Python Deep Learning Projects

Feature engineering

So, before we feed the raw audio data into our model, we need to transform the data into numerical representations that are features. In this section, we will explore various techniques to extract features from the speech data that we can use to feed into the model. The accuracy and performance of the model vary based on the type of features we use. As an inquisitive deep-learning engineer, it's your opportunity to explore and learn the features with these techniques and use the best one for the use case at hand.

The following table gives us a list of techniques and their properties:

Techniques	Properties
Principal component analysis (PCA)	Eigenvector-based method Non-linear feature extraction method Supported to ...