DOCUMENT CLASSIFICATION USING INFORMATION THEORY AND A FAST BACK-PROPAGATION NEURAL NETWORK

In this paper, a fast back-propagation neural network is developed to build document classifiers and the information gain method is used for feature selection. According to the rank of the information gain of all the words contained in the documents, those words that contain more information to classify the documents are selected as the input features of the artificial neural network (ANN) classifiers. The neural network developed assumes a three-layer structure with a fast back-propagation learning algorithm. Because of the information contained in the vectors selected, (fie learning efficiency of the developed ANN is very high. For the output of the ANN, Shannon entropy is used to tune the threshold of (lie binary classifiers. The classifiers are tested using the Reuters corpus. Two performance measures are used to evaluate the performance of the classifiers and generally the results of this study are better than those claimed in literature.