首页 | 本学科首页   官方微博 | 高级检索  
     

多种搜索行为联合分析方法研究
引用本文:张引,高克宁,张斌. 多种搜索行为联合分析方法研究[J]. 东北大学学报(自然科学版), 2014, 35(9): 1249-1252. DOI: 10.12068/j.issn.1005-3026.2014.09.008
作者姓名:张引  高克宁  张斌
作者单位:(1 东北大学 信息科学与工程学院, 辽宁 沈阳110819; 2 东北大学 医学影像计算教育部重点实验室, 辽宁 沈阳110819)
基金项目:国家自然科学基金资助项目(61073062);沈阳市科学技术计划项目(F11-264-1-35,F11-264-1-33);中央高校基本科研业务费专项资金资助项目(N120304002)
摘    要:针对现有搜索行为分析方法只能分析单一类型搜索行为,无法有效获取用户兴趣的问题,提出多种搜索行为联合分析方法.通过使用M5模型对页面停留时间、鼠标点击次数、页面重访问次数及滑块移动次数4种类型的用户行为进行联合分析,从多个角度获取用户行为信息用于分析用户兴趣,并实现了对多种搜索行为构成的高维数据进行联合分析,同时满足了在线行为分析中对结果计算的实时性要求.实验表明该方法可以提供比Belkin方法更高的行为分析质量.

关 键 词:搜索引擎  个性化搜索  用户行为分析  搜索行为特征  用户行为模型  

Joint Analysis Method of Multiple Search Behaviors
ZHANG Yin;GAO Ke-ning;ZHANG Bin. Joint Analysis Method of Multiple Search Behaviors[J]. Journal of Northeastern University(Natural Science), 2014, 35(9): 1249-1252. DOI: 10.12068/j.issn.1005-3026.2014.09.008
Authors:ZHANG Yin  GAO Ke-ning  ZHANG Bin
Affiliation:1 School of Information Science & Engineering, Northeastern University, Shenyang 110819, China; 2 Key Laboratory of Medical Image Computing of Ministry of Education, Northeastern University, Shenyang 110819, China.
Abstract:Aiming at the fact that available search behavior analysis methods could only analyze one single type of search behavior, which leads to the problems that the user interest could not be effectively get, a joint analysis method of multiple search behaviors was proposed. By combining the analysis of page dwell time, mouse click times,page revisit times and slider move times using M5 model, clues from more angles other than one single type of user behavior were gained for user interest analysis, and joint analysis of high dimension data composed of multiple user behaviors was realized with timeliness ensured for online behavior analysis. Experiment results show that the proposed method can analyze the behaviors better than compared approaches.
Keywords:search engine   personalized search   user behavior analysis   search behavior character   user behavior model  
本文献已被 CNKI 等数据库收录!
点击此处可从《东北大学学报(自然科学版)》浏览原始摘要信息
点击此处可从《东北大学学报(自然科学版)》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号