Fuzzy Bayesian Belief Network for Analyzing Medical Track Record

Bayesian Belief Network (BBN), one of the data mining classification methods, is used in this research for mining and analyzing medical track record from a relational data table. In this paper, the BBN concept is extended with meaningful fuzzy labels for mining fuzzy association rules. Meaningful fuzzy labels can be defined for each domain data. For example, fuzzy labels secondary disease and complication disease are defined for disease classification. We extend the concept of Mutual Information dealing with fuzzy labels for determining the relation between two fuzzy nodes. The highest fuzzy information gain is used for mining association among nodes. A brief algorithm is introduced to develop the proposed concept. Experimental results of the algorithm show processing time in the relation to the number of records and the number of nodes. The designed application gives a significant contribution to assist decision maker for analyzing and anticipating disease epidemic in a certain area.