Customer-Retention-Deep-Learning

• Dataset is about bank customers in which we evaluate the customers retaining prediction.
• We will measure why customers are leaving.
• We will build a deep learning model to predict the retention
• At the end we use precision,recall, f1-score to measure the performance of our model.

Bulit a Model by using Tensorflow

• Used RELU and Softmax as an activation
• Choose optimizers as Adam and loss as binary_crossentropy
• We tried 10 epochs on model training
• Model secued 85% performance.

Conclusion

Truth

• 1552 times found correct and 55 times uncorrect at 0 which is 97% we can see on Recall 0
• 144 times found correct and 249 times uncorrect at 1 which is 37% we can see on Recall 1

Predicted

• 1552 times found correct and 249 times uncorrect at 0 which is 86% we can see on Precision 0
• 144 times found correct and 55 times uncorrect at 1 which is 72% we can see on Precision 1