Naive Bayes Classifier

for the MNIST handwritten digit dataset

How to run the code

python dataClassifier.py 

Tested to have an accuracy of 74.0% on validation set & 75.0% on test set on default settings. By default smoothing_constant is 2, training set size is 1000 and test set size is 100.

python dataClassifier.py -k [smoothing_constant] -t [training_set_size] -x [test_set_size]

Tested to have an accuracy of 64.1% on validation set & 64.3% on test set (k = 2, t = 10000, x = 1000)

skeleton was taken from source