【摘要】 随着问答社区信息的长期积累,越来越多的过时信息充斥在其中并被搜索引擎检索,给信息需求者带来不便。用户的网页浏览日志中隐性地包含用户的行为习惯,通过分析得到这些信息对判断网页信息时效性有着重要意义。文中提出针对网页浏览日志的查询过程划分方法,并在划分的基础之上对大量真实用户的浏览行为习惯做了统计分析。结果显示,用户查询一次信息平均浏览8.05个页面,用时6.28分钟,有将近1/3的查询在交替并发中进行,另外用户对于网站站内搜索的依赖较高。从浏览日志数据集中选取了一个社区网站的浏览记录来进行初步的网页信息时效性分析,结果表明造成用户不满意的原因主要是查询相关度不高,而过时信息只是其中一小部分。更多还原

【Abstract】 With the long-term accumulation of the Q & A community information,there is more and more outdated information indexed by search engines,bringing inconvenience to users.The log of a user’s browsing-behaviors contains the user’s behavioral intentions and habits,which can help analyze timeliness of the information.This paper proposed a query-process-division method for users’ browsing logs.Based on this method,a large number of real users’ browsing historical records were statistically analyzed.The results show that in average,a user browses 8.05Web pages in 6.28 minutes for one query.In addition,nearly 1/3of total queries carry out concurrently and alternately.It is also found that users rely on inner-site searching more.By analyzing the browsing historical records of a community site,we found that the users are not satisfied with the query results mainly because of the non-high-related results.Out-of-date information is only a small part in the query results.更多还原