Churn Prediction with Sequential Data and Deep Neural Networks. A Comparative Analysis
This work addresses churn prediction for financial services providers, but it is incremental as it applies existing methods (LSTM) to a specific domain without introducing new techniques.
The paper tackled churn prediction by comparing LSTM neural networks with logistic regression using sequential RFM data from a financial services provider, finding that LSTM models achieved higher top-decile lift and profit metrics and improved logistic regression performance by 25% when used as a feature.
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.