This is an class kaggle competition on predicting the price of wine.
-
Predicting the price of wine based on a collection of reviews and other product features.
-
We use the Random Forest Regressor and the XGBoost Algorithm to predict the prices of wine.
-
Feature Engineering : We perform the following feature engineering ;
- Label encoding
- One Hot Encoding
- TF-IDF
-
10 Fold cross validation applied in both models.
-
Random Forest performs better than XGBoost.
-
Combining the models gives even better results.
-
We used two models after cross validation.
-
Their predictions were Averaged and RMSE of 20.4 was achieved.