# Extreme Point Machine (EPM): 2-D 2-Class Version

by Ahmad Masadeh and Mohamad Hassoun (May 1998)

###
This (restricted complexity) version of our EPM employs a polynomial preprocessing (mapping) layer or a mapping layer of perceptrons with random Gaussian weights. Training the output perceptron
weights is accomplished using the AHK learning rule (click here
for details).