Dynamic Calculation Method of Degree of Association Considering the Common Attributes of Target Concepts

DOI IR Web Site 7 Citations Open Access

Bibliographic Information

Other Title
  • 比較対象概念の共通属性を重視する動的関連度計算方式
  • ヒカク タイショウ ガイネン ノ キョウツウ ゾクセイ オ ジュウシ スル ドウテキ カンレンド ケイサン ホウシキ

Search this article

Description

We human beings associate when we are talking with each other. For example, someone says `cars', we can associate relative words, `Tire', `Engine', `Accident' …and so on. It is important to give association mechanism to a computer for conversation with human and computer. If computers can evaluate the depth of relationship between concepts, association mechanism is established by present relational words to input-word from various words. As a method of quantifying depth of relationship with words using concept-base, vector space model and a calculation method of degree of association have been proposed. In these methods, each method quantifies depth of relationship with words by evaluating weights of static attributes of target concepts. However, like multisence words, the concept has various meanings, and attributes are different each meaning generally. So degree of association between concepts depends on selection of attributes.In this paper, it is proposed that calculation method of degree of association by dynamic selection of attributes on target concepts, based on calculation method of degree of association that can evaluate depth of relationship between concepts by various points of view, and it is shown by experiment that proposed method more closely evaluate relationship between concepts like human beings evaluation.

Journal

Citations (7)*help

See more

Keywords

Details 詳細情報について

Report a problem

Back to top