Check out the colab notebook.
- Why waveplots aren't used to understand audio and its features.
- What is special about spectrograms and why are they used for classfication of audio data.
In this project we are trying to classify the sentences as loaded and non-loaded first. From this classification we get an understanding about which sentences spoken were important or were impactful for sentiment analysis. Doing this will reduce the amount of sentences that we would have to go through during training and evalution process. This should eventually increase the efficieny of the whole process.