Wrapped input selection using multilayer perceptrons for repeat-purchase modeling in direct marketing

In this paper, we try to validate existing theory on and develop additional insight into repeat-purchase behavior in a direct marketing setting by means of an illuminating case study. The case involves the detection and qualification of the most relevant RFM (Recency, Frequency and Monetary) variables, using a neural network wrapper as our input pruning method. Results indicate that elimination of redundant and/or irrelevant inputs by means of the discussed input selection method allows us to significantly reduce model complexity without degrading the predictive generalization ability. It is precisely this issue that will enable us to infer some interesting marketing conclusions concerning the relative importance of the RFM predictor categories and their operationalizations. The empirical findings highlight the importance of a combined use of RFM variables in predicting repeat-purchase behavior. However, the study also reveals the dominant role of the frequency category. Results indicate that a model including only frequency variables still yields satisfactory classification accuracy compared to the optimally reduced model.