lorenzori
commented
6 years ago tried with different kernels on Nigeria LSMS, no improvement:
model = KernelRidge(alpha=1.0, kernel='poly')
clf = GridSearchCV(estimator=model, param_grid={"alpha":[0.001,0.01,0.1,1,10,100,1000], 'degree':[1,2,3,4,5,6,7,8,9,10]}, cv=inner_cv, scoring=r2_pearson)
clf.fit(data_features, y)
clf.grid_scores_
cross_val_score(clf, data_features, y, scoring=r2_pearson, cv=outer_cv).mean()
try to use a non linear regressor as http://scikit-learn.org/stable/modules/kernel_ridge.html