Week 1: Data Preprocessing and Exploration
Week 2: Model with Classical ML Algorithms
Week 3: Model with Neural Networks
- Exploration of both tabular and audio data.
- Performing thorough preprocessing on audio data, including cleaning, feature engineering, and feature extraction.
- For Ugrads: use EMODB and merge with RAVDESS; For Grads: Merge any of the other datasets or create your own emotional voice data (using your own voice recording) and merge that with RAVDESS dataset.
- Analyze the data in terms of gender balance and emotional category balance, show in a graph or visual representation and comment 1-2 sentences on the possible impact of them both
- Clone the jupyter notebook into your repository on github.
- Perform detailed analysis on the combined dataset. [Feature analysis and cleaning, write about the differences between RAVDESS and your dataset]
- Commit the changes in your repository.
Write a 1–2-page detailed report on the analysis that you have performed. You can include issues you encountered, how did you solve problems. Upload your report on canvas in PDF format (along with Jupyter Notebook file). [between 500-1000 word limit] Cite all your resources, including the datasets as references (outside of page/word limit .