Data mining for fuzzy decision tree structure with a genetic program

Abstract

A resource manager (RM), a fuzzy logic based expert system, has been developed. The RM automatically allocates resources in real-time over many dissimilar agents. A new data mining algorithm that uses a genetic program, an algorithm that evolves other computer programs, as a data mining function has been developed to evolve fuzzy decision trees for the resource manager. It not only determines the fuzzy decision tree structure it also creates fuzzy rules while mining scenario databases. The genetic program’s structure is discussed as well as the terminal set, function set, the operations of cross-over and mutation, and the construction of the database used for data mining. Finally, an example of a fuzzy decision tree generated by this algorithm is discussed.