Software for the implementation of an intelligent system to solve the problem of “cold start”

Abstract

As a result of the research, one of the approaches to building an intelligent information system based on the recommendation of products to users with a solution to the cold start problem is described and modeled. The conducted research takes into account the advantages and disadvantages of the methods, as well as their compatibility, when combining them, which is an important factor for the speed of the system and the efficiency of the algorithm. The implementation of the hybrid method for the construction of an intelligent information system, as well as its performance testing in comparison with the classical k-means algorithm, was carried out. Based on the received analysis, a practical comparison of the efficiency of the system with the basic approach to solving the problem and the hybrid one was carried out.