Neural Network Regression for Predicting U and V

This project builds and trains a neural network using TensorFlow to predict the values of U and V from input variables x and y.

Files

• train.py — Loads and preprocesses data, trains the model, saves predictions and training plot.
• test.py — Loads the saved model, evaluates performance, and generates a scatter plot comparing predicted and actual values.
• Data.xlsx — Input data file (must contain columns: x, y, U, V).

Outputs

• final_model.keras — Saved trained model.
• final_prediction.xlsx — Predicted values on training data.
• final_loss_plot.png — Log-scaled loss plot during training.
• test_predictions.xlsx — Model predictions and ground truth comparison.
• test_scatter.png — Scatter plot of true vs predicted values.
• Terminal output of R², MSE, RMSE, and MAE.

Plots

Test	Training Loss Plot

---

Author: Muhammed T. (@2mdtln)
Date: 2025-05-18

License

This project is licensed under the MIT License.