Support vector machine and neural network with others for two dimensional classification problem: an empirical study

摘要

The performance analysis of most commonly used dashers Support Vector Machine (SVM) and Neural Networks (NN) are compared with others six different classifiers on seventeen quite different, standard and extensively used datasets in terms of classification error rates and computational times. It is found that the average error rates for a majority of the classifiers are doses with each other but the computational times of the classifiers differ over a wide range. SVM algorithm based on statistical naming has the lowest average error mm and computationally it as faster than NN but computationally expensive than other classifiers. SVM and NN have also been considered with changing parameter values.