Input Relevance in Multi-Layer Perceptron for Fundraising

In this contribution, we consider a Multi-Layer Perceptron (MLP) methodology for predicting specific gift features, particularly the count of donations and the gift amounts. Moreover, we use Garson's indicator to evaluate the relative importance of the input variables to the output(s) in the MLP model with the aim of enhancing the effectiveness of fundraising campaigns. In the discussed application, the Donors' behaviors are estimated using a simulated dataset that includes individual characteristics and information about donation history.