Research Article
Enhancing Music Recommendation Accuracy with Hybrid Autoencoder, CNN and RNN Models
@INPROCEEDINGS{10.4108/eai.28-4-2025.2358010, author={Anitha Surada and Yarra Veera Venkata Lakshmi and Neerukonda Sahithi and Dadi Bindu Tanya and Koppisetti Chandu Sai Venkata Ganesh and Appala Anuradha}, title={Enhancing Music Recommendation Accuracy with Hybrid Autoencoder, CNN and RNN Models}, proceedings={Proceedings of the 4th International Conference on Information Technology, Civil Innovation, Science, and Management, ICITSM 2025, 28-29 April 2025, Tiruchengode, Tamil Nadu, India, Part II}, publisher={EAI}, proceedings_a={ICITSM PART II}, year={2025}, month={10}, keywords={hybrid model matrix factorization neural collaborative filtering multi-layer perceptron recommendation system accuracy dataset personalization}, doi={10.4108/eai.28-4-2025.2358010} }- Anitha Surada
Yarra Veera Venkata Lakshmi
Neerukonda Sahithi
Dadi Bindu Tanya
Koppisetti Chandu Sai Venkata Ganesh
Appala Anuradha
Year: 2025
Enhancing Music Recommendation Accuracy with Hybrid Autoencoder, CNN and RNN Models
ICITSM PART II
EAI
DOI: 10.4108/eai.28-4-2025.2358010
Abstract
A hybrid model for user-item recommendation systems is proposed, combining the strengths of Matrix Factorization (MF) and Neural Collaborative Filtering (NCF). By integrating collaborative filtering techniques with deep learning models, specifically using Multi-Layer Perceptron (MLP) fusion, the model captures complex non-linear relationships between users and items. The objective is to improve recommendation accuracy by leveraging both linear and non-linear interactions present in user-item data. The model is evaluated using well- known datasets, including the Million Song Dataset and the Last.fm dataset. With an accuracy of 88%, a Mean Absolute Error (MAE) of 0.30, and a Root Mean Square Error (RMSE) of 0.90, experimental data show that the hybrid model works better than conventional techniques these results validate the model’s effectiveness in providing personalized and accurate recommendations, making it suitable for large-scale recommendation systems.
