RELATED DATA GENERATION DEVICE, RELATED DATA GENERATION METHOD AND PROGRAM

• Main Authors: HIROI KAZUE, HORIBE YASUKI, SAWAJIRI HARUHIKO, ISHIGURO MASAO, HAYASHI AKIO
• Format: Patent
• Language: English; Japanese
• Published: 28.09.2015
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

PROBLEM TO BE SOLVED: To generate related data having high relevance to a predetermined keyword and including a related word with higher freshness.SOLUTION: A related data generation device includes: co-occurrence word data generation part for generating co-occurrence word data including a co-occurrence word that is a vocabulary being used together with a predetermined keyword in submission data for all periods in submission data which are submitted for a plurality of periods different from each other, and an appearance frequency of the co-occurrence word; and a related data generation part for generating related data storing the co-occurrence word as a normal related word in the case where temporal fluctuation of the appearance frequency of the co-occurrence word is lower than a first threshold value and the appearance frequency is higher than a second threshold value. 【課題】所定のキーワードに高い関連性を有し、より鮮度の高い関連語を含む関連データを生成することができる。【解決手段】 相互に異なる複数の期間に投稿された投稿データのうち、全ての期間の投稿データに所定のキーワードと共に用いられている語彙である共起語と、該共起語の出現頻度とを格納した共起語データを生成する共起語データ生成部と、前記共起語の出現頻度の時間的変動が第１の閾値よりも小さく、かつ、出現頻度が第２の閾値よりも高い場合、該共起語を通常関連語として格納した関連データを生成する関連データ生成部と、を備える。【選択図】図1