NEURAL NETWORKS IN PREDICTING INDIVIDUAL CUSTOMER PROFITABILITY
Dražena Gašpar1, Ivica Ćorić2, Mirela Mabić3
This paper presents how neural networks can be used for the analysis of current and predicting of future individual customer profitability. Based on historical data stored in the data warehouse, neural networks predict individual customer profitability by defining belonging of particular customer to one of predefined customer segments. Neural networks were used as a tool for analyzing and predicting customer profitability because of their ability to learn from historical data and to make a valid generalization. Since one of the main features of neural networks is their possibility to work with nonlinear and nonfinancial data, it makes them an appropriate tool for predicting individual customer profitability. A presented use of neural networks in predicting individual customer profitability was tested using empirical data from a production company which operates in the Southeast European market.