Exploring Web Search Behavior Patterns to Personalize the Search Results

Tagging refers to the metadata that many users added in the form of keywords on photos, videos, and other resources for sharing the contents via the internet. However, there are several difficulties with tagging that come from tag variation, tag ambiguity and flat organization. This paper presents the integration of social tagging and the collaborative web search to improve users' or groups of users' satisfaction in search results. The characteristics of collaborative web search enable remote collaboration among small groups of users who are working either synchronously or asynchronously on web search tasks. Our main purpose is to analyze the users' behavior patterns from the search keywords obtained from the same and different groups of users. The patterns of web search behaviors are learned and used for generating the tag-based association rules. The results of analysis can be used to improve the quality of web search for a social search community. We designed and implemented a web application for effective use of tags to improve users' experience on web search. The users are asked to assign weights to the web links based on the relevancy of the web content to the tag. These weights are used to re-rank the original search results so that tags that best serving a user's needs are at the top of the ranked list. When a new user performs a web search, the system will generate a personalized tagging recommendation to the user for selection.