Title

Extraction of Associative Attributes from Nouns and Quantitative Expression of Prototype Concept

Authors

Maya Ando (Graduate School of Media and Governance, Keio University 5322 Endo, Fujisawa-shi, Kanagawa, 252-8520, Japan)

Jun Okamoto (Graduate School of Media and Governance, Keio University 5322 Endo, Fujisawa-shi, Kanagawa, 252-8520, Japan)

Shun Ishizaki (Graduate School of Media and Governance, Keio University 5322 Endo, Fujisawa-shi, Kanagawa, 252-8520, Japan)

Session

WP2: Lexicons

Abstract

One of the purposes of this research is to formalize similarity among nouns by using attributes associated from the nouns, and then using the similarity, to formalize prototypes of categories. The other purpose is to extract features of nouns by using adjectives or adjective-like words obtained by the association experiments and to formalize importance of the nouns with the words. We constructed an associative concept dictionary using many kinds of attributes associated from nouns. Similarity among nouns was calculated by using their associated attributes with inner product methods, where the nouns were organized in a hierarchical structure using generalized or specific relations. This paper discusses similarity between nouns using their attributes. We found that the similarity of nouns located at lower levels has a high score in many cases. Then prototypes are quantitatively formalized among Japanese noun concepts. It uses similarities of part/material concepts, features, and action concepts, and distance values between the noun and its lower-level concepts. Such formalized prototypes are compared with a result of human questionnaire experiments to obtain a good correspondence among them. 

Keywords

Nouns, Extraction

Full Paper

270.pdf