developed a model that can predict air temperature according to atmospheric pressure.
In this project we used "Saudi Arabia weather history" dataset from Kaggle. We took 2 attributes (temp, barometer), type(int64, float64), and 5000 random examples, 80% for training and 20% for testing.
We used 3 different models:
- Linear Regression
- Batch Gradient Descent
- Polynomial Regression
Tried different polynomial degrees, learning rates, and number of iterations.
We compared them by using:
- Mean absolute error (MAE)
- Mean sum of squares error (MSSE)
- Root mean square error (RMSE)