The impact of data augmentation on the performance of session-based recommender systems

Abstract

In response to the challenges related to the large volume of data that application users encounter while interacting with internet services, recommender systems have emerged as a valuable solution. The content provided by recommenders is tailored to the specific choices of each user. Despite the proposal of novel models in the existing literature, there is also an emergence of studies addressing alternative factors that may enhance recommendation performance. Data augmentation is one of the steps involved in data pre-processing that affects the final accuracy. The objective of this paper is to examine the most efficient methods of data augmentation on a range of session-based recommender systems. Furthermore, novel strategies for data enhancement are proposed and verified.