DETECTION OF LOCAL SYSTEM STRUCTURES FROM COMPLEX DATABASE

摘要

In this paper, we discuss a methodology about detection of local system structures from complex database. In the complex database, some heterogeneous structures are often included together. In addition, some symbolic attributes and some numerical attributes are often stored together in the databases. In system modeling, the detection of local system structures is important task, because it influences not only the accuracy of a rule-based model but also the expression of the each rule. Requirement of complex systems modeling techniques is increasing under these conditions. This paper describes an agent-based approach to complex systems modeling. Each agent has a role to collect data points (objects) based on similarities. The similarity is introduced so that agent can treat mixed database. The interaction between the agents focuses on squabbling about objects considering all attributes. The agent identifies an if-then rule based on the collected data. The agent tries to determine variables for rule description. There is no objective function differing from the conventional mining methods. Taking this approach, there are some advantage points of the proposed method. One of the advantage points is that it is not required to determine the number of the clusters before performing the process.