README

Name: Shouman Das
Email: shouman.das@rochester.edu
Course: CSC446
Homework: HW2
Implement perceptron for the adult income dataset using Python. Data available in /u/cs246/data/adult. Experiment with performance as a function of number of iterations.

************ Files *********

das_perceptron.py
README
das_perceptron.png

************ Algorithm *****

We implement perceptron algorithm on the adult income dataset.
repeat
	for n = 1...N do
		if t_n != y_n then
		w = w + y_n * x_n
		end if
	end for
until for all n t_n = y_n or maxiters

************ Instructions ***

The main algorithm is on das_perceptron.py which is executable with --nodev and --iterations argument. 
This script also experiments with different number of iterations. If --nodev and --iterations is not given we run the loop for 100 times and plot the training accuracy and dev accuracy. 

************ Results *******

Comparing our results to the previous homework, we see that perceptron gives us around 80% accuracy on the test data. we choose our final iteration number as 10. And report our training testing error for that.

************ Your interpretation **** (Attention: important!)

From this implementation we see that iterations and learning rate is highly important for SVM algorithm. 

************ References ************
1. Lecture Notes(https://www.cs.rochester.edu/~gildea/2018_Spring/notes.pdf)
2.  Book: Bishop, Christopher, Pattern Recognition and Machine Learning