Creating an Energy Switcher Prediction Model

Abstract

The aim of this research is to investigate the use of an Energy Switcher Prediction Model for clients of the Volksbank. The model predicts to which energy supplier a client will go, based on a clients transaction data. To make this prediction possible, the model should be technically feasible enough, legally allowed and ethically responsible. If the outcome of this research is positive, more services based on personal transaction data could be developed. On the technical side of this research the collection of the dataset is described and the results of the multilayer perceptron. The legal and ethical side of the research question is placed in a framework which could be used for other machine learning applications. The results show that an Energy Switcher Prediction Model is technical, legal and ethical not yet feasible and future research is needed.