A  Neuro-Fuzzy-Genetic Architecture for Data Mining 

Korakot  Hemsathapat


This research presents a method for rule extraction from a neuro-fuzzy-genetic based data mining architecture.  In the architecture, all input variables are first preprocessed and all continuous variables are fuzzified.  Principal Component Analysis (PCA) is then applied to reduce the dimension of the input variables in finding combinations of variables, or factors, that describe major trends in the data.  The reduced dimension of input variables are then used to train Probabilistic Neural Networks (PNNs) to classify the dataset according to the classes considered.  A rule extraction technique is then applied in order to extract explicit knowledge from the trained neural networks and then represent the knowledge in the form of crisp and fuzzy If-Then-rules.  In the final stage a genetic algorithm is used as a rule pruning module to eliminate weak rules that are still in the rule bases even though the classification accuracy of the rule bases has not changed or improved.  The application of the neuro-fuzzy-genetic based data mining architecture is currently implemented with the meningoencephalitis diagnosis dataset in JSAI KDD Challenge 2001.


* For detailed information please visit this research's homepage

 Return to SESL homepage