Application of neural networks for the sensitivity analysis of customer loyalty
| Authors | |
|---|---|
| Year of publication | 2012 |
| MU Faculty or unit | |
| Citation | |
| Description | The research topic deals with analysis of specific conditions for application of customer relationship management (CRM) indicators derived from historical information of customer behaviour extracted from the transactional data, widely available in the enterprises. The neural network analysis was applied for the classification task of distinguishing potentially returning customers from those, who tend to leave the company. The experimental research was performed by mining customer database of the travel agency. The experimental evaluation revealed that the neural network model could not be uniformly applied throughout all the customer lifecycle, as the classification model had increasing ability to recognize the reliable customers (from 69.5 % to 95.5%), but its performance to predict further churn of the customer was weakened (from 63.4% to 22.2%). Therefore the new model was suggested based on the dynamics of sensitivity of the customer indicators, which allowed to improve its accuracy and provide new research insights. |
| Related projects: |