Inputs

The inputs section is highlighted below:

This section controls what features you want to use to perform regression or classification of the data where each feature is an input neuron in the input layer. Therefore, you are configuring the input layer in this section.

The data consist of two dimensional data, where X1 is the horizontal axis and X2 is the vertical. There are a variety of features to choose from. Simply toggle which features you want to use. Each feature is accompanied by a preview image of the output over a two-dimensional grid of coordinates if you were to choose that feature that are right beside the feature name itself. Values of white in each preview box are the classification boundary, meaning that this is where the output is close to 0. Features that are selected will have a gray shadow surrounding that preview box once you bring focus away from that box.

The available features are the following:

• X1: The feature X1 itself (i.e. the first dimension of the data)
• X2: The feature X2 itself (i.e. the second dimension of the data)
• X1^2: The feature X1 but squared
• X2^2: The feature X2 but squared
• X1*X2: Both features of X1 and X2 multiplied together
• sin(X1): The feature X1 with the sin operation applied to it
• sin(X2): The feature X2 with the sin operation applied to it