发明名称 SEQUENTIAL IMPORTANT WORD EXTRACTION DEVICE, SEQUENTIAL IMPORTANT WORD EXTRACTION METHOD AND PROGRAM
摘要 <P>PROBLEM TO BE SOLVED: To provide a technology, in discriminating an important word from a word set, capable of preventing an increase in a storage capacity and updating TFIDF upon arrival of a packet. <P>SOLUTION: A HTTP data assembly section 22 links fragmented HTTP data stored in a packet received from a packet reception section 21 through a terminal 10 and restores the HTTP data to an original state thereof. A keyword extraction section 23 extracts a keyword by applying a morphological analysis to the original HTTP data. A calculation section 24 calculates a level of importance by acquiring a parameter for each received word necessary for a calculation from a keyword parameter DB25. The level of importance is expressed in a form of a recurrence formula which requires only a last value for the calculation and thereby preventing an increase in a storage capacity and making a real time calculation possible. An important word transmission section 26 packetizes the calculated levels of importance or words having high levels of importance and transmits the packetized data to a service device 40. <P>COPYRIGHT: (C)2012,JPO&INPIT
申请公布号 JP2011238081(A) 申请公布日期 2011.11.24
申请号 JP20100109858 申请日期 2010.05.12
申请人 NIPPON TELEGR & TELEPH CORP <NTT> 发明人 KONDO SATORU;SAKURADA REIKO;MIYAGI YASUTOSHI;MORIYA TAKAAKI
分类号 G06F17/30 主分类号 G06F17/30
代理机构 代理人
主权项
地址