Customer Churn Prediction in Computer Security Software
Dang, Quynh (2019)
Dang, Quynh
2019
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2019060414552
https://urn.fi/URN:NBN:fi:amk-2019060414552
Tiivistelmä
One of the most valuable assets of any company is its customer; hence it is vital for a company to focus on customer retention strategy by its advantage over customer acquisi- tion. Customer churn prediction is a tool for increasing customer retention by identifying potential churners prior to them leaving. There are many studies on churn prediction over the past decade on various sectors but still lacking churn prediction study on the computer software industry, especially, in the context of the combination between telecommunica- tion and security software. The paper attempted to find a solution to this problem by ap- plying state-of-the-art machine learning models to real-world data. The results show the best model performs 4.6 times better than random as well as correctly identify 100 percent of churn in the top five percent of ranked customers. The results also demonstrate the im-
portance of handling the imbalanced class issue in predicting customer churn.
portance of handling the imbalanced class issue in predicting customer churn.