Implementation of Recommendation Algorithm based on Recommendation Sessions in E-commerce IT System

Abstract

The aim of the paper is to present a study as the implementation of the author’s Algorithm of the Recommendation Sessions ARS in an operating e‐commerce information system and to analyse basic parameters of the recommendation system created as a result of the implementation. The first part of the study contains a synthetic description of the area of recommendation systems. The next section presents the proprietary ARS recommendation algorithm based on recommendation sessions. The third part of the paper describes the mathematical model of the recommendation session built on the basis of the theory of graphs and networks, which such model makes the input data for the algorithm in question. The next part of the publication describes the possibilities of representing graph structures and the method of implementing a G graph (constituting a set of the recommendation session) in a relational database. The implementation of the ARS algorithm, based on the SQL standard, was also presented. The implementations in question have been developed on the basis of a working information system of the e‐commerce class. As a result of the implementation of the algorithm, a fully functional recommendation system was created, which can be adapted to various e‐commerce IT systems. The positive result of the work was confirmed by the research on the parameters of the recommendation system, included in the last part of the study.