For a computer to perform intelligent information processing requires functions that can extract concepts from words, as humans do, and then associate those concepts with related concepts. In order to implement this association Junction, it is necessary to quantify the degree of association between two concepts. In the present paper, we propose a method for quantifying degree of association incorporating viewpoint that uses a concept base (a knowledge base that expresses concepts as a collection of pairs, each pair consisting of an attribute word used to describe the concept and a weighting that expresses the word's importance). Here, "Viewpoint" is the perspective from which a concept is viewed; for example, consider the degree of association between "airplane" and "automobile", and the degree of association between "airplane" and "bird". From the viewpoint of "vehicle", "airplane" and "automobile" are highly related, while from the viewpoint of "flight", "airplane" and "bird" are highly related We present herein a comparison of two methods for calculating degree of association incorporating viewpoint, and demonstrate that the method involving modulation of attribute weightings based on viewpoint results in degree of association calculations that are closer to human senses.
展开▼
