Personalized Music Recommendation System based on Streamer Streaming Trends

Abstract

This distinctive research paper investigates a fresh slant to process personalized music recommendation to a streamer based on the customized streaming trends focusing on today's music streaming services. Music Recommendation Systems (MRS) are one of the sole criteria that gives streaming application such as Spotify, Apple Music and other current major streaming companies a better edge over others. Thus, research into building more efficient MRS is majorly invested upon. The dependency on the streamer and item-based activity is the sole idea behind personalized recommendations. The study is focused on initiating algorithms based on Machine Learning (ML) models & Data Sciences (DS). The model has been worked so as to choose the effective model when required. Major models that the research has focused on is Genre Classification using metrics, Content-Based filtering, Collaborative filtering and the major new research ideology based on Forgetting curve Theory (FCT) that focuses on finding a correlation between model the concept that hypothesizes the decline of memory retention in time with music retention capabilities. Our Dataset is much or less based and augmented from the Spotify released Million Track Dataset[1] that has been published and licensed to Kaggle for research purpose.