Kurniawan, Kevin and Erwin, Alva and Kho, I Eng (2017) A Study On Data Mining Algorithms For Churn Prediction In An Organization. Bachelor thesis, Swiss German University.

Preview	Text Kevin Kurniawan 11302010 TOC.pdf Download (251kB) | Preview
	Text Kevin Kurniawan 11302010 1.pdf Restricted to Registered users only Download (228kB)
	Text Kevin Kurniawan 11302010 2.pdf Restricted to Registered users only Download (808kB)
	Text Kevin Kurniawan 11302010 3.pdf Restricted to Registered users only Download (663kB)
	Text Kevin Kurniawan 11302010 4.pdf Restricted to Registered users only Download (657kB)
	Text Kevin Kurniawan 11302010 5.pdf Restricted to Registered users only Download (185kB)
Preview	Text Kevin Kurniawan 11302010 Ref.pdf Download (228kB) | Preview

Abstract

Employee churn and customer churn are two common problems for an organization as it indicates organization loss of resources and profits. Hence, implementing churn prediction helps to minimize those problems by detect potential churn employees or customers and apply targeted strategies to prevent them churning. By minimizing churn rate could maximize organizations revenue and productivity. This research presents churn prediction for human resource as a tool to prevent loss of valuable employees and for marketing as a tool to prevent customers’ loss in telecom industry. The author compared three common data mining algorithms on both datasets in predicting churn. The experiment result shows the overview prediction performance of each data mining algorithms in different unit of analysis. Overall, the best prediction was performed by Random Forest with PCC score 85.9% in employee churn and 95.4% in customer churn, F-Measure score 26.2% in employee churn and 82.2% in customer churn. However, Random Forest's F-Measure score in employee was not the highest as Neural Network has F-Measure score 47.1%. From this research, it can be concluded that Random Forest is a good data mining algorithm to perform churn prediction in different industries. Moreover, more comparison is recommended for real implementation in order to achieve best prediction model.

Item Type:	Thesis (Bachelor)
Uncontrolled Keywords:	Customer Churn; Employee Churn; Data Mining; Churn Prediction; Telecom
Subjects:	Q Science > QA Mathematics > QA76 Computer software >
Divisions:	Faculty of Engineering and Information Technology > Department of Information Technology
Depositing User:	Astuti Kusumaningrum
Date Deposited:	12 May 2020 01:47
Last Modified:	12 May 2020 01:47
URI:	http://repository.sgu.ac.id/id/eprint/273

Actions (login required)

View Item