A Comparison among Three Neural Networks for Text Classification
Zhan Wang, Yifan He, Minghu Jiang · 2006
In this paper the effectiveness of three neural networks, the competitive, the backpropagation (BP) and the radial basis function (RBF), in text classification is examined. The competitive network is a kind of unsupervised learning which is used in data clustering. The BP network is one of the most widely used models among artificial neural network patterns and the RBF network has also showed its vitality in recent years. All of the three are fit for pattern classification and function approximation. The three networks are independently used automatic text classification. Experimental results show that BP and RBF network outperform competitive network because of the application of supervised learning. Besides its much shorter training time than BP, the RBF network makes precision and recall rates that are almost at the same level as BP's. Thus RBF network deserves more attention in the use of text classification