A hybrid recommender system to support UI/UX designers

A hybrid recommender system has been developed to support UI/UX designers, which is aimed at optimizing the search for relevant information and reducing the time for practical creation of web interfaces. The combination of collaborative filtering and content filtering methods can significantly increase the accuracy and relevance of recommendations for users. The system for clustering publications analyzes their keywords and the time to read the content, which helps to avoid the problem of "cold start". To create a hybrid recommender system, a complex methodology has been developed, which consists of the following stages: obtaining input data, pre-processing, clustering and building a classifier, the next step is building a knowledge base and obtaining recommendations, as well as obtaining data about users of the recommender system, forming a database, analysis user requests to prevent “cold start”. The applied clustering and classification methods, including machine learning, ensured the accuracy of the system's recommendations at the level of 97 %, which is confirmed by the low root mean square error (RMSE = 1.21) and high ratings from users. This helps to increase the efficiency of the work of designers by offering individual recommendations based on their requests.