Support vector machine classification for large datasets using decision tree and Fisher linear discriminant

The training of a support vector machine (SVM) has a time complexity between O(n) and O(n). Most training algorithms for SVM are not suitable for large data sets. Decision trees can simplify SVM training, however the classification accuracy becomes lower when there are inseparable points. This paper introduces a novel method for SVM classification. A decision tree is used to detect low entropy regions in input space. We use Fisher’s linear discriminant to detect the data near to support vectors. Experimental results demonstrate that our approach has good classification accuracy and low standard deviation, the training is significantly faster than other training methods.