EXTRACTION SYSTEM AND PROGRAM

• Main Authors: MATSUSHIMA HIROYASU, HIRANO MASANORI, IZUMI KIYOSHI, KATO ATSUO, SAKACHI YASUNORI, MORIOKA TSUGUTO, KIMURA SHOKO, NAGAO SHINTARO
• Format: Patent
• Language: English; Japanese
• Published: 09.07.2020
• Subjects: CALCULATING; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

To provide an extraction system capable of reducing a work load on a user who extracts an extraction target related to a theme, improving work efficiency, and optimizing an extraction result.SOLUTION: After determining a relevant word related to a theme by using a similarity degree Sword between a word of a theme and another word, an extraction system 10 uses the similarity degree Sword of each related word and a text data reference degree of relevance Pword that is a matching rate obtained by matching with unique text data indicating features of each extraction target, calculates a correction degree of relevance SPword of each related word, determines a focused word based on a magnitude of the correction degree of relevance SPword, obtains the number of occurrences COUNTword of each word for each focused word or theme word in the unique text data, and uses these to calculate extraction target/theme relevance FS for each extraction target.SELECTED DRAWING: Figure 1 【課題】テーマに関連する抽出対象を抽出するユーザの作業負担の軽減、作業効率の向上、抽出結果の適正化を図ることができる抽出システムを提供する。【解決手段】抽出システム１０では、テーマの単語と他の単語との間の類似度Ｓｗｏｒｄを用いて、テーマに関連する関連単語を決定した後、各関連単語の類似度Ｓｗｏｒｄと、各抽出対象の特徴を示す固有テキストデータとの照合により得られた適合率であるテキストデータ基準関連度Ｐｗｏｒｄとを用いて、各関連単語の修正関連度ＳＰｗｏｒｄを算出し、この修正関連度ＳＰｗｏｒｄの大小で着目単語を決定し、固有テキストデータにおける各着目単語やテーマ単語についての単語別出現回数ＣＯＵＮＴｗｏｒｄを求め、これらを用いて各抽出対象についての抽出対象・テーマ関連度ＦＳを算出する。【選択図】図１