Simple Naive Bayes implementation from scratch on UIUC TA evaluation dataset (http://archive.ics.uci.edu/ml/datasets/Teaching+Assistant+Evaluation). Does not use pandas. Implements both Gaussian and Empirical methods. uses add one laplace smoothing for empirical implementation. 5 fold cross validation is als performed on data and results analyzed. train-test split tweaked and accuracy plotted. Note the paths in main.csv are hard-coded and should be changed.