A GNN Model with Adaptive Weights for Session-Based Recommendation Systems

Begüm Özbay,Dr.Resul Tugay,Prof. Dr. Şule Gündüz Öğüdücü

2024-08-09

Abstract:Session-based recommendation systems aim to model users' interests based on their sequential interactions to predict the next item in an ongoing session. In this work, we present a novel approach that can be used in session-based recommendations (SBRs). Our goal is to enhance the prediction accuracy of an existing session-based recommendation model, the SR-GNN model, by introducing an adaptive weighting mechanism applied to the graph neural network (GNN) vectors. This mechanism is designed to incorporate various types of side information obtained through different methods during the study. Items are assigned varying degrees of importance within each session as a result of the weighting mechanism. We hypothesize that this adaptive weighting strategy will contribute to more accurate predictions and thus improve the overall performance of SBRs in different scenarios. The adaptive weighting strategy can be utilized to address the cold start problem in SBRs by dynamically adjusting the importance of items in each session, thus providing better recommendations in cold start situations, such as for new users or newly added items. Our experimental evaluations on the Dressipi dataset demonstrate the effectiveness of the proposed approach compared to traditional models in enhancing the user experience and highlighting its potential to optimize the recommendation results in real-world applications.

Information Retrieval,Artificial Intelligence

What problem does this paper attempt to address?

The paper aims to address the issue of prediction accuracy in Session-based Recommendation Systems (SBRs). Specifically, the authors propose a new method to improve existing Graph Neural Network (GNN)-based session recommendation models (SR-GNN) by introducing an adaptive weighting mechanism. The main objectives of this method include:

1. **Improving Prediction Accuracy**: By dynamically adjusting the importance of each item in the session, the accuracy of predicting the next item is enhanced.
2. **Addressing the Cold Start Problem**: In cases of new users or newly added items, dynamically adjusting the importance of items provides better recommendations in cold start scenarios.
3. **Enhancing Personalized Recommendations**: By integrating side information, a better understanding of user interests is achieved, leading to more personalized recommendations.

The researchers hypothesize that this adaptive weighting strategy can significantly enhance the performance of SBRs in various scenarios. This hypothesis is validated through experiments on the Dressipi dataset. The results show that, compared to traditional models, the proposed SR-GNN model with adaptive weights excels in improving user experience, particularly in handling real-time data and short-term preference changes.