Since we used a shallow autoencoder with two nodes in the hidden layer in the middle, it would be worth using the dimensionality reduction to explore our feature space. We can extract this hidden feature with the scoreDeepFeatures() method and plot it to show the reduced representation of the input data.
The scoreDeepFeatures() method scores an auto-encoded reconstruction on-the-fly, and materialize the deep features of given layer. It takes the following parameters, frame Original data (can contain response, will be ignored) and layer index of the hidden layer for which to extract the features. Finally, a frame containing the deep features is returned. Where number of columns is the hidden ...
