发明名称 |
INFORMATION EXTRACTING SERVER, INFORMATION EXTRACTING CLIENT, INFORMATION EXTRACTING METHOD, AND INFORMATION EXTRACTING PROGRAM |
摘要 |
According to one embodiment, an information extracting method includes: collecting a text in which a keyword of interest appears, the keyword of interest, and a time of creation of the text; extracting a keyword included in the text except for the keyword of interest and configured to extract the time of creation; extracting the keyword having a time score obtained on the basis of an appearance frequency of the keyword in a time interval exceeding a first threshold value and a local score obtained from the appearance frequency of the keyword in a predetermined local area exceeding a second threshold value as a local hot word, and also extract the extracted time interval of the extracted keyword and the keyword of interest corresponding to the keyword; and storing the extracted local hot word, the time interval, and the keyword of interest. |
申请公布号 |
US2014188883(A1) |
申请公布日期 |
2014.07.03 |
申请号 |
US201314138759 |
申请日期 |
2013.12.23 |
申请人 |
Kabushiki Kaisha Toshiba |
发明人 |
SASAKI Kenta;NAGANO Shinichi;UENO Koji;CHO Kenta |
分类号 |
G06F17/30 |
主分类号 |
G06F17/30 |
代理机构 |
|
代理人 |
|
主权项 |
1. An information extracting server comprising:
a collecting unit configured to collect a text in which a keyword of interest indicating information of interest that a user is interested in appears, the keyword of interest, and a time of creation of the text; a keyword extracting unit configured to extract a keyword included in the text except for the keyword of interest, and the time of creation; a local hot word extracting unit configured to extract the keyword having a time score obtained on the basis of an appearance frequency of the keyword in a time interval, which is a period T1 backward from the time of creation, exceeding a first threshold value and a local score on the basis of the appearance frequency of the keyword in a local area indicating a range of a specific topic exceeding a second threshold value as a local hot word, and also extract the extracted time interval of the extracted keyword and the keyword of interest corresponding to the keyword; and a local hot word storing unit configured to store the extracted local hot word, the time interval, and the keyword of interest. |
地址 |
Minato-ku JP |