Gender Classification Based On Text Mining

People's lives have been changed a lot with the development of the networktechnology, so do the original life style, and business model and communicationmodel.People trough various forms of communication tools for instant communication.So this non-verbal communication patterns hava similar difference in gender with theverbal communication.Is it possible to classifty the gender trough the text classification,and sum up the male and female specific word net in the non-verbal communicationenviornment.What's more, the non-verbal communication partterns enlarge thedifference of those hidend by the people in ordinary life by the usage of terms and theway of expressing the feeling and so on.Gender difference in the network environment includes not only the attitude andwillingness of usage to computer but also the shopping behavior in electroniccommerce.In order to verify the existence of the gender difference in the non-verbalcommunication enviornment,I select the content of the message form car and stockforum in TianYa,and with the approach of hyposis test verify the user's postingbehavior,with some kind of text classification method classify the gender,and by theassociation rule mining approach extract the feature word of the car and stock of themale and female.By this feature words,the related manufactures can formulate themarketing strategy.Through the similarity of words and mutual information of featureestabilish the dictionary to gender.