The churn_modelling dataset is of Kaggle which predicts for a binary output whether or not a customer leaves a bank or not based on a few parameters. Deep learning classification algorithms are involved which predict the output by training the ANN model for two hidden layers. For the output prediction a 'sigmoid' activation function has been used which gives the exact probability of a person staying with the bank or not. However, for the sake of simplicity, in this particular model I have considered 1/2 as threshold (i.e) the people who have the probabilities greater than 1/2 will stay with the bank and others would leave.
- 0 --> Customer left the bank
- 1 --> Customer stays with the bank