This repository contains a linear regression model built with PyTorch to predict gasoline prices in Turkey based on three features: USD exchange rate, policy interest rate, and oil prices. The project also includes a 3D and 2D visualization of the model's predictions and the actual gasoline prices.
This will:
Train a simple linear regression model with three features.
Predict gasoline prices based on the trained model.
Plot 3D and 2D visualizations of the real and predicted gasoline prices.
Generate the prediction for a new data point (9th month values).
The model predictions are visualized in two plots:
3D Plot: Shows the relationship between normalized USD exchange rate, policy interest rate, and gasoline prices. A surface plot is also created to represent the linear regression model.
2D Plot: Compares real gasoline prices and predicted prices over time.