Enhancing Recommendation Performance via Stacking Ensemble and Synthetic Data Augmentation

Recommendation systems are essential tools that primarily aim to help users navigate through a large volume of information. They simplify the decision-making process by suggesting relevant items based on users’ historical behaviour. However, their performance is often affected by common problems such as data sparsity. This work proposes a stacking-based ensemble recommendation system that integrates multiple machine learning models to enhance the model’s predictive performance. A new synthetic data augmentation technique is introduced to address the sparsity issue in the user–item rating matrix. This method uses the Naïve Bayes algorithm to predict additional ratings for each user. These are then added to the original dataset to reduce the sparsity in the user–item rating matrix. Additionally, matrix factorisation models are included as base models to extract latent features. The results of the base models are fed as input to the stacking model to generate the final predictions. Experiments were performed on five benchmark datasets using MAE and RMSE as metrics to assess the performance. The results demonstrate that our proposed StackGBR-SDA model significantly improved the predictive performance compared to the individual models that made up the ensemble. Specifically, it achieved RMSE values of 0.1545, 0.2912, 0.8635, 0.4816, and 0.7163 on the datasets Amazon Food, Yelp, MovieLens100K, CiaoDVD, and FilmTrust, respectively. Moreover, our model outperformed methods from previous studies in terms of RMSE. These findings confirm the effectiveness of ensemble learning, as well as our data augmentation approach, in alleviating the sparsity problem and improving the recommendation performance.