Binary Classification using scikit-learn

We took up a binary classification task and used scikit-learn to achieve it. We extensively studied the tradeoffs between performance, accuracy and the level of preprocessing required for the following algorithms (and ensemble methods).

• Decision Tree
• SVM (with linear and RBF kernels)
• k-Nearest Neighbors
• Naive-Bayes
• Random Forest
• Adaboost

A technical report for the project detailing the results can be found here. The report includes precision-recall plots for the aforementioned learning algorithms. All of the plots can be found under plots/

Team Members

• Adithya Bhat - responsible for pre-processing and feature extraction
• Srinivasan Ravichandran - responsible for implementing scikit-learn machine learning algorithms

References:

• Dataset
• scikit-learn