Song Hit prediction model based on various acoustic and lyrical features. ML Course Project Monsoon 2021.
Tested on various Machine Learning classifers including :
- Logistic Regression
- Support Vector Machine
- Decision Trees
- Random Forest
Apart from the above, Artificial Neural Networks (ANNs) were also tested as a classifier and gave really good results with 2 hidden layers.
Lyrics were extracted from Google APIs. Some analysis were done on them too :
- Topic modelling (LDA)
- K-Means Clustering (Along with other features)
- Sentiment Analysis
All the results are documented in the models/
folder jupyter notebooks for reference. All the Reports and PPTs submitted for the project have been added to the reports/
folder. The weights and plots can be found in the weights/
and plots/
folders respectively. The methods used for extracting the multiple datasets have been stored in data_extraction/
.
Team Number - 6
Team Members :
- Ansh Arora 2019022
- Jishnu Raj Prashar 2019048
- Nandika Jain 2019064
- Tushar Mohan 2019393