Ischemia prediction using ANFIS

In this paper, a novel algorithm to make use of Adaptive Neuro Fuzzy Interference System (ANFIS) in order to predict Ischemia diseases in Electrocardiogram(ECG) signals is presented. Pre-processing for ECG signal has been performed in order to detect QRS complex. Then, baseline wandering and noise suppression is done. With the intention of extract influential features in Ischemia disease an ANFIS is employed as a predictor to predict Ischemia beats in ECG signals. Root Mean Square Error criterion is used to evaluate validity of predictor accuracy. Class of predicted beats is also recognized by an ANFIS classifier. They are classified as normal or Ischemia beats. Performance of prediction is evaluated in relation to computed Sensitivity (Se) and Specificity (Sp). Several recordings of ECG signals from European Society of Cardiology for ST-T database used in this study. Results of study were satisfactorily suitable in all Sensitivity (Se) and Specificity (Sp) factors.