关键词: 元搜索引擎, 模糊C均值, 关联规则, 结果聚类, 关联词矩阵

Abstract: After making word segmentation on some results which are returned from component search engine and extracting the main keywords，then it uses association rules to build up an associated word matrix，and adopts the FCM（Fuzzy C-Means） algorithm to make fuzzy clustering for the results based on the matrix.Finally，it depends on the degree of relevance to return the results for users.The paper compares this method above with K-means（K-means clustering） algorithm through the experimentation and demonstrates that the method above can ensure operating efficiency and the effectiveness of clustering，further more，it improves the ranking position of the related results，so can well satisfy the requirements of the users.

Key words: meta search engine, Fuzzy C-Means（FCM）, association rules, results clustering, associated word matrix