Towards new directions of data mining by evolutionary fuzzy rules and symbolic regression

There are various techniques for data mining and data analysis. Among them, hybrid approaches combining two or more fundamental methods gain importance as the complexity and dimension of real world problems and data sets grows. Fuzzy sets and fuzzy logic can be used for efficient data classification by the means of fuzzy rules and classifiers. This study presents an application of genetic programming to the evolution of fuzzy rules based on the concept of extended Boolean queries. Fuzzy rules are used as symbolic classifiers learned from data and used to label data records and to predict the value of an output variable. An example of the application of such a hybrid evolutionary-fuzzy data mining approach to a real world problem is presented.