Employee Attrition Prediction using Machine Learning

Data mining plays an important role in the internal Human Resource management processes of any firm [1]. These processes include prevention and prediction of employee attrition. This paper presents a machine learning based approach to attrition prediction for individual employees, by training different machine learning models on attrition data. In this paper, machine learning models like the logistic regression model, the linear regression model, the decision tree model, the random forest model, K nearest neighbors model, radius nearest neighbors model, the naïve bayes classifier model and Bayesian ridge model are trained on data obtained from Kaggle, an online community platform for data scientists and machine learning enthusiasts. The resulting models then predict whether an employee with particular attributes will attrit, and if so, within how many years they can be expected to do so. Prior to training the models, the data is cleaned (outlier removal, feature removal), scaled, and categorical variables are converted to numerical ones. Then, predictions are carried out and feature importances are found using random forest and decision tree models. The utilities of these models are then compared against each other based on accuracy, precision, recall, f-beta score, kappa score and three self made metrics. The results show that, surprisingly, the nearest neighbor models outperformed all others by a large margin (possibly due to data scaling being a part of preprocessing), and the logistic regression model was unable to predict attrition very satisfactorily. The results showed that satisfaction_level, average_monthly_hours and last_evaluation_rating are the most important features when predicting attrition or years. The research also shows that it is viable to use traditional ML models to predict the time in which an employee will attrit, and using the methodology defined in this research on a real dataset will provide useful information to the corporation applying it.