GA-Driven Fuzzy Set-Based Polynomial Neural Networks with Information Granules for Multi-variable Software Process

摘要

In this paper, we investigate a GA-driven fuzzy-neural networks-Fuzzy Set-based Polynomial Neural Networks (FSPNN) with information granules for the software engineering field where the dimension of dataset is high. Fuzzy Set-based Polynomial Neural Networks (FSPNN) are based on a fuzzy set-based polynomial neuron (FSPN) whose fuzzy rules include the information granules obtained through Information Granulation. The information Granules are capable of representing the specific characteristic of the system. We have developed a design methodology (genetic optimization using real number type gene Genetic Algorithms) to find the optimal structures for fuzzy-neural networks which are the number of input variables, the order of the polynomial, the number of membership functions, and a collection of the specific subset of input variables. The augmented and genetically developed FSPNN (gFSPNN) with aids of information granules results in being structurally optimized and information granules obtained by information granulation are able to help a GA-driven FSPNN showing good approximation on the field of software engineering. The GA-based design procedure being applied at each layer of FSPNN leads to the selection of the most suitable nodes (or FSPNs) available within the FSPNN. Real number genetic algorithms are capable of reducing the solution space more than conventional genetic algorithms with binary genetype chromosomes. The performance of GA-driven FSPNN (gFSPNN) with aid of real number genetic algorithms is quantified through experimentation where we use a Boston housing data.