RECOMMENDER SYSTEM BASED ON GRAPH NEURAL NETWORKS
DOI:
https://doi.org/10.24867/25BE15NjegomirKeywords:
graph neural networks, recommender systems, GraphSAGEAbstract
This study aimed to create a machine-learning model based on graph neural networks that can provide item recommendations for users of an e-commerce system. For model training, the data had to be converted to a heterogenous bipartite graph, where users and products were represented as nodes, while edges represented the interactions between them. Models were tested on the Brazilian E-commerce dataset. The first model used one encoder containing graph neural network layers for encoding nodes and one decoder for edges, which aimed to make rating predictions that the user would assign to the product. The second model used separate encoders for users and products and had an edge decoder. The model with two encoders achieved better RMSE than the single encoder model and had a more stable training process.
References
[2] Koren, Y., 2009. The bellkor solution to the netflix grand prize. Netflix prize documentation, 81(2009), pp.1-10.
[3] Koren, Y., Bell, R. and Volinsky, C., 2009. Matrix factorization techniques for recommender systems. Computer, 42(8), pp.30-37.
[4] Scarselli, F., Gori, M., Tsoi, A.C., Hagenbuchner, M. and Monfardini, G., 2008. The graph neural network model. IEEE transactions on neural networks, 20(1), pp.61-80.
[5] Xu, K., Hu, W., Leskovec, J. and Jegelka, S., 2018. How powerful are graph neural networks?. arXiv preprint arXiv:1810.00826.
[6] Olist and André Sionek, 2018, Brazilian E-Commerce Public Dataset by Olist Data set. Kaggle. https://doi.org/10.34740/KAGGLE/DSV/195341.
[7] Ying, R., He, R., Chen, K., Eksombatchai, P., Hamilton, W.L. and Leskovec, J., 2018, July. Graph convolutional neural networks for web-scale recommender systems. In Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining (pp. 974-983).
[8] Wang, X., He, X., Wang, M., Feng, F. and Chua, T.S., 2019, July. Neural graph collaborative filtering. In Proceedings of the 42nd international ACM SIGIR conference on Research and development in Information Retrieval (pp. 165-174).
[9] Wang, X., Ji, H., Shi, C., Wang, B., Ye, Y., Cui, P. and Yu, P.S., 2019, May. Heterogeneous graph attention network. In The world wide web conference (pp. 2022-2032).
