The scikit-learn DummyRegressor
class implements several strategies for random guessing, which can serve as baseline for regressors. The strategies are as follows:
mean
: This predicts the mean of the training set.median
: This predicts the median of the training set.quantile
: This predicts a specified quantile of the training set when provided with thequantile
parameter. We will apply this strategy by specifying the first and third quartile.constant
: This predicts a constant value that is provided by the user.
We will compare the dummy regressors with the regressors from Chapter 9, Ensemble Learning and Dimensionality Reduction, using R-squared, MSE, MedAE, and MPE.
The imports are as follows:
import numpy as np from sklearn.dummy import DummyRegressor import ch10util from sklearn import metrics import dautil as dl from IPython.display import HTML
Load the temperature data as follows:
y_test = np.load('temp_y_test.npy') X_train = np.load('temp_X_train...