Development and research of models and software for the recommender system of consumer goods

Abstract

There have been proposed investigation of the problem of creating recommendations with technical description for building the Recommender System of consumer goods with help of modern algorithms, approaches, principles and contains the investigation of the most popular methods. It was defined, that the deployment of Recommender Systems is one of the rapidly developing areas for improving applied information technolog ies, tools for automatic generating offers service based on the investigation of the personal needs and profile of customers. It was investigated, that such systems have started to play a very important role in the fast growing Internet, as they help users to navigate in a large amount of information, because users are not able to analyze a large amount of information, because it is very difficult and takes a lot of time and effort, but due to such systems, namely Recommender Systems that are able to filter a large amount of information, and provide for users the information and recommendations their likes the problem can be solved and instead of providing the static information, when users search and, perhaps, buy products, Recommender Systems increase the degree of interactivity to expand the opportunities provided to the user. It was defined, that Recommendation systems form recommendations independently for each specific user based on past purchases and searches, and also on the basis of the behavior of other users with help of recommendation services, which collect different information about a person using several methods and at the same time all systems are shared. An overview of content-based, collaborative filtering and hybrid methods was performed. An overview of Alternating Least Squares and Singular Value Decomposition recommendation algorithms was performed. The design of the Recommender System of consumer goods software component was described. The main features of software implementation and programming tools for the system which is being developed were explained. The conclusions about the problems of Recommender Systems and the review of existing algorithms were made.