Prediction of Customer Churn in the Banking Industry Using Artificial Neural Networks: Prediction of Customer Churn in the Banking Industry Using Artificial Neural Networks

Willi Akbar Satria (Universitas Nasional)
Iskandar Fitri (Universitas Nasional)
Sari Ningsih (Universitas Nasional)

Jurnal Mantik
Vol. 4 No. 1 (2020): May: Manajemen, Teknologi Informatika dan Komunikasi (Mantik)

The Artificial Neural Network (ANN) is an architect in deep learning inspired by how the human brain works. As the name suggests, this technique works like a biological neural network possessed by a living creature, by accepting input data, processing it in a node/neuron, and displaying it in the output. ANN works by utilizing the large number of layers and the number of existing nodes/neurons to perform tasks, such as performing feature extraction, pattern recognition, regression, and classification. In this journal ANN's architect was used in the design of deep learning model, to solve classification problem, in case of banking churn prediction. The number of layers (layers) used by 4 layers, namely input layer, 2 hidden layer, and output layer with the number of nodes/neurons in sequence as much as 10.6, 6, 1. The final result is a model that is already (ready) deployed into a web-based application that can predict the churn (banking) with an accuracy of 84% on the customer's input data