add new jupyter notebook for random forest

[Ruixi-Zhang]

• Performed hyperparameter tuning with the following parameter grid
  • n_estimators: [100, 200]
  • max_depth: [None, 10, 20]
  • min_samples_split: [2, 5]
  • min_samples_leaf: [1, 2]
• Best parameter:
  • min_samples_leaf=2, min_samples_split=5, n_estimators=200, random_state=42
• Result
  • Mean squared error: 0.1757393572711894
  • Coefficient of determination: 0.6579122100506867
• Also trained by city.

(More detailed results and graphs are in the written report)