The dramatic development of social networks in recent years has made it possible for users to express their interests and opinions on various issues on a daily basis. As such, a large amount of user information is available and extracted from this information can be used in a variety of applications, including recommender systems and digital marketing. In this project, using neural networks and modeling of texts sent and users’ networks, each user’s profile is estimated with a set of predefined features and user profiles. ) Is obtained. Features can include people’s personal information, people’s views on various topics, fake user profiles, as well as the amount of user influence on the social network. The information from the network that will be used can be from the texts sent by the user on the network, the network of users communicating with each other and the reaction of users to each other’s texts. Due to the relative ease of data collection and previous work in this area, the social network Twitter will be used to collect user data. Information disclosed by users on this network or similar social networks is considered to be actual user information. To evaluate the model, the accuracy criterion for multi-category features and the mean absolute error and Pearson correlation coefficient for quantitative characteristics will be used. The language of user submissions can be English or Persian. Initially, the focus will be on English and, if possible (getting the right data) in Farsi.