Analyzing Query Trails and Satisfaction Based on Browsing Behaviors

摘要

Nowadays, more and more web users use global search engines, such as Google and Bing, to hunt required information through query. At the same time, many Community-based Question Answering (C-Q&A) web sites supply large repositories of valuable information, knowledge and inner search engines to response the queries sent from users. In this paper, we do an exploratory analysis of the users' browsing behaviors to answer the following three questions. a)In users' search actions, which pattern is most used? They use the web sites that have inner search engine directly to get information, or use global search engines first, then go to certain web sites to get information, or use other patterns. b)How often can users find satisfying answers in C-Q&A web sites when web pages of C-Q&A web sites are listed as searching results in searching processes? c)Whether different categories in C-Q&A web sites have almost same satisfaction rate? In addition, is there any feature showing that the information of different categories has different life cycle? The analysis of query trails and satisfaction is based on browsing behaviors containing 81,168,263 access records which are created by 24,498 panel users. Analyzing results show that most frequent pattern of users' searching trails is from "inner-site search engines" to web pages. If web pages of C-Q&A web sites occurred in a searching process, about 50% are satisfying pages completely or on certain level. Experimental results also show that different categories of checked C-Q&A web site have different satisfaction rate.