USER FEEDBACK RATING COMPUTATION BASED ON A HYBRID RECOMMENDER SYSTEM

artificial intelligence

The exponential growth of information offered on the Internet leads to an exponential increase in products. Therefore, it does not make sense to display all products to the user through the site, so it needs to provide useful things that match the interests of this user. Thus, there is a need for a tool to filter the product according to the interests of some users, and the recommendation system provides these services to users, so the importance of optimizing the recommendation system is necessary to provide recommendations that are more suitable for their preferences. User profiles are commonly used to predict product ratings which are not taken into account. This system also faces several challenges or problems related to large data, failure to collect complete information about users, or adding a movie or a new user to the system. In this thesis, a hybrid system consisting of collaborative filtering and content-based filtering will be implemented in order to recommend products to different users, where the matrix factorization technique will be implemented to divide the data into two matrices in order to solve the problem of data scalability as well as solve the problem of data sparsity where according to previous studies the solutions were This problem is solved by implementing a clustering where the data is divided into several groups and each group is dealt with separately. Either to solve the cold start problem, the document content will be used, and the film type will be used.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.