基于用户聚类的框计算改进模型研究  

An Improved Box-computing Model Based on Users Clustering

作　　者：丁向荣[1] 

机构地区：[1]淮安信息职业技术学院,江苏淮安223003

出　　处：《科技通报》2012年第12期57-59,共3页Bulletin of Science and Technology

摘　　要：随着互联网信息量的急剧膨胀,搜索引擎逐渐成为人们获取信息的重要途径。为迎合这一趋势,百度推出的框计算的核心在于准确识别用户需求,提高搜索引擎的响应速度和结果的准确度。然而,随着用户需求的日益复杂和多样化,如何满足其个性化需求是框计算研究的重点和难点之一。针对此问题,提出了一种基于浏览行为的用户聚类方法,并结合框计算精准性制约因素设计了一种改进的框计算模型。With the rapid exploration of information on the Internet,search engine has become one of the major information accessing channels for users.The ＂box computing＂ technique developed by Baidu aims to precisely identify the information needs of web users,and improve the response speed and efficacy of search engines.With the increasingly complexity and diversity of users＇ information requests,satisfying the personalized needs of users has become one of the major challenges for box-computing.This paper presents a novel user clustering method based on users＇ browsing behaviors.An improved box-computing model is developed by integrating the proposed user clustering method with the accuracy constraint of box-computing.

关 键 词：框计算 个性化搜索 WEB日志挖掘 用户聚类 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]