DOIONLINE 
DOIONLINE Publish In |
International Journal of Advance Computational Engineering and Networking (IJACEN)-IJACEN |
 Journal Home Volume Issue |
||||||||
Issue |
Volume-7, Issue-8 ( Aug, 2019 ) | |||||||||
Paper Title |
Music Genre Classification on AWS Cloud Platform | |||||||||
Author Name |
Astha Baranwal, Priya.V | |||||||||
Affilition |
School of Information Technology and Engineering, VIT, Vellore, India | |||||||||
Pages |
10-15 | |||||||||
Abstract |
Music genre classification is a trending topic in the music industry. Genre classification is crucial to the music industry as it plays a vital role in song recommendation systems. This paper uses the AWS Cloud platform’s machine learning model to classify songs into genres (viz. rock, pop, blues, country, classical and hip-hop) based on the predicted scores. AWS Machine Learning models are fast and easy to build. Stochastic Gradient Descent is used for classification for the AWS Machine Learning multi-class model. The predicted scores for each genre for a song can help determine multi-labeled classification for a song as well. Real-time classification of songs is possible through the AWS Machine learning model created. Random forest model is built for the dataset using Python machine learning libraries. This model is then hyper parameter tuned and the performance metrics are calculated. F1-scores for the two models are then compared. Keywords – Machine learning, Multi-classification, Cloud Computing, Multi-label classification, Stochastic Gradient Descent, Random Forest, AWS Machine Learning | |||||||||
| View Paper | ||||||||||
