A New Hybrid Recommender System Integrating Density-Based Clustering and Neural Networks to Addressing the Cold Start Problem

Subject Areas : International Journal of Data Envelopment Analysis

Hanie Saleh Tabari ¹ , Mousa Nazari ^{2 *} , Mohammad M. AlyanNezhadi ³ , Hesamoddin Pourrostami ⁴ , Seyyed Mohammad R. Hashemi ⁵

1 - Department of Computer Science, University of Science and Technology of Mazandaran, Iran
2 - University of Science and Technology of Mazandaran: Behshahr, Mazandaran, IR
3 - گروه علوم کامپیوتر، دانشگاه علم و فناوری مازندران، بهشهر، ایران
4 - Department of Computer Science, University of Science and Technology of Mazandaran, Iran
5 - گروه مهندسی کامپیوتر و فناوری اطلاعات، دانشگاه صنعتی شاهرود، شاهرود، ایران.

Received: 2025-05-19 Accepted : 2025-07-05 Published : 2025-02-08

Keywords: Recommender Systems, Cold Start Problem, Density Clustering, Neural Network, Hybrid Filtering Techniques,

Abstract :

Recommender systems have become essential for filtering the vast amounts of online content available today, yet they often falter when confronted with the cold start problem: new users or items lack the historical data required for accurate suggestions. To overcome this, our study introduces a hybrid framework that fuses density-based clustering with a Multi-Layer Perceptron (MLP) neural network. Initially, density-based clustering groups users and items according to demographic attributes and content features, revealing latent structure even in sparse settings. Those cluster assignments, together with any available ratings, are then fed into the MLP, which learns complex, non-linear relationships to predict user preferences. We evaluated our approach on the MovieLens 1M dataset, comprising one million ratings from 6,040 users on 3,952 films, and compared it against classic collaborative filtering and content-based baselines. Across metrics such as RMSE, our method consistently produced lower error rates and higher top-k recommendation accuracy, with the most pronounced gains observed for users with fewer than five prior ratings or newly introduced movies. By leveraging both side information and rating patterns, this hybrid model delivers more personalized recommendations in data-sparse scenarios, ultimately improving user satisfaction and engagement on digital platforms.

References:

[1] Khusro, S., Ali, Z., & Ullah, I. (2016). Recommender systems: Issues, challenges, and research opportunities. In Information Science and Applications (ICISA) 2016 (pp. 1179–1189). Springer.
[2] Eirinaki, M., Gao, J., Varlamis, I., & Tserpes, K. (2018). Recommender systems for large-scale social networks: A review of challenges and solutions. Information Sciences, 78, 413–418. Elsevier.
[3] Peška, L., Tashu, T. M., & Horváth, T. (2019). Swarm intelligence techniques in recommender systems—A review of recent research. Swarm and Evolutionary Computation, 48, 201–219.
[4] Nilashi, M., Ibrahim, O., & Bagherifard, K. (2018). A recommender system based on collaborative filtering using ontology and dimensionality reduction techniques. Expert Systems with Applications, 92, 507–520.
[5] Pancholi, D., & Selvi, C. (2023). Study of cold-start product recommendations and its solutions. In International Conference on Data Management, Analytics & Innovation (pp. 45–58). Springer.
[6] Panteli, A., & Boutsinas, B. (2023). Addressing the cold-start problem in recommender systems based on frequent patterns. Algorithms, 16(4), 182.
[7] Tsikrika, T., Symeonidis, S., Gialampoukidis, I., Satsiou, A., Vrochidis, S., & Kompatsiaris, I. (2018). A hybrid recommendation system based on density-based clustering. In Internet Science: INSCI 2017 International Workshops, IFIN, DATA ECONOMY, DSI, and CONVERSATIONS (pp. 49–57). Springer.
[8] Hao, B., Yin, H., Zhang, J., Li, C., & Chen, H. (2023). A multi-strategy-based pre-training method for cold-start recommendation. ACM Transactions on Information Systems, 42, 1–24. https://doi.org/10.1002/int.23011
[9] Rahman, M. M., Sifat, F. F., Islam, R., Molla, S., & Khan, M. R. K. (2024). Hybrid recommendation systems using adaptive clustering to address cold-start problems. In 2024 International Conference on Electrical, Computer and Energy Technologies (ICECET) (pp. 1–6). IEEE.
[10] Tanwar, A., & Vishwakarma, D. K. (2023). A deep neural network-based hybrid recommender system with user-user networks. Multimedia Tools and Applications, 82(10), 15613–15633.
[11] Alizadeh, S. H. (2018). A hybrid recommender system using multi-layer perceptron neural network. In 2018 8th Conference of AI & Robotics and 10th RoboCup IranOpen International Symposium (IRANOPEN) (pp. 7–13). IEEE.
[12] Tahmasebi, F., Meghdadi, M., Ahmadian, S., & Valiallahi, K. (2021). A hybrid recommendation system based on profile expansion technique to alleviate cold start problem. Multimedia Tools and Applications, 80, 2339–2354.
[13] Şeref, B., Bostanci, G. E., & Güzel, M. S. (2021). Evolutionary neural networks for improving the prediction performance of recommender systems. Turkish Journal of Electrical Engineering and Computer Sciences, 29(1), 62–77.
[14] Duan, L., Wang, W., & Han, B. (2021). A hybrid recommendation system based on fuzzy c-means clustering and supervised learning. KSII Transactions on Internet and Information Systems (TIIS), 15(7), 2399–2413.
[15] Chen, J., Zhao, C., Uliji, & Chen, L. (2020). Collaborative filtering recommendation algorithm based on user correlation and evolutionary clustering. Complex & Intelligent Systems, 6, 147–156.
[16] Bahramian, Z., Abbaspour, R. A., & Claramunt, C. (2017). A cold start context-aware recommender system for tour planning using artificial neural network and case-based reasoning. Mobile Information Systems, 2017, Article 2017.
[17] Camacho, L. A. G., & Alves-Souza, S. N. (2018). Social network data to alleviate cold-start in recommender system: A systematic review. Information Processing & Management, 54(4), 529–544.
[18] Sedhain, S., Menon, A., Sanner, S., Xie, L., & Braziunas, D. (2017). Low-rank linear cold-start recommendation from social data. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 31, No. 1).
[19] Hernando, A., Bobadilla, J., Ortega, F., & Gutiérrez, A. (2017). A probabilistic model for recommending to new cold-start non-registered users. Information Sciences, 376, 216–232.