Firstly I have implemented a Naïve Bayes Classifier and evaluated its performance on the Large Movie review dataset, for different values of the hyper-parameter alpha. Below are my results Hyper-Parameter (ALPHA) Train Accuracy (%) Test Accuracy (%) 0.1 97.768 81.132 0.5 95.876 82.148 1.0 94.512 82.284 5.0 90.665 82.772 10.0 88.912 82.816
Nextly, I have implemented the Perceptron and Averaged Perceptron Algorithm and evaluated its performance on the dataset. I have evaluations for different values of the hyper-parameter(number of iterations) Below are the results for the Perceptron and Average Perceptron Iterations Perceptron Accuracy (%) Averaged Perceptron Accuracy (%) Train Test Train Test 1 75.04 74.23 85.92 83.84 10 95.04 86.11 94.16 87.2 50 99.7 86.2 98.9 87.1 100 99.9 86.2 99.8 86.8 1000 100 86.1 99.89 86.77