IMPROVING WEB SEARCH BY EXPLOITING SEARCH LOGS

With the increased use of Web search engines, acute needs evolve for more adaptive and more personalizable Information Retrieval (IR) systems. This study proposes an innovative probabilistic method exploiting search logs to gather useful data about contexts and users to support adaptive retrieval. Real users' search logs from an operational Web search engine, Infocious, were processed to obtain past queries and click-through data for adaptive indexing and unified probabilistic retrieval. An empirical experiment of retrieval effectiveness was conducted. The results demonstrate that the log-based probabilistic system yields statistically superior performance over the baseline system.