Customer Churn Prediction using TensorFlow/Keras
-
Project Overview
This project uses an Artificial Neural Network (ANN) to predict customer churn based on a dataset of telecom customers. The goal is to identify users who are likely to cancel their service so the business can take action to keep them.
-
Model Architecture
I built a Sequential model with three layers:
Input Layer: Processes the customer features (tenure, monthly charges, etc.).
Hidden Layers: Two layers with 32 and 64 neurons using the ReLU activation function to find patterns.
Output Layer: A single neuron with a Sigmoid activation to output a probability between 0 and 1.
- Training & Performance
Optimizer: Stochastic Gradient Descent (SGD).
Loss Function: Binary Crossentropy.
Epochs: 200.
Final Accuracy: 78.1% on the test dataset.
- How to Use
To use this model, you will need TensorFlow installed. You can load the my_model.keras file using the following Python code:
Python
from tensorflow.keras.models import load_model
model = load_model('my_model.keras')
