Author: Mohammadreza Ebrahimi
Email: [email protected]
This repository is included as a perfect example of Machin learning.
It is about electron mass prediction by utilizing CERN electron collision data. We examined different models as follows
- LinearRegression
- DecisionTree
- RandomForest
Then, after training the model by using them, we examined them
for data which prepared by Polynomial Feature.
At the end, we would train several perfect models with low-cost function_ where the best RMSE would be gotten as 1.6 for the test set.
Still, this model can give us the better MSE or RMSE
by tuning Hyperparameter.
If you have any questions or suggestions please contact me.
Prediction of Mass
Mass of electron (min=2.00 - max=109.99)
Cross Validation of Training dataset
Shape of data : (100000, 19)
| Model | RMSE |
|---|---|
| Linear Regression | 19.29 |
| LASSO (alpha=0.01) | 19.50 |
| Decision Tree | 12.07 |
| Random Forest | 6.51 |
| Polynomial Features (degree=2) | |
| Linear Regression | 4.57 |
| Decision Tree | 2.69 |
| Random Forest | 1.68 |
| Test dataset | |
| Random Forest | 1.60 |