Towards the development of a recommender system for product de-livery using graph databases and related algorithms

Authors

  • Zaiker Nassima ETIS Laboratory, cy-tech Engineering School, CY-CERGY University in Paris, France
  • Lamghari Zineb LRIT Associated Unit to CNRST (URAC 29), Faculty of Sciences, Rabat IT Center, Mohammed V University in Rabat, Morocco

DOI:

https://doi.org/10.6977/IJoSI.202206_7(2).0004

Keywords:

Recommendation system, process model, Hybrid filtering, Graph database, Neo4j, Cypher, Python

Abstract

Recommendation systems are among the promising strands of machine learning that have revolutionized information retrieval operations. These systems are designed to make recommendations to users based on different factors. The realization of a recommender system requires a study of the users’ needs and the metrics that may influence each recommendation, as well as the attributes that can be entered into the application but that have no effect on the system's functioning. In this context, SoftCentre1 aims to develop a delivery recommender system using graph databases and related algorithms, in order to figure out the best path for each delivery to its destination. In this context, the deliverer will respect deadlines, specifications, and deduce the best itinerary to travel on. Therefore, our project revolves around the design, modeling, and implementation of a recommendation system based on these main phases: 1)Data collection and preprocessing, 2)Graph database creation, and 3)Applying recommendation and optimization algorithms.

Published

2022-06-30