US2020004790A1PendingUtilityA1

Method and system for extracting sentences

44
Assignee: UBERPLE CO LTDPriority: Sep 9, 2015Filed: Sep 12, 2019Published: Jan 2, 2020
Est. expirySep 9, 2035(~9.2 yrs left)· nominal 20-yr term from priority
G06F 16/313G06F 16/93G06F 40/211
44
PatentIndex Score
0
Cited by
0
References
0
Claims

Abstract

Methods and apparatus for extracting sentences are provided, one of methods comprises, receiving a document summarization request, which does not comprise a keyword for extracting important sentences from a document, from a user terminal, selecting a keyword, which reflects the preference of a user of the user terminal, based on a profile of the user, extracting important sentences from the document based on the selected keyword; and generating a summary of the document based on the extracted important sentences.

Claims

exact text as granted — not AI-modified
What is claimed is: 
     
         1 . A document summarization method comprising:
 receiving a document summarization request, which does not comprise a keyword for extracting important sentences from a document, from a user terminal;   selecting a keyword, which reflects the preference of a user of the user terminal, based on a profile of the user;   extracting important sentences from the document based on the selected keyword; and   generating a summary of the document based on the extracted important sentences.   
     
     
         2 . The document summarization method of  claim 1 ,
 wherein the selecting comprises selecting a first keyword and a second keyword which reflect the preference of the user,   wherein the extracting comprises:
 extracting k important sentences from the document based on the first keyword; and 
 extracting m important sentences from the document based on the second keyword, and 
   wherein the generating comprises generating a summary based on identical sentences among the k important sentences and the m important sentences.   
     
     
         3 . The document summarization method of  claim 1 ,
 wherein the selecting comprises selecting a first keyword and a second keyword which reflect the preference of the user,   wherein the extracting comprises:
 extracting k important sentences from the document based on the first keyword; 
 generating a first summary based on the k important sentences; 
 extracting m important sentences from the first summary based on the second keyword; and 
 generating a second summary based on the m important sentences. 
   
     
     
         4 . The document summarization method of  claim 1 ,
 wherein the selecting comprises selecting a first keyword and a second keyword which reflect the preference of the user, the first keyword being different from the second keyword,   wherein the extracting comprises:
 extracting a first important sentence from the document based on the first keyword; and 
 extracting a second important sentence from the document based on the second keyword, 
   wherein the generating comprises generating a summary which comprises the first important sentence and the second important sentence, the first important sentence being different from the second important sentence.   
     
     
         5 . A document summarization apparatus comprising:
 a network interface;   one or more processors;   a memory which is loaded with a computer program executed by the processors; and   a storage device which stores one or more files,   wherein the computer program comprises:
 an operation of receiving a document summarization request, which does not comprise a keyword for extracting important sentences from a document, from a user terminal; 
 an operation of selecting a keyword, which reflects the preference of a user of the user terminal, based on a profile of the user; 
 an operation of extracting important sentences from the document based on the selected keyword; and 
 an operation of generating a summary of the document based on the extracted important sentences. 
   
     
     
         6 . The document summarization apparatus of  claim 5 ,
 wherein the operation of the selecting comprises an operation of selecting a first keyword and a second keyword which reflect the preference of the user,   wherein the operation of the extracting comprises:
 an operation of extracting k important sentences from the document based on the first keyword; and 
 an operation of extracting m important sentences from the document based on the second keyword, and 
 wherein the operation of the generating comprises an operation of generating a summary based on identical sentences among the k important sentences and the m important sentences. 
   
     
     
         7 . The document summarization apparatus of  claim 5 ,
 wherein the operation of the selecting comprises an operation of selecting a first keyword and a second keyword which reflect the preference of the user,   wherein the operation of the extracting comprises:
 an operation of extracting k important sentences from the document based on the first keyword; 
 an operation of generating a first summary based on the k important sentences; 
 an operation of extracting m important sentences from the first summary based on the second keyword; and 
 an operation of generating a second summary based on the m important sentences. 
   
     
     
         8 . The document summarization apparatus of  claim 5 ,
 wherein the operation of the selecting comprises an operation of selecting a first keyword and a second keyword which reflect the preference of the user, the first keyword being different from the second keyword,   wherein the operation of the extracting comprises:
 an operation of extracting a first important sentence from the document based on the first keyword; and 
 an operation of extracting a second important sentence from the document based on the second keyword, 
   wherein the operation of the generating comprises generating a summary which comprises the first important sentence and the second important sentence, the first important sentence being different from the second important sentence.

Cited by (0)

No later patents cite this yet.

References (0)

No backward citations on record.