Study on Handwritten Digit Recognition using Support vector machine

Abstract

A machine learning model of handwritten digit recognition based on SVC is established in this paper. Then the influence of sample number, kernel function parameters, penalty coefficients and other parameters on the prediction model is analysed. This results show that training samples have a significant impact on the model. There is an acceptable training number. Different kernel functions have a different effect on the accuracy of the model. The radial basis function is the best in recognition model. The recognition rate increases continuously with C, while the recognition rate increases first with gamma increases, and when gamma increases to a certain value, precision begins to decline.