Bishop's University

CS 566

1. Introduction
In this project, you will design three classifiers: a perceptron classifier, a large-margin (MIRA) classifier,
and a slightly modified perceptron classifier for behavioral cloning. You will test the first two classifiers
2
on a set of scanned handwritten digit images, and the last on sets of recorded pacman games from various
agents. Even with simple features, your classifiers will be able to do quite well on these tasks when given
enough training data. Optical character recognition (OCR) is the task of extracting text from image sources.
The data set on which you will run your classifiers is a collection of handwritten numerical digits (0-9).
This is a very commercially useful technology, similar to the technique used by the US post office to route
mail by zip codes. There are systems that can perform with over 99% classification accuracy (see LeNet-5
for an example system in action). Behavioral cloning is the task of learning to copy a behavior simply by
observing examples of that behavior. In this project, you will be using this idea to mimic various pacman
agents by using recorded games as training examples. Your agent will then run the classifier at each action
to try and determine which action would be taken by the observed agent. The code for this project includes
the following files and data, available as a zip file.
a. Dataset
data.zip: Data file, including the digit and face data.
b. Files you'll edit
naiveBayes.py: The location where you will write your naive Bayes classifier.
perceptron.py: The location where you will write your perceptron classifier.
mira.py: The location where you will write your MIRA classifier.
dataClassifier.py: The wrapper code that will call your classifiers. You will also write your enhanced feature
extractor here. You will also use this code to analyze the behavior of your classifier.
answers.py: Answer to Question 2 goes here.
c. Files you might want to look at
classificationMethod.py: Abstract super class for the classifiers you will write.
(You should read this file carefully to see how the infrastructure is set up.)