Lecture Objective

The objective of this lecture is classification of categorical data. Use these interactive pages to understand how logistic regression and threshold-based decisions classify observations into categories.

Simple Sigmoid (Page 1)

Basic illustration of the function 1/(1+np.exp(-z)) with a single z control.

Simple Sigmoid

Confusion Matrix Practice (Page 3)

Practice threshold-based 2-class and 3-class predictions, then build confusion matrix and key metrics.

Confusion Matrix

Sigmoid Function

Interactive plot of p = 1 / (1 + exp(-(b0 + b1x))) with sliders and chart actions.

Sigmoid Function

Visualization for Cost Function

Compare two logistic-model cases and track gradient descent trajectory on the cost surface.

Cost Function