Neural Networks for the Recognition of Traditional Chinese Handwriting

In recent years, Support Vector Machine [1] has been a popular machine learning algorithm. SVM has the better recognize capability and faster calculation speed than the general neural network. Furthermore, it does not have the over-learning situation. There are so many researches prove that SVM have good performance of recognition. Probabilistic neural network (PNN) is a kind of neural network based on Bayesian decision theory. PNN´s highly regarded due to its short training time, and also, it does not have the iterative process. In this paper, we used PNN and SVM as the recognition of Traditional Chinese handwriting tool. The database were made of 20 people´s hand-writing in Traditional Chinese, according to everyone´s handwriting habits and their using of different quantization methods, in order to explore the feasibility of using handwriting recognition as an identification identity. Experimental results show that the best rate to use SVM to recognize is 75% while the best rate for PNN best rate is 80%.