Summary: No Free Lunch For Noise Prediction
Pasadena, CA 91125
February 20, 2000
-- Neural Computation, Vol 12, No. 3, 2000 --
No Free Lunch theorems have shown that learning algorithms
cannot be universally good. We show that No Free Lunch exists
for noise prediction as well. We show that when the noise is ad
ditive and the prior over target functions is ``uniform'', a prior on
the noise distribution cannot be updated, in the Bayesian sense,
from any finite data set. We emphasize the importance of a prior
over the target function in order to justify superior performance
for learning systems.
Keywords: No Free Lunch, Noise Prediction, Bayesian Prior.
Which data set has more noise?