3. Polynomial and Higher order Features:

Let us use polynomial features with the Perceptron. Consider the dataset shown below. [Hint: The dataset is not separable]. Note that this dataset consists of 2-dimensional points x = [x1, x2]

Part 1: Write down the perceptron loss function with quadratic features. First write down what will be the features, the dimensionality of the expanded (quadratic) feature set and the loss function. Is the dataset linearly separable with quadratic features?

Part 2: Draw approximately the output of the perceptron algorithm on this dataset.