tailieunhanh - Báo cáo khoa học: "Query log analysis with GALATEAS LangLog"

This article describes GALATEAS LangLog, a system performing Search Log Analysis. LangLog illustrates how NLP technologies can be a powerful support tool for market research even when the source of information is a collection of queries each one consisting of few words. We push the standard Search Log Analysis forward taking into account the semantics of the queries. The main innovation of LangLog is the implementation of two highly customizable components that cluster and classify the queries in the log. . | Query log analysis with GALATEAS LangLog Marco Trevisan and Luca Dini CELI trevisan@ dini@ Igor Barsanti Gonetwork Eduard Barbu Universita di Trento Nikolaos Lagos Xerox Research Centre Europe Frederique Segond and Mathieu Rhulmann Ed Vald Objet Direct Bridgeman Art Library fsegond@ mruhlmann@ Abstract This article describes GALATEAS LangLog a system performing Search Log Analysis. LangLog illustrates how NLP technologies can be a powerful support tool for market research even when the source of information is a collection of queries each one consisting of few words. We push the standard Search Log Analysis forward taking into account the semantics of the queries. The main innovation of LangLog is the implementation of two highly customizable components that cluster and classify the queries in the log. 1 Introduction Transaction logs become increasingly important for studying the user interaction with systems like Web Searching Engines Digital Libraries Intranet Servers and others Jansen 2006 . Various service providers keep log files recording the user interaction with the searching engines. Transaction logs are useful to understand the user search strategy but also to improve query suggestions Wen and Zhang 2003 and to enhance the retrieval quality of search engines Joachims 2002 . The process of analyzing the transaction logs to understand the user behaviour and to assess the system performance is known as Transaction Log Analysis TLA . Transaction Log Analysis is concerned with the analysis of both browsing and searching activity inside a website. The analysis of transaction logs that focuses on search activity only is known as Search Log Analysis SLA . According to Jansen 2008 both TLA and SLA have three stages data collection data preparation and data analysis. In the data collection stage one collects data .