Text orientation classification can be widely applied to information retrieval

product online tracking

diagnoses of public opinions and chat systems.Two kinds of hybrid feature selecting methods based on words with restricted part of speech and information gain

and sentiment words and information gain are proposed respectively.A classifier based on support vector machine is designed for text orientation classification.As the training and testing data

review texts about car products

basketball games and Sino-Japanese relationship are used to validate the methods.The results indicate that the hybrid feature selecting method based on words with restricted part of speech and information gain is superior to the another method when the classifier is based on support vector machine.