Classification of ECG waveform using feature selection algorithm

The ECG classification problems have been solved by means of a methodology, which has the capability to enhance the ECG classification performance. This method reduces the computational complexity which mainly occurs during the feature selection. The computational requirements of exhaustive search method (those which test all possible subsets) increase exponentially with the number of features in the original set. The proposed system use particle swarm optimization for the selection of feature subset. PSO is attractive for feature selection, in that particle swarms will discover best feature combination as they fly within the best subset space. Some classifiers such as MLP, which start at random chosen point and then adjust weights to move in the direction. Although the training phase takes long time. Thus SVM is used for classification, which is based on local approximation strategy. It reduces the number of operations in learning mode and it is well suited for larger datasets.