Binary Recommender System with Artificial Intelligence Aids

Abstract

An approach to the development of a binary recommender system based on intelligent methods and their evaluation based on quality metrics is proposed. The dynamics of the popularity of recommendation systems is analyzed. Binary classification models have been developed using neural networks, support vector machines and random forest. An assessment of the modeling stage is carried out using calculations of such quality metrics as accuracy, precision, recall, F score. The most effective model for building recommendations based on an experimental data set from the education sector has been determined. The prospects for the further development of the proposed approach in the context of incomplete data are identified.