Multiagent Data Mining for Adaptive Monitoring of Business Information

 

Rey-Long Liu

 

ABSTRACT

Information monitoring (IM) is an essential activity for management and control. It inquires current statuses of critical information items from various kinds of information sources (e.g. individual persons and servers). Once an update is found, it validates whether an exception happens and triggers suitable actions such as exception handling, event logging, and user notification. Since data inquiry may be costly and uncontrolled inquiry may even exhaust the information sources, IM should be conducted under an inquiry-bounded constraint (i.e. limiting the times of inquiry). Under the constraint, an IM system should inquire right targets at right time by adapting to the update behavior (i.e. rough update frequency at each time period) of each information item. This requirement brings a significant challenge to data mining: mining the imprecise data that is sampled in an inquiry-bounded manner. The data is imprecise in the sense that the monitor cannot know when, how many, and how frequently updates really happened. In this project, we tackle the challenge by developing a scalable multiagent model in which each agent performs autonomous mining and monitoring. It aims to capture a higher percentage of updates in a timelier manner by conducting fewer times of inquiry. The framework is to be theoretically and empirical evaluated. It is also to be applied to a real-world application in which its performances and contributions to the manager may be investigated. The integrative study may bring significant impacts to the related disciplines including data mining, multiagent technology, and management by exceptions.

Keywords: data mining, imprecise data, multiagent technology, information monitoring

 

     Back to Research Project