datascience-visualize-grid-cv

datascience-visualize-grid-cv.py

def visualize_grid_cv_params(grid_cv):
    df = pd.DataFrame(grid_cv.cv_results_['params'])
    df['score'] = grid_cv.cv_results_['mean_test_score']

    fig, axes = plt.subplots(1, len(grid_cv.param_grid), sharey=True, figsize=(15,4))

    i = 0
    for param in grid_cv.param_grid:
        data = df.groupby(param).mean()['score'].to_dict()
        param_values = list(data.keys())
        score_values = list(data.values())

        ax = axes[i]
        ax.plot(param_values, score_values, marker='o')
        ax.set_xlabel(param)    
        ax.set_ylabel('Score')
        ax.grid(True)
        i = i+1
        
# Example:
import pandas as pd
import sklearn
from sklearn import model_selection, ensemble
%pylab inline

param_grid = {
    'n_estimators' : [5,10,15,50,100],
    'max_features' : [5, 10, 20, 40, 50, X.shape[1]],
    'max_depth' : [2,5,10,20,100],
}
random_forest_estimator = ensemble.RandomForestClassifier()
grid_cv = model_selection.GridSearchCV(random_forest_estimator, param_grid, scoring = 'accuracy', cv=3, return_train_score=True, verbose=2)
grid_cv.fit(X, Y)