| Abstrakt: |
The Indian insurance industry is found to be dynamically changing with the global market, and the increase in the competition is observed based on the entry of several players post market liberalization. Therefore, it is imperative for the insurance companies to maintain the portfolios of the existing customers along with attaining new customers. Besides, it is noticed that a useful technique in retaining the current customers would make the insurance company more profitable by reducing the overall cost incurred in the marketing and advertisements (cost of customer acquisition) to offset the recruitment of new customer [1]. A recent study, indicate that the customer retention is a significant aspect of improving operational efficiencies and the diverse service sectors are required to focus better on retaining its existing customers. In the last decade, the customer relationship management is seen to have growing attention in terms with the prediction of customer churn prediction. In recent studies, several data mining techniques are employed to predict the customer churn by considering the characteristics of the datasets, validation, and evaluation of the behaviour of the customers. With the emergence of newer technologies in data mining like Big Data, Machine Learning, Neural Networks etc., such models are gaining dimensions which couldn't have been explored earlier. In this paper, a study on different data mining techniques and machine learning models of customer attrition prediction in the insurance sector are described. Generally, used prediction models are logistic regression, decision trees, artificial neural networks, support vector machines and random forests. Various authors/researchers have used these models in their study to predict the customer churn in service sectors. This compares the various prediction models used to predict the customer churn in the insurance and telecom sector. [ABSTRACT FROM AUTHOR] |
