Recent advances on kernel fuzzy support vector machine model for supervised learning

A most fashionable off-the-shelf classifier is the Support Vector Machine. It is a powerful recent proceed of the data mining practitioner. A computing world has a lot to gain of new generation learning system. Statistical learning theory is a latest advances in supervised learning community. It is a feed forward network and binary learning machine with highly elegant properties. It plays a vital role in the reduction of machine learning problem into optimization problem, convex problems, linear programming, smaller quadratic programming, convex analysis, second order cone programming and so on. The SVM has enthused and established by the way of kernel learning algorithm. That maps data into some dot product feature space perform the linear algorithm. This kernel function is implemented by Platt´s sequential minimal optimization algorithm used in function estimation that can train efficiently and fast with Polykernel, Normalized poly kernel, Pearson VII function based Universal Kernel (PUK), Radial basis function. In this paper we focus Kernel Fuzzy Support Vector Machine to tune the kernel parameters to gain high performance from the classifier model. The trendy parameter technique k fold cross validation adopted under various kernel function and efficiency is empirically evaluate and observed a significant progress whilst the dataset is not linearly separable.