Using Data Mining for Learning and Clustering FCM

Abstract — Fuzzy Cognitive Maps (FCMs) have successfully been applied in numerous domains to show relations between essential components. In some FCM, there are more nodes, which related to each other and more nodes means more complex in system behaviors and analysis. In this paper, a novel learning method used to construct FCMs based on historical data and by using data mining and DEMATEL method, a new method defined to reduce nodes number. This method cluster nodes in FCM based on their cause and effect behaviors. Keywords — Clustering, Data Mining, Fuzzy Cognitive Map (FCM), Learning . I. I NTRODUCTION N decision making process, there are some critical components. In most of the time, numbers of these critical components are numerous and they affect each other, so analyzing them is not easy. The efficiency of decision-making depends largely on the ability of decision-makers to analyze the complex cause and effect relationships and take productive actions based on the analysis. In complex systems, different components affect each other, and these cause and effect relations show system behavior. Cause and effect are two different concepts. Causes tell the reason why something happened, whereas effects are the results of that happening. In most of the systems, managers draw a system conceptualization graph to understand all of the system aspect. This diagram shows the cause and effect relations between system components. The information about these relations generated and enriched over time with the experience of managers who are expert in that field. There are two big challenges, at first, if there is no expert to construct the above mental model how this must be drawn and secondly, if there is a way to construct that diagram with more components, how they could be analyzed. Therefore, a new mechanism must be used to bridge these two gaps and constituted with experts in first case and cluster the components into similar categories based on their behaviors for the second one.