An evolutionary multi-objective local selection algorithm for customer targeting

In an increasingly competitive marketplace, one of the most interesting and challenging problems is how to identify and profile customers who are most likely to be interested in new products or services. At the same time, minimizing the number of variables used in the prediction task is important with large databases. We consider a novel application of evolutionary multi-objective algorithms for customer targeting. Evolutionary algorithms are considered effective in solving multi-objective problems because of their inherent parallelism. We use ELSA, an evolutionary local selection algorithm that maintains a diverse population of solutions approximating the Pareto front in a multi-dimensional objective space. We use artificial neural networks (ANNs) for customer prediction and ELSA to search for promising subsets of features. Our results on a real data set show that our approach is easier to interpret and more accurate than the traditional method used in marketing.