欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
搜索引擎成为当今在互联网上进行信息检索最常用的工具.主流搜索引擎以与用户查询的相关度排序返回搜索结果,且自然语言中存在的＂一义多词＂和＂一词多义＂现象,用户很难清楚表达他们的意图,导致往往花费较长时间从结果列表中选择所感兴趣的话题.针对这种状况,采用网页聚类技术对标题和摘要进行聚类后,并可视化地以树和图的方式向用户快速、全貌和直观地展示搜索结果,明显改善了用户搜索体验.在此基础上设计了网页聚类原型系统ECE（effective clustering engine）,实验结果表明该算法具有聚类结果可读性好以及聚类准确度比较高的优点.
英文摘要：
Nowadays search engines are the most common tools for information retrieval on the internet.However,there are several limitations such as low search coverage and dynamic characteristic of web pages,it is the reason why no breakthrough made on users＇ searching experience recent years.The leading search engines will return a long list of records that are sorted by the correlation with the queries,the phenomena of synonymy and polysemy make users express their intention difficultly and spend much time on selecting web pages they are interested in.This paper aims at enhancing searching experience using data analysis technologies.Through clustering and visualizing web search results,then grouping the clustering results according to some criterions,it makes users locate their interested information quickly.The data structure related to suffix tree are being widely used in string processing and text compression.The clustering algorithm based on suffix tree which makes it easy to recognize the shared phrases among web pages can be used to cluster web pages,it improves the clustering efficiency as not to calculate the similarities between pair-wise documents,and assigns meaningful labels for the clustering results to enhance the readability,also improves end users＇ searching experience through visualization.An effective clustering engine prototype system named effective clustering engine has been built on this approach.The algorithm is quite efficient,and the clustering results are readable and accurate verified by the experiments.
同期刊论文项目
同项目期刊论文
期刊信息
位置：