A hybrid mRMR-genetic based selection method for the prediction of epileptic seizures

Seizure forecasting would significantly improve the quality of life of epileptic patients. Predictive algorithms use high dimensionality data to evaluate the likelihood of an impending seizure. Dimensionality reduction is a key step towards the development of portable prediction systems. In this work, a comparative study of feature selection and classification methods was performed. Based on a Support Vector Machine and an Adaptive Neuro Fuzzy inference system, data reduction was performed by combining a minimum redundancy maximum relevance approach for electrodes selection and a genetic algorithm for features selection. The results show that the selected subset of features operates equally and sometimes even better than the whole features set.