Implemented a Logistic Regression classifier (from scratch), K-Fold Validation, and Feature Engineering Pipeline in Python. Achieved accuracies of 93% & 80% on 2 distinct test sets. (Non-Kaggle)
Implemented a Bernoulli Naïve Bayes’ classifier (from scratch) with Binary Feature Encoding, TFIDF Normalization, and 𝑿𝟐 Feature Selection in Python for Reddit Posts Classification. Achieved a 91.7% accuracy in the Kaggle competition. https://www.kaggle.com/c/reddit-classification/leaderboard
Implemented a VGG-16 Convolution Neural Network (from scratch) with Image Augmentation, Pixel Normalization, and Dropout Regularization in Pytorch for a customized MNIST Fashion Dataset. Achieved a 96% accuracy in the Kaggle competition. https://www.kaggle.com/c/ecse-551-f20-image-understanding/leaderboard