To reduce a processing time spent for extraction a sequence pattern by efficiently counting an appearance number of the sequence pattern.
A template generation part 120 defines the sequence pattern including only an item value and an arbitrary item value permitting an arbitrary value as a generalized pattern, and a sequence pattern count part 140 counts the number of pieces of sequence data matching each generalized pattern as a generalized pattern appearance number. Thereafter, a sequence pattern finding part 150 extracts the generalized pattern wherein the counted generalized pattern appearance number is a prescribed threshold value or above, defines the sequence pattern wherein a selection item value is set to the arbitrary item value included in the extracted generalized pattern, and counts the number of the pieces of the sequence data matching each sequence pattern as the sequence pattern appearance number, and a sequence pattern output part 160 outputs the sequence pattern wherein the sequence pattern appearance number counted by the sequence pattern finding part 150 is a prescribed number or above as an extraction result.
COPYRIGHT: (C)2007,JPO&INPIT
Aoki Okamoto
Tatsuya Asai
高野洋、外3名,大規模な時系列テキストデータからのイベント時系列パターンの発見,第64回 知識ベースシステム研究会資料(SIG-KBS-A304),日本,社団法人人工知能学会,2004年 3月 1日,p.233-238
櫻井茂明、外3名,時系列イベントパターンマイニングにおける時間制約の導入,DEWS2005論文集 [online],日本,(社)電子情報通信学会データ工学研究専門委員会,2005年 5月 2日,p.1-8,[DL from http://www.ieice.org/iss/de/DEWS/DEWS2005/procs/papers/6C-o1.pdf]