Downloads: 84 | Views: 211
Research Paper | Software Engineering | India | Volume 5 Issue 8, August 2016
Attrition Prediction Using Machine Learning to Help in Astute Decision
Reshad Abdullah | Sachin Bojewar
Abstract: Industries, especially IT (Information Technology) today, are experiencing high employee attrition rate. The employee leaving voluntarily is not good for organization or to project in which they are working. Hence HR and senior managers and the policy makers of any industry are working together to reduce this voluntary exit. A good leader senses and understands employee needs and work with them and HR to fix the issues. However not all attrition causes are known to managers and when it actually happens it turns out as a surprise, then they are not able to do much. Some amount of attrition is certain and bound to happen like employee retiring or death of employee hence the scope of this work is only restricted to voluntary exit [1]. Organization and HR department has felt that if they would have known earlier, or they could have picked the sign of exit, they might have prevented good employees leaving. With vast amount of historical data available within the organization, and through analytics & machine learning it is possible to predict attrition. These tools not only predict but also show some clear pattern in attrition. Many organizations today uses cots attrition prediction tool or build their own in-house prediction tool. The scope of this work is implementation of my theory paper published Attrition Prediction- Need of the Hour for Companies [1]. A tool is developed to predict attrition and it also predicts reason for attrition, this tool is based on decision tree algorithm and developed in R language. The factor or reasons for attrition are then effectively used by managers and HR department to design a retention strategy for the employee or proactively find his replacement. At the same time management becomes aware of the situation and are in position to predict how much new backup recruitment can be done in future.
Keywords: Attrition, COTS, C45, C50, ID3, exp, Model, Machine learning, Notice period, Retention, r_dt10
Edition: Volume 5 Issue 8, August 2016,
Pages: 1366 - 1370