An Optimized early churn prediction for food industry by deep neural decision forests

Abstract

Customer retention is one of the main goals in large-scale food industries. This is achieved by predicting churn customers in advance and satisfying their needs. This research focuses on providing a higher accurate model than the other previous models. The deep neural decision forest model is a combination of Random Forest and Convolutional Neural Network models that help in fulfilling the objective. The dataset is taken from a US food industry to train and test the models. This model has achieved 92% accuracy in predicting the churn customers.