Music Sentiment Analysis

Description:

Using a dataset of annotated music examples and extracted static features, we attempt to use machine learning techniques to predict emotion classifications in the valence arousal space. Firstly, we replicate results obtained by Professor Renato Panda by using the abovementioned static features evaluated using a support vector machine. Then we design both a convolutional neural network and a bidirectional long-short term memory network trained using mel-spectrograms and raw audio as input respectively. Finally, we compare our results with Panda’s and offer next steps to improve them even further.