Planning of Subject Database Based on the Similarity of Entities

The planning of subject database is an important process of information resource planning. In connection with the problems of affinity´s two-way nature and the aggregation´s inaccurate measurement of entities, a new concept of measuring the similarity of entities was proposed. On this basis, the formula of measuring the similarity between entity and entity group is given. The improved k-means algorithm is applied to plan the subject database. This method considers the overall relationship among entities and between entity and entity group abundantly. It reflects the relevance and tendency of entities more clearly and makes the result of the planning of subject database more reasonable.