基于查询意图聚类的实时搜索建议

对于搜索引擎返回的结果太多且较少考虑用户个性差异等缺陷,提出根据用户查询意图,实时给予多个主题的搜索建议,帮助用户更准确地描述所需信息,修正查询词与真实意图之间的差距,提高检索效率。同时运用K-means算法,对资源类别的意图特征值相似用户进行聚类,缩小查找目标对象最近邻居的范围,提高搜索建议的实时响应速度。实验结果表明,该方法是可行的。

Real-Time Search Suggestions Based on the Clustering of the User’ s Query Intent

Aimed at the defects that the search engine offers too many results and is lack of considering the differences between the user’s personalities, this paper offers a way to give users real-time search suggestions of multi theme according to the user’s search intent in order to help the users describe the information in need more accurately, as well as narrow the gap between the query word and the user’s real intentions to increase the search efficiency. At the same time, it uses K-means to cluster users who are similar in their intent eigenvalue of resources categories, narrow the range of the nearest neighbor of the searching target, as well as to speed up the real-time response of the search suggestions. The experiment result shows that this method is practical.