Web Based Parallel/Distributed Medical Data Mining Using Software Agents

Using Software Agents Hillol Kargupta, Brian Sta ord, and Ilker Hamzaoglu Computational Science Methods Group X Division, Los Alamos National Laboratory P.O. Box 1663, MS F645, Los Alamos, NM, 87545 This paper describes an experimental parallel/distributed data mining system PADMA (PArallel Data Mining Agents) that uses software agents for local data accessing and analysis and a web based interface for interactive data visualization. It also presents the results of applying PADMA for detecting patterns in unstructured texts of postmortem reports and laboratory test data for Hepatitis C patients. Introduction Data mining involves extraction, transformation, and presentation of data in useful form. As we move more and more toward a paper-less society, each of these components of data mining is likely to face the challenges of dealing with large volume of data and the very distributed nature of the data storage and computing environments. Medical databases are often ideal candidates for large scale, possibly distributed data mining applications. In this paper we describe an experimental software agent based system for parallel/distributed data miningPADMA (PArallel Data Mining Agents). PADMA is characterized by agent based distributed data accessing, distributed data analysis, and web based interactive data visualization. This paper also presents results of applying PADMA in medical databases. Section presents a general overview of the PADMA system. The parallel relational database accessing operations of PADMA agents are described in Section . Section describes the data analysis capabilities of the agents. Section describes the web-based Disk Disk DM DM