Implementing a Fuzzy Classifier and Improving its Accuracy using Genetic Algorithms

The use of fuzzy logic to describe abstract concepts and when designing decision making systems to make decisions that are much closer to the way a human does is an interesting and useful area to explore. To effectively implement these types of systems expert knowledge is required in the application area. This can be a problem if the people designing a system of this nature are not experts in the application area since it is not always possible to have an expert working continuously with a project. This problem can be alleviated by using data supplied by an expert to let the system learn from it and thereby improve the system without continuous input from experts. In this paper the implementation of a fuzzy classifier that classifies the specie of the iris flower, based on the length and width of the sepal and petal, is presented. Its accuracy is then improved using a genetic algorithm to fine tune the membership functions.