Lecture2-1Perceptron Hypothesis Set

Credit Approval Revisited

A simple hypothesis set: Perceptron

For x=(x1,x2,...,xd), ‘features of customer’, compute a weight ‘score’ and

- approve credit if ∑di=1wixi>threshold

- deny credit if ∑di=1wixi<threshold

:{+1good,−1bad}, 0 ignored -linear formula h∈ are

h(x)=sign((∑di=1wixi)−threshold) are called perceptron

Vector Form of Perceptron

h(x)=sign((∑i=1dwixi)−threshold)=sign((∑i=1dwixi)−threshold×(+1))=sign(∑i=1dwixi)=sign(wTx)

Perceptrons in ℝ2

Perceptron⟺Linear Classifiers