Hey,
Just going through example test, for semi supervised learning, when the regularisation parameter tends to infinty, the f* is nothing but the first eigen vector , as the model will give less importance to correctness of solution.Is this intution correct?