def select_features(rf):
"""
return the index of the features that have importance above average
"""
ft_names = feature_names()
idx = np.arange(len(ft_names))
importances = rf.feature_importances_
ft = zip(idx, 100.0 * importances/importances.max())
ft_sorted = sorted(ft, key=lambda x: x[1], reverse=True)
return [tup[0] for tup in ft_sorted]
def select_features(rf):
"""
return the index of the features that have importance above average
"""
ft_names = feature_names()
idx = np.arange(len(ft_names))
importances = rf.feature_importances_
ft = zip(idx, 100.0 * importances / importances.max())
ft_sorted = sorted(ft, key=lambda x: x[1], reverse=True)
return [tup[0] for tup in ft_sorted]
def var_importance(rf):
ft_names = feature_names()
importances = rf.feature_importances_
ft = zip(ft_names, 100.0 * importances/importances.max())
ft_sorted = sorted(ft, key=lambda x: x[1])
return np.asarray(ft_sorted)
def var_importance(rf):
ft_names = feature_names()
importances = rf.feature_importances_
ft = zip(ft_names, 100.0 * importances / importances.max())
ft_sorted = sorted(ft, key=lambda x: x[1])
return np.asarray(ft_sorted)