EXTRACTION DEVICE FOR LANGUAGE CHARACTERISTICS, EXTRACTION DEVICE FOR UNIQUE EXPRESSIONS, EXTRACTION METHOD, AND PROGRAM

• Main Authors: TOMITA JUNJI, SAITO KUNIKO, KOBAYASHI NOZOMI
• Format: Patent
• Language: English; Japanese
• Published: 31.10.2019
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

To skillfully absorb difference in characteristics to be taken account for each language and embodying common unique expression extraction as a processing system.SOLUTION: A language characteristic extraction device 11 comprises a language characteristic extraction section which: selects an extraction rule corresponding to the characteristics of a target language from a collection of the extraction rules common to a plurality of languages; deems a method to extract the detailed origin tailored to the target language, and what is defined as an output condition, as a characteristic extraction rule by language; defines the characteristic extraction rule by language for each of a plurality of target languages; refers to the characteristic extraction rule by language for extracting an origin relating to an expression or a part of speech that is defined against a morpheme analysis result of an input sentence and against the language of the input sentence, and that is further included in the morpheme analysis result; extracts the origin corresponding to the language; and outputs thereof as a language characteristic extraction result.SELECTED DRAWING: Figure 1 【課題】言語ごとに考慮すべき特徴の違いをうまく吸収し、処理系としては共通の固有表現抽出を実現することができるようにする。【解決手段】言語特徴の抽出装置１１は、複数の言語に共通した抽象ルールの集合から、対象言語の特徴に応じた抽象ルールを選択し、前記対象言語に合わせた具体的な素性の抽出方法、及び出力条件として定義したものを言語別特徴抽出ルールとし、複数の対象言語の各々に対し、前記言語別特徴抽出ルールを定義し、入力文の形態素解析結果に対し、前記入力文の言語に対して定義され、かつ前記形態素解析結果に含まれる表記又は品詞に関する素性を抽出するための前記言語別特徴抽出ルールを参照して、前記言語に応じた素性を抽出し、言語特徴抽出結果として出力する言語特徴抽出部を備える。【選択図】図１