Siamak hekmatiuan zadeh pour - Master of Information Technology, University of Qom
Amir jalaly Bidgoly - Assistant Professor at Qom University

The widespread use of the Internet among people has led to the exchange of a large amount of comments in cyberspace. On most Internet networks users Anonymously record their opinions. However, the author s characteristics, such as age and gender, can be widely used in better analysis of the author s opinion or in identifying the views of different sections of society. In this research, the problem of automatically identifying the gender of the author has been examined from the text of written comments in Persian language in Myket site . The work done to identify a gender focuses more on the text of the comment. In this study, the side features along with the comment text were also used. To categorize comments Multilayer perceptron neural network and Naive Bayes algorithms were used. We showed that the use of side features along with the comment can improve the accuracy of the model. The highest accuracy achieve Detection Gender is the use of side features of 86.4.

Gender Identification, Naive Bayes, neural network, side features