Logistic Regression on the Iris Dataset

The Iris dataset used in the literature on classification methods and widely used in statistics and machine learning. There are four features: sepal length, sepal width, petal length, and petal width. The data set contains 3 classes of 50 instances each, where each class refers to a type of iris plant: Iris Setosa, Iris Versicolour, or Iris Virginica.

The following is a visualization of the logistic regression model trained on the Iris dataset. A given input goes through 3 transformations:

1. Scaling: Normalizing the input to a mean of 0 and a standard deviation of 1, a.k.a, Z-score normalization
2. Computing Logits: We calculate the logits as a linear combination of all the scaled inputs
3. Softmax: The logits aren't useful by themselves, we can apply the softmax function to convert it to a probability distribution
4. Prediction: Our prediction is the class with the largest probability

Notice that there is a fifth extra feature which is always 1. This gives us a convenient way to encode the bias terms.