Churn Prediction with Sequential Data and Deep Neural Networks. A Comparative Analysis

TL;DR

This paper compares neural network models, especially LSTM, with traditional methods for churn prediction using sequential RFM data, demonstrating superior performance of neural networks in capturing time-varying features.

Contribution

It provides a comparative analysis showing that LSTM neural networks outperform logistic regression in churn prediction with sequential data, and enhances logistic regression performance by incorporating LSTM probabilities.

Findings

LSTM with RFM data outperforms logistic regression in key metrics.

Using LSTM probabilities as features improves logistic regression performance by 25%.

Neural networks effectively utilize time-varying features for churn prediction.

Abstract

Off-the-shelf machine learning algorithms for prediction such as regularized logistic regression cannot exploit the information of time-varying features without previously using an aggregation procedure of such sequential data. However, recurrent neural networks provide an alternative approach by which time-varying features can be readily used for modeling. This paper assesses the performance of neural networks for churn modeling using recency, frequency, and monetary value data from a financial services provider. Results show that RFM variables in combination with LSTM neural networks have larger top-decile lift and expected maximum profit metrics than regularized logistic regression models with commonly-used demographic variables. Moreover, we show that using the fitted probabilities from the LSTM as feature in the logistic regression increases the out-of-sample performance of the latter by 25 percent compared to a model with only static features.