AN EFFECTIVE MACHINE LEARNING MODEL FOR CUSTOMER ATTRITION PREDICTIONIN MOTOR INSURANCE USING GWO-KELM ALGORITHM

The prediction of customers' churn is a challenging task in different industrial sectors, in which the motor insurance industry is one of the well-known industries. Due to the incessant upgradation done in the insurance policies, the retention process of customers plays a significant role for the concern. The main objective of this study is to predict the behaviors of the customers and to classify the churners and non-churners at an earlier stage. The Motor Insurance sector dataset consists of 20,000 records with 37 attributes collected from the machine learning industry. The missing values of the records are analyzed and explored via Expectation Maximization algorithm that categorizes the collected data based on the policy renewals. Then, the behavior of the customers are also investigated, so as to ease the construction process training classifiers. With the help of Naive bayes algorithm, the behaviors of the customers on the upgraded policies are examined. Depending on the dependency rate of each variable, a hybrid GWO-KELM algorithm is introduced to classify the churners and non-churners by exploring the optimal feature analysis. Experimental results have proved the efficiency of the hybrid algorithm in terms of 95% prediction accuracy; 97% precision; 91% recall & 94% F-score.

Efficient End-to-End Asynchronous Time-Series Modeling With Deep Learning to Predict Customer Attrition

Modeling time-series data with asynchronous, multi-cardinal, and uneven patterns presents several unique challenges that may impede convergence of supervised machine learning algorithms, or significantly increase resource requirements, thus rendering modeling efforts infeasible in resource-constrained environments. The authors propose two approaches to multi-class classification of asynchronous time-series data. In the first approach, they create a baseline by reducing the time-series data using a statistical approach and training a model based on gradient boosted trees. In the second approach, they implement a fully convolutional network (FCN) and train it on asynchronous data without any special feature engineering. Evaluation of results shows that FCN performs as well as the gradient boosting based on mean F1-score without computationally complex time-series feature engineering. This work has been applied in the prediction of customer attrition at a large retail automotive finance company.

Customer Attrition and Retention

In a competitive environment, organizations and firms are susceptible to customer attrition. Customer attrition and customer retention terms are widely spoken about. Customer retention which is quite the opposite of attrition is important for the company’s sustainability in today’s market. Many studies have come up with an attempt to find factors that influence customer retention. Firms have long desired to know who might end their relationship with them. Similarly, companies try to find how many existing customers did not return to purchase. Customer attrition is a problem that deals with clients and customers who are attrited from a particular brand or firm. In simple terminology it deals with the loss of profit associated with companies. This paper deals with the different ways to overcome the increasing attrition rate among customers. It also includes the implementation of customer segmentation using RFM model and K-means clustering. It also includes the prediction of customer retention using logistic regression.

Employee's Attrition Prediction Using Machine Learning Approaches

In any industry, attrition is a big problem, whether it is about employee attrition of an organization or customer attrition of an e-commerce site. If we can accurately predict which customer or employee will leave their current company or organization, then it will save much time, effort, and cost of the employer and help them to hire or acquire substitutes in advance, and it would not create a problem in the ongoing progress of an organization. In this chapter, a comparative analysis between various machine learning approaches such as Naïve Bayes, SVM, decision tree, random forest, and logistic regression is presented. The presented result will help us in identifying the behavior of employees who can be attired over the next time. Experimental results reveal that the logistic regression approach can reach up to 86% accuracy over other machine learning approaches.

Models for Optimisation of Delivery Timing and Volumes under Uncertainty Encompassing Penalty and Customer Attrition Risks

The paper proposes a model for inventory management and specifically for determining the optimum volume and timing of deliveries, encompassing the uncertainty of demand. The criteria of efficiency are the minimisation of integral costs and maximisation of profit with due regard for the risks of penalties and customer attrition. The triangular distribution is a reference for the distribution pattern of the stochastic demand and delivery timing fluctuations as it is one of the most common choices in case of insufficient statistical data

Customer attrition analysis in the securities industry: a large-scale field study in Korea

Purpose The purpose of this paper is to identify the attributes that predict customer attrition behavior in the brokerage and investment banking sectors. Design/methodology/approach The authors analyze the complete stock trading records and customer profiles of 458,098 retail customers from a Korean brokerage house. The authors develop customer attrition prediction models and further explore the practicality of these models using statistical classification techniques. Findings The results from three different binary selection models indicate that customer transaction patterns effectively explain the attrition of active retail customers in subsequent periods. The study results demonstrate that monetary value variables are the most critical for predicting customer attrition in the securities industry. Research limitations/implications This study contributes to the customer attrition literature by documenting the first large-scale field-based evidence that confirms the practicality of the canonical recency, frequency and monetary (RFM) framework in the investment banking and brokerage industry. The findings advance previous survey-based studies in the financial services industry by identifying the attributes that predict customer attrition behaviors in the securities industry. Practical implications The outcomes can be easily operationalized for attrition prediction by practitioners in financial service firms. Moreover, the ex post density of inactive customers in the top 10 percent most-likely-to-churn group is estimated to be five to six times the ex ante unconditional attrition ratio, which ascertains that the attributes recognized in this study work well for the purpose of target marketing. Originality/value While the securities industry is regarded as one of the most information-intensive industries, detailed empirical investigation into customer attrition in the field has lagged behind partly due to the lack of suitable securities transaction data and demographic information at the customer level. The current research fills this gap in the literature by taking advantage of a large-scale field data set and offers a starting point for more elaborate studies on the drivers of customer attrition in the financial services sector.

Decision Rule Induction: Relieving Complexity in Detecting Defection

Customer attrition has become a serious problem globally, particularly in telecom service, resulting into substantial revenue decline. Attrition may result in accumulation ofdues as a resultof payment defaults.Proactive identification of potential attrite will help in retention as well as minimizing loss of revenue.For attrition detection many robust but complex algorithms are used. Depending on the severity of error, the complexity can be lessened and thus cost. Two methods of decision rules (1R& C5.0) are used to predict the attrition and predictive accuracy is judged withconfusion matrix. Comparison between models is made by sensitivity and specificity. It was found that 1R has a sensitivity of .60 against .69 for C5.0 and hence, the performance is not significantly different. It is suggested that 1R could be used instead of more complex algorithmsand also it can be adopted for benchmarking