Random forest

# Model 1: Ranbdom Forest Rgressor
import sklearn
from sklearn.model_selection import GridSearchCV
from sklearn.ensemble import RandomForestRegressor

# Split training data into features (x_train) and labels (Y_train)
x_train = Train_encoded.drop(columns=['Id','SalePrice','SalePrice_log'])
Y_train = Train_encoded.SalePrice_log


# Set paramters for Grid Search
param_grid =  {'n_estimators':[200, 300, 400, 500, 600],
               'max_features':[0.1, 0.3, 0.6]
              }
# Initialise the random forest model 
RandForest = RandomForestRegressor(n_jobs= -1, random_state = 0, bootstrap=True)

# Initialise Gridsearch CV with 5 fold corssvalidation and neggative root_mean_squared_error
Tuned_RandForest = GridSearchCV(estimator=RandForest, param_grid=param_grid, scoring='neg_root_mean_squared_error', cv=5)

# Fit model & Time the process for training the model
start_time = time.process_time()
Tuned_RandForest.fit(x_train, Y_train)
# End of fit time
print(time.process_time() - start_time, "Seconds")

# Record the results for all models in a pandas dataframe and keep only the best model
Results = pd.DataFrame(Tuned_RandForest.cv_results_)
Results_Best = Results.loc[Results.rank_test_score==1]
    
print('Random Forest Regressor')
#Results = Results.loc[Results.rank_test_score==1]

# Create a plot to show all models 

col = ['param_max_features']

for col in col:
    grid = sns.FacetGrid(Results, col=col, hue='rank_test_score', palette="tab20c", legend_out=False,
                         col_wrap=5, height=5)

    # Draw a horizontal line to show the starting point
    grid.map(plt.axhline, y=0, ls=":", c=".5")

    # Draw marker on plot and decide what parameters to plot
    grid.map(plt.plot, "param_n_estimators", "mean_test_score", marker="o")

    # Adjust the arrangement of the plots
    grid.fig.tight_layout(w_pad=1)
    
    # Add legend to gird 
    grid.add_legend()