One of the hot topics in the current era is the use of recommender systems. A recommender system is a subclass of an information filtering system that seeks to predict a user's rank or preference in a particular case. Examples of such systems include movie streaming websites that offer movies to users based on their history, or an online shopping site that offers another product based on a user's previous purchase. In such systems, the computer uses the data it has already collected from its users to try to predict the vague opinions of users on specific topics and use this conjecture to persuade the user to use the platform more. The purpose of this project was to implement a recommender system and adjust its parameters to obtain the minimum error and compare its performance with other recommender systems. The algorithm used to implement this system in the project was matrix factorization algorithm. In this algorithm, two matrices P and Q are formed and attempts are made to adjust the indices of these two so that the product of the multiplication of these two matrices is close to the user-vote matrix. Once these two matrices are formed, it is easy to predict users' votes for movies they have not seen. After implementing this algorithm, its hyper parameters were tuned so that the output had the lowest error, and then this error was compared to the error of other methods such as RBM, which performs this operation using neural networks. The results of this study showed that in this implementation, the predicted output differs from the actual output by an average of about 14% and this error is accepted and appropriate for such systems. Also, this method performed better than the RBM method which had an error of about 17%.