Naive Bayes and Perceptron implementation in Python3.6 to classify faces and digits represented in text files.
The Naive Bayes tests and Perceptron tests can be run using python bayesAlgo.py and python bayesAlgo.py respectfully.
For each algorithm, the algorithm will use predifined training and validation/testing data sets to train the algorithms and then test prediction accuracy, with 5 trials each using training sample sizes from 10%-100%.
Defines a class to create a bayes classifier given training/test image files, and thier metadata (size). Interesting variables to experiment with include the Laplace smoothing constant, k.
Defines a class to create a perceptron classifier given training/test image files, and thier metadata (size/feature enumeration definitions). Interesting variables to experiment with include the weight incrementing threshold constant.