Data mining: artificial intelligence in data analysis

Abstract

Summary form only given. Data mining is a fast-growing area. The first Knowledge Discovery in Databases Workshop was held in August 1989, in conjunction with the 1989 International Joint Conference on Artificial Intelligence, and this workshop series became the International Conference on Knowledge Discovery and Data Mining (KDD) in 1995. In 2003, there were a total of 15 data mining conferences, most of which are listed at http://www.kdnuggets.com/meetings/meetings-2OO3-past.html. These 15 conferences do not include various artificial intelligence (AI), statistics and database conferences (and their workshops) that also solicited and accepted data mining related papers, such as DC AI, ICML, ICTAI, COMPSTAT, AI & Statistics, SIGMOD, VLDB, ICDE, and CIKM. Among various data mining conferences, KDD and ICDM (the IEEE International Conference on Data Mining) are arguably (or unarguably) the two premier ones in the field. ICDM was established in 2000, sponsored by the IEEE Computer Society, and had its first annual meeting in 2001. This work reviews the topics of interest from ICDM from an AI perspective, and analyze common topics in data mining and AI, including key AI ideas that have been used in both data mining and machine learning. We also discuss two current research projects on (1) user-centered agents for biological information exploration on the Web, and (2) dynamic classifier selection in dealing with streaming data. Both projects apply data mining techniques for intelligent analysis of large volumes of data.