Altice Labs, in partnership with Polytechnic Institute of Viseu, contributed to a paper where mechanisms based on Artificial Intelligence are proposed for the construction of recommendation systems, i.e., mechanisms capable of analyzing user behavior and making content recommendations.
This paper, entitled “Weight-based Dynamic Hybrid Recommendation System for Web Application Content”, written by Margarida Jerónimo and Rui Pedro Duarte, from the Polytechnic Institute of Viseu, and Filipe Cabral Pinto, from Altice Labs’ Innovation and Technology Strategy department, was presented at ICICT 2022 – International Congress on Information and Communication Technology.
In more technical terms, this work presents a prototype of a web application content recommendation system applied to movie recommendations, having been proposed the combination of collaborative filtering techniques with content-based filtering. In content-based filtering, the TF-IDF technique was applied to allow, from the existing descriptions, the extraction of different topics associated with video contents. For collaborative filtering, the kNN algorithm was used to calculate the similarities between clients. The system was built on Django, a python-based web platform for the development and instantiation of web applications.