ssen7 / sys6018-final-project

sys6018-final-project

SYS 6018 Final Project: Music Genre Prediction System using Ensemble models

1. audio_analysis/Audio EDA v0.1ss.ipynb: The code contains:

• Exploratory audio data analysis

2. audio_analysis/Audio Models+Ensemble v0.1ss.ipynb:

• Audio Modeling - QDA, Logistic, SVM

• Ensemble models

3. Text Modeling (5000 words, 1000 word) - text_analysis/final_text_analysis_Saurav.ipynb The code contains:

• Text Modeling - Random Forest, Gradient Boosting, Multilayer Perceptron

• Ensemble Model

4. Text EDA - data_eda/word_cloud.ipynb

5. Tree_MLP_Oversampling Analysis.ipynb: The code contains:

• Boosting, random forest, knn(No PCA), MLP, Oversampling for Audio

6. Ensemble Boosting.ipynb: The code contains:

• Boosting for ensemble

7. LyricBagofWordsClassifier.ipynb This code contains:

• Text modeling (500 word) - Random Forest, SVM, KNN, Logistic Regression

Datasets Used:

1. Genre Annotations:

• http://www.tagtraum.com/msd_genre_datasets.html (File name: msd_tagtraum_cd2c.cls.zip)

2. Audio Dataset:

• https://labrosa.ee.columbia.edu/millionsong/tasteprofile

3. Lyrics Dataset:

• https://labrosa.ee.columbia.edu/millionsong/musixmatch