中文摘要：

用户在利用搜索引擎进行信息检索时，查询条件往往存在歧义，这导致搜索结果的多样性和冗余性。传统的方法主要是基于语义分析或构建知识库，此类方法在实际应用中的可行性不高。本文基于搜索引擎的分类信息，实现了一个简单有效的分类搜索系统。它首先根据用户的查询条件，将返回的搜索结果进行分类，并以树形目录的形式展示给用户，而后根据用户的点击数据，逐步确定用户的搜索意图，从而达到了查询歧义消减的目的。论文详细介绍了系统的设计思想、架构和工作流程。测试实例表明，该系统可以在一定程度上确定用户的查询意图，为用户返回更加准确的搜索结果。

英文摘要：

Users＇ query ambiguity on information retrieval leads to the duplication and redundancy of search results. The traditional approaches are mainly based on semantic analysis or certain knowledge bases, and wouldn＇ t be viable in practice. This paper presents a simple and effective classification search system based on the classes＇ information provided by search engines. It first groups the initial returned search results into some classes, and then displays them in the form of tree directories to users. After getting the users＇ click-through data, it can distinguish users＇ search intention step by step, hence achieves the aim of query ambiguity reduction. This paper elaborates the system design ideas, architecture and work flow. The test example shows that this system is able to determine the users＇ search intention to a certain degree and return more accurate search results to users.