ML-Algorithm-Comparison

This project serves as a replication of the work done in the Caruana and Niculescu-Milzi (CNM06) paper written in 2006 by comparing three different supervised machine learning algorithms across four different datasets. The algorithms utilized are logistic regression, SVM, and KNN, and the datasets were collected from the UC Irvine repository of machine learning datasets. These three algorithms were repeated per each dataset in order to find which one performed better in comparison to the others. Each dataset underwent one-hot-encoding in order to clean up the data and were put through hyperparameter tuning per each algorithm. The results seem to mimic the findings of CNM06 as SVM led the way with the best performance followed by KNN and Logistic Regression.