Song Recommendation System Using Machine Learning Algorithms and Data Analysis

Abstract

The present paper proposes a song recommendation system that employs machine learning methods to suggest various songs to users. By analyzing user searched song name, the system generates song recommendations similar to the user's searched song name. The proposed approach combines collaborative filtering, which recommends songs based on similar user searches, and content-based filtering, which suggests songs based on their attributes. Furthermore, a hybrid approach that integrates both techniques is utilized to generate a more comprehensive set of recommendations. The study evaluates the system's performance using precision and recall metrics, with the results indicating that the hybrid approach surpasses individual techniques. The proposed system can enhance the user experience by delivering appropriate recommendations that cater to their musical interests. In this project, we utilized a sample dataset of songs to establish relationships between users and songs. The objective was to recommend new songs to users based on their searched song. We utilized various libraries such as NumPy and Pandas to implement this project. Additionally, we employed Count Vectorizer in combination with Cosine similarity to analyze and measure the similarity between songs.