:, and
transmitting the extracted data to an application configured to output the extracted data including extracted content describing the object with improved reliability according to the reliability data of the extracted data.

2. The information extracting device according to claim 1 , wherein the word vector is based on a product of a frequency of appearance of the content and the first weight.

3. The information extracting device according to claim 1 , the circuitry further configured to execute operations comprising:
calculating reliability data of each of sub-classifications that are classifications by type of content of the data included in the classification, for each classification; and
extracting data from the data group on the basis of each reliability of the sub-classifications.

4. The information extracting device according to claim 3 , the circuitry further configured to execute operations comprising:
calculating the reliability data of each of the sub-classifications on the basis of:
the frequency of appearance of the content in the sub-classification, and
the average of second weights of a plurality of information sources in the sub-classification for each of the classifications;

extracting a label indicating the content, wherein the reliability data the content is equal to or higher than a predetermined threshold value; or higher; and
extracting and outputting data of the data group including the content of the label, along with the label.

5. A computer-implemented method for extracting information, the method comprising:
acquiring, using a device as an information source, a data group, wherein the device comprises a sensor capturing data of an object, and the data group comprises:
the data including a content describing the object, a location where the content was recorded, and a date and time at which the content was recorded;

generating a word vector of the data of the data group,
wherein the word vector is based at least on a textual expression of:
the date and time,
the location,
the content according to a first weight, and
a type of the content, and

the first weight is based on a difference between the date and time and a current date and time of performing the generating a word vector operation;

determining a distance among a plurality of word vectors of a plurality of data in the data group;
classifying each data of the data group on the basis of the distance among the plurality of word vectors to generate one or more clusters of the data according to classification;
determining, based at least on a frequency of appearance of the classified each data of the data in the classification and a second weight of each information source, reliability data of the classification;
extracting data from the data group on the basis of the reliability data; and
transmitting the extracted data to an application configured to output the extracted data including extracted content describing the object with improved reliability according to the reliability data of the extracted data.

6. A computer-readable non-transitory recording medium storing computer-executable program instructions that when executed by a processor cause a computer system to execute a operations comprising:
acquiring, using a device as information source, a data group, wherein the device comprises a sensor capturing data of an object, and the data group comprises:
the data including a content describing the object, a location where the content was recorded, and a date and time at which the content was recorded;

generating a word vector of the data of the data group,
wherein the word vector is based at least on a textual expression of:
the date and time,
the location,
the content according to a first weight, and
a type of the content, and

the first weight is based on a difference between the date and time and a current date and time of performing the generating a word vector operation;
determining a distance among a plurality of word vectors of a plurality of data in the data group;
classifying each data of the data group on the basis of the distance among the plurality of word vectors to generate one or more clusters of the data according to classification;
determining, based at least on a frequency of appearance of the classified each data of the data in the classification and a second weight of each information source, reliability data of the classification;
extracting data from the data group on the basis of the reliability data; and
transmitting the extracted data to an application configured to output the extracted data including extracted content describing the object with improved reliability according to the reliability data of the extracted data.

7. The information extracting device according to claim 1 , wherein the content of the data of the object indicates a condition of a road, and the condition including one of flat or stepped.

8. The information extracting device according to claim 1 , wherein the classifying is based on a hierarchical clustering without a predetermined number of clusters.

9. The information extracting device according to claim 1 , the circuitry further configured to execute operations comprising:
outputting, based on the reliability data, the extracted data of the data group including the content of the label.

10. The computer-implemented method according to claim 5 , wherein the word vector is based on a product of a frequency of appearance of the content and the weight based on the date and time.

11. The computer-implemented method according to claim 5 , the method further comprising:
calculating the reliability data of each of sub-classifications that are classifications by type of content of the data included in the classification, for each classification; and
extracting data from the data group on the basis of each reliability of the sub-classifications.

12. The computer-implemented method according to claim 5 , wherein the content of the data of the object indicates a condition of a road, and the condition including one of flat or stepped.

13. The computer-implemented method according to claim 5 , wherein the classifying is based on a hierarchical clustering without a predetermined number of clusters.

14. The c

Assignees

• Nippon Telegraph & Telephone

Inventors

• Abe Naoto
• Seshimo Hitoshi
• Konishi Hiroshi

Classifications

• G06F16/35 Primary

  Clustering; Classification · CPC title

• G06F16/387

  using geographical or spatial information, e.g. location · CPC title

• G06F16/285 Primary

  Clustering or classification · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/2237

  Vectors, bitmaps or matrices · CPC title