Simple Linear Discriminant Analysis
Overview
Theory
A linear discriminant is a functon {% g(\vec{x}) %}
{% g(\vec{x}) = \vec{w}^T \vec{x} %}
sometimes the function includes a bias
{% g(\vec{x}) = \vec{w}^T \vec{x} + \vec{b} %}
The discriminant function assigns labels to points by computing a real number as above and assigning
a label based on a set of ranges. A common linear discriminant is to classify a binary problem
by assigning all points where the discriminant is positive a certain label, and all points less than
zero a different label. The binary classifier is effectively a
perceptron.
