Recommended System Optimization in Social Networks based on Cooperative Filter with Deep MVR Algorithm

Today, social networks have become very popular due to their high usage in communicating with each other. But this popularity requires the development of a backend to communicate with each other. Hence, a topic called identifying users is created by making recommendations or propositional systems, and so on, link prediction. The most important issue is the new users' social networks so that they can offer suggestions. In this research, we tried to provide a system of recommendations for introducing new users to previous users and vice versa based on the principles of machine learning. The proposed method is that the data is entered into the program and then the keywords are extracted from them. Then a sampling between the data is performed based on the Pearson and Cronbach method. In the process of extraction operations along with diminishing dimensions, selection and finally extraction of the best features is done using the cooperative filter which is named here based on deep learning- Modified Vector Rotational (MVR) algorithm and its operators. In the following, due to the lack of probabilistic and statistical training in Deep and Reinforcement Learning with a random structure that is used to select users and also to offer users concerning the tastes of the extracted, there is an optimization algorithm for MVR to consider the best features with training. In the following, a series of evaluation criteria have been used to ensure the proposed approach, indicating the appropriate results of the proposed method.