| The study and application of Decision Support Systems (DSS) has become one of the major topics of MIS. An underlying component of a DSS is the decision model. Data mining is a useful technique to discover the hidden pattern or relationship among data. Thus, a decision model can be built upon the findings. However, unlike machine learning, a comprehensive data mining process covers from data collection to the use of discovered knowledge. How to gather the data from different resources becomes the first problem. To some industries, which are mainly consisted of small-scale business organizations, data resources are widely separated within each organization, even though the data logically belong to one data scheme. In this case the distributed databases exist but the distributed database management system does not. The data warehousing approach is introduced to overcome the problem of data acquisition. The other difficulty of data mining is the demand upon computing resources, and it can be computation intensive.; This dissertation casts the above difficulties to a global network environment, namely the Internet. The goal is to integral a data warehouse, data mining, distributed databases and distributed computation into a single framework under the Internet infrastructure. This work fully develops this framework and implement a prototype system referred to as D4 that covers data acquisition, data mining and knowledge deployment. |
