The recommender systems are the popular personalization tool for helping users to find pertinent information based on preferences kept in individual profiles. A user profile plays an essential role in the success of the recommendation processes, so the recommender systems must design a profile to identify the user’s needs. The accuracy of the user profile affects the overall performance of the recommender system. Personalizing through creating a user profile is considered a challenge because people’s interests are changing over time. In this paper, a learning method is proposed that uses user feedback to improve the accuracy and precision of the recommendation list. This method utilizes learning automata to complete the user profile. The user preferences represent in the form of a weight vector. This vector is the action probability vector of the learning automata; it is updated according to user feedback. Experimental results based on Movie Lens 100k, Movie Lens 1M, and Netflix datasets show that the proposed approach is superior to existing alternatives.
