A framework for a World Wide Web-based Data Mining system

The phenomenal growth of the World Wide Web (WWW) introduces new vistas for trade and commerce. In addition to overcoming the difficulties of distances, cultures and languages, the WWW provides another important business resource: information. Spurred on by the development of better data storage and retrieval techniques, companies are beginning to provide more information about themselves and their activities. This widespread information dissemination necessitates the development of tools which can process it. The application of machine learning algorithms to this vast quantity of data can yield useful information, which can be used for individual and corporate decision support. One example of many possible applications of Data Mining, the system described in this paper, deals with stock-market data available on certain Web-sites. It classifies these stocks asbuyorsellusing the decision tree technique. The user interaction is through a Web browser, which makes the Web-based nature of the data gathering and processing transparent to the user. A number of experiments are conducted to gauge the performance of the classification.