Skip to content

Latest commit

 

History

History
10 lines (6 loc) · 376 Bytes

README.md

File metadata and controls

10 lines (6 loc) · 376 Bytes

audio-emotion-classification

In this project we classify negative audio emotions , data is combined between ( RAVDESS/ SAVEE/TESS and CREMA-D)

After extracting features we perform PCA to reduce dimensions)

4 models were tested : CNN, LSTM, KNN and SVM

Best results were obtained with SVM

You can find the pretrained model in file 'finalized_model.sav' ( SVM version)