Distributed representation of items in time-dependent recommender systems

• Grant number: 20/09354-3
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: September 01, 2020
• End date: August 31, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Tiago Agostinho de Almeida
• Grantee: Amanda Carnio Pascon
• Host Institution: Centro de Ciências em Gestão e Tecnologia (CCGT). Universidade Federal de São Carlos (UFSCAR). Campus de Sorocaba. Sorocaba , SP, Brazil

Abstract

With the high amount of information and ease of access, finding relevant information is a task that is becoming increasingly difficult. From this problem, recommender systems emerged, which analyze users' previous behavior to issue personalized item recommendations. Currently, most companies have some form of recommendation in their channels or services. Even so, many problems in the area have not been properly solved, such as the high dimensionality and sparseness that current recommendation problems have. Additionally, although promising results have been reported in studies that made use of temporal information to enrich the recommender algorithms, little attention has been paid to the subject. Based on the aforementioned scenario, this project presents temporal variations for recent methods in the area of recommendation that address the dimensionality and sparseness problems. Its main goal is to yield personalized recommendations through low-dimensional distributed representations of items that considers temporal information during the training phase.