Programming Machine Learning

Bending the Boundary

Let’s switch from the perceptron to the neural network. What does its decision boundary look like?

If you bother to try the neural network on the linearly separable dataset, you’ll find that it gets the same perfect accuracy as the perceptron, and a similarly straight decision boundary. On the second dataset, however, things get interesting.

Here’s the code that trains the neural network on the non-linearly separable dataset:

12_classifiers/plot_neural_network_boundary.py

import numpy as np
import neural_network as nn
x1, x2, y = np.loadtxt('non_linearly_separable.txt', skiprows=1, unpack=True)
X_train = X_test = np.column_stack((x1, x2))
Y_train_unencoded = Y_test = y.astype(int).reshape(-1, 1)