Predict Employee Retention

December 31, 2016
By using open Human Resource dataset hosted on kaggle here we have created a simple yet powerful ML model to predict employee churn.
Predicting Employee Retention is one of the hottest problems that machine learning models are solving these days. Recently kaggle published an open dataset for Human Resource Analytics. This has IN and OUT status of about 15,000 employees. Definitely this is a good stuff for any Data Scientist to experiment with. This dataset has following features. - Employee satisfaction level - Last evaluation - Number of projects - Average monthly hours - Time spent at the company - Whether they have had a work accident - Whether they have had a promotion in the last 5 years - Sales - Salary - Whether the employee has left It is easy to create a machine learning model in Microsoft Azure ML with less efforts and in short time. Azure Machine Learning simplifies and streamlines the process from building to deploying a predictive model in production. Azure Machine Learning offers free-usage tier, you can create your free account to experiment with it. Simple and Powerful Two-Class Decision Forest Model This is a fairly simple model that takes HR data in uploaded csv file and trains a Two-Class Decision Forest Classifier. ![enter image description here][1] Dataset was divided into two parts (train = 80%, test = 20%) and model was trained on 80% dataset (training dataset) and evaluated on 20% (test dataset). Following is the accuracy metrics. ![enter image description here][2] With **97% Accuracy** model is performing extremely well. ![enter image description here][3] **Feature Importance** If you have downloaded the experiment in your own azure ml account then can run the experiment, click on the output port of Permutation Feature Importance, and select visualize to inspect the output results of the module. The following figure shows the list of features sorted in descending order of their permutation importance scores ![enter image description here][4] From this we can see 'satisfaction level' and 'number of projects' are two major factors along with others in ranking. Share this article with your BOSS to tell her that keep your employees satisfied and engaged in projects to retain them ;) Afroz works as a Data Scientist at ACTS Inc, Tel: (904) 317-2140 Fax: (904) 317-2139 Email:​​ Web: [1]: [2]: [3]: [4]: