Hossein Jamali - Islamic Azad University, Arak Branch, Arak, Iran
Mehdi haghighizadeh - Islamic Azad University, Arak Branch, Arak, Iran

Recommender systems affecting the in user leadership and guidance, in the midst of huge amount of choices possible, to reach the useful and interesting option for him. With the expansion of Internet world, many and different information stored by supplied product sites but recommender systems usually ben-efit small part of this information. One of the information that is largely ignored user text comments which contain important information of users that they will take the feelings and interests. In this paper we provide method that all information available with text comments accommodates in system. With this work is created for each user profile that is personalized for each individual. The proposed algo-rithm determines the effect of the types of information and obtains better results than other methods.

Recommender systems, Collaborative filtering, Comments mining, Information analysis