aspg-1-795-1791417426 20261007235706 American Scientific Publishing Group admin@americaspg.com American Scientific Publishing Group American Journal of Business and Operations Research AJBOR 2770-0216 2692-2967 2021 4 2 Taif Khalid Shakir Ascencia Business School, College de Paris, France Ahmed N. Al Masri American University in the Emirates, Dubai, United Arab Emirates Customer churn prediction (CCP) is a crucial problem in telecom industry which helps to improve the revenue of the company and prevent the loss of customers. Customer churn is an important issue in service sector with highly competitive services. At the same time, the prediction of users who are probably leaving the company can be identified at an earlier stage to prevent loss of revenue. Several works have used machine learning (ML) techniques for predicting the existence of customer churn in different industries. With this motivation, this paper presents an optimal long, short-term memory with stacked autoencoder (OLSTM-SAE) technique for CCP in telecom industry. The OLSTM-SAE technique encompasses three subprocesses namely preprocessing, classification, and parameter optimization. The OLSTM-SAE technique classifies the preprocessed data into churn and non-churn customers. In addition, the grey wolf optimization (GWO) technique is used to adjust the variables involved in the LSTM-SAE model. For examining the enhanced performance of the OLSTM-SAE technique, an extensive simulation analysis takes place, and the outcomes are inspected with respect to various measures. The experimental results highlighted the betterment of the OLSTM-SAE technique in terms of different evaluation parameters. 2021 57 64 795 https://creativecommons.org/licenses/by/4.0/ 10.54216/AJBOR.040202 https://www.americaspg.com/journal/1/article/795 https://www.americaspg.com/storage/9845.pdf