This is a project I developed for Fall 2017 QMSS4058 Data Mining at Columbia University.

I led a two-student team effort on optimizing a binary classification problem of kaggle data: https://www.kaggle.com/ludobenistant/hr-analytics-1.

We were interested in predicting whether or not an employee leaves prematurely. We employed 11 models including various regression and tree models as well as neural network and found the best prediction rate at 98.4% with random Forest.