The fully connected part takes the deep representation from either the RNN or CNN and transforms it into the final output classes or class scores. This component is comprised of fully connected layers along with batch normalization and optionally dropout layers for regularization.

We have now seen the general meta-architecture of a deep learning text classifier. In the following sections, we will dive into hands-on examples of text classifiers, incorporating this architecture as we go. Specifically, we will look at identifying spam in YouTube video reviews and classifying news articles.