Trang Nguyen
Machine Learning
Homework 2

Files:
README
Nguyen_perceptron.py

Experiment:
- After running the program with different number of iterations provided that --nodev (or train data) or dev data is being used, the performance fluctuates a lot.
- Hence, we can clearly see that the performance is being overshoot.
- If we decrease the learning rate, the performance will still fluctuate for train and dev data but has a better accuracy.

How to run:
./Nguyen_perceptron.py --nodev --iterations (number) --lr (number)