Rastomi Pamungkas, Permata Permata, Rakhmat Dedi Gunawan, Adhie Thyo Priandika

Hybrid Music Recommendation System Using K-Means Clustering and Neural Collaborative Filtering for Spotify Playlist Personalization

Introduction

Hybrid music recommendation system using k-means clustering and neural collaborative filtering for spotify playlist personalization. Enhance Spotify playlist personalization with a hybrid music recommendation system. Integrating K-Means clustering and Neural Collaborative Filtering, this study boosts accuracy and addresses cold start.

Abstract

Personalizing music recommendations has become a significant challenge on music streaming platforms such as Spotify due to the vast number of available songs and the limitations of conventional recommendation systems in accurately capturing user preferences. In addition, traditional single-method recommendation approaches often face the cold start problem, which reduces the effectiveness of generated recommendations. Therefore, this study aims to develop and evaluate a hybrid recommendation system that integrates the K-Means Clustering algorithm and Deep Collaborative Filtering based on Neural Matrix Factorization to improve the relevance of music playlist recommendations. The dataset used in this study consists of more than 15,151 Spotify songs obtained from the Spotify dataset available on Kaggle. The dataset was processed through several stages including data inspection, data cleaning, feature selection, and standardization. Audio features used in the analysis include danceability, energy, acousticness, instrumentalness, valence, tempo, and duration. The optimal number of clusters was determined using the Elbow Method and Silhouette Score, resulting in five clusters with a relatively balanced data distribution. The clustering results were then used as the basis for Cluster-Based Filtering to narrow the search space of candidate songs before being processed by the Neural Matrix Factorization model. Performance evaluation was conducted using Hit Ratio at rank 10 and Normalized Discounted Cumulative Gain at rank 10. The proposed model achieved values of 0.1110 and 0.0507, respectively, indicating that the integration of clustering and deep collaborative filtering can improve the effectiveness and personalization of music recommendation systems. This study contributes by proposing a hybrid recommendation framework that integrates clustering-based item grouping with deep collaborative filtering to improve recommendation efficiency and playlist personalization in large-scale music streaming platforms.

Review

This paper addresses the significant challenge of personalizing music recommendations on platforms like Spotify, acknowledging the limitations of traditional single-method systems and their susceptibility to the cold start problem. To overcome these issues, the authors propose a novel hybrid music recommendation system. This innovative approach integrates K-Means Clustering with Deep Collaborative Filtering, specifically using Neural Matrix Factorization, with the explicit goal of enhancing the relevance and personalization of music playlist recommendations for a more refined user experience. The methodology outlines a comprehensive approach, utilizing a dataset of over 15,000 Spotify songs, which underwent rigorous preprocessing including feature selection focusing on audio attributes such as danceability, energy, and acousticness. K-Means Clustering was applied, with the optimal number of five clusters determined using the Elbow Method and Silhouette Score, ensuring a relatively balanced data distribution. Crucially, these clustering results were then employed for Cluster-Based Filtering, effectively narrowing the search space for candidate songs before they were fed into the Neural Matrix Factorization model. Performance evaluation, conducted using Hit Ratio at rank 10 and Normalized Discounted Cumulative Gain at rank 10, yielded promising results of 0.1110 and 0.0507, respectively, indicating the effectiveness of the proposed hybrid model in improving recommendation quality. This study presents a valuable contribution to the field of music recommendation systems by demonstrating a practical and effective hybrid framework. The integration of clustering for initial item grouping significantly improves efficiency by reducing the computational load for the deep learning model, while simultaneously enhancing personalization. The findings suggest that combining content-based filtering (via clustering of audio features) with collaborative filtering (via neural networks) provides a robust solution for large-scale music streaming platforms. While the absolute metric values might appear modest in isolation, their positive indication of improved effectiveness and personalization, especially within a complex hybrid context, underscores the potential of this architecture for addressing critical recommendation challenges.

Full Text

Comments

You need to be logged in to post a comment.