Abstract

Abstract: This paper introduces stacked generalization, a scheme for minimizing the generalization error rate of one or more generalizers. Stacked generalization works by deducing the biases of the generalizer(s) with respect to a provided learning set. This deduction proceeds by generalizing in a second space whose inputs are (for example) the guesses of the original generalizers when taught with part of the learning set and trying to guess the rest of it, and whose output is (for example) the correct guess. When used with multiple generalizers, stacked generalization can be seen as a more sophisticated version of cross-validation, exploiting a strategy more sophisticated than cross-vali-dation’s crude winner-takes-all for combining the individual generalizers. When used with a single generalizer, stacked generalization is a scheme for estimating (and then correcting for) the error of a generalizer which has been trained on a particular learning set and then asked a particular ques-tion. After introducing stacked generalization and justifying its use, this paper presents two numer-ical experiments. The first demonstrates how stacked generalization improves upon a set of sepa-rate generalizers for the NETtalk task of translating text to phonemes. The second demonstrates how stacked generalization improves the performance of a single surface-fitter. With the other ex-perimental evidence in the literature, the usual arguments supporting cross-validation, and the ab-stract justifications presented in this paper, the conclusion is that for almost any real-world gener-alization problem one should use some version of stacked generalization to minimize the general-ization error rate. This paper ends by discussing some of the variations of stacked generalization, and how it touches on other fields like chaos theory. Key Words: generalization and induction, combining generalizers, learning set pre-processing, cross-validation, error estimation and correction.

Citations

2888	Induction of Decision Trees - Quinlan - 1986
2289	Adaptation in Natural and Artificial Systems - Holland - 1975
1515	A Theory of the Learnable - Valiant - 1984
439	Toward memory-based reasoning - Stanfill, Waltz - 1986
218	Stochastic complexity and modeling - Rissanen - 1986
130	Nonlinear prediction of chaotic time series - Casdagli - 1989
128	NETtalk: a parallel network that learns to read aloud - Sejnowski, Rosenberg - 1986
114	Methods for Solving Incorrectly Posed problems - Morozov - 1984
78	Exploiting chaos to predict the future and reduce noise. Los Alamos preprint LA-UR-88-901 - D, Sidorowich - 1988
78	Efficient algorithms with neural network behavior - Omohundro - 1987
60	How neural nets work - Lapedes, Farber - 1988
38	Asymptotics for and against cross-validation - Stone - 1977
35	Computers and the theory of statistics: Thinking the unthinkable - Efron - 1979
16	From Stein's unbiased risk estimates to the method of generalized cross validation - Li - 1985
15	Constructing a generalizer superior to NETtalk via a mathematical theory of generalization. Neural Networks - Wolpert - 1990
10	Mit progress in understanding images - Poggio - 1988
7	The relationship between Occam’s razor and convergent guessing - Wolpert - 1990
4	Comparison of predicted and experimentally determined structure of adenyl kinase," Nature 250 - Schulz - 1974
4	A mathematical theory of generalization: part - Wolpert - 1990
3	Machine learning. Annual review of computer science - Dietterich - 1990
3	A benchmark for how well neural nets generalize - Wolpert - 1989
3	A mathematical theory of generalization: part II - Wolpert - 1990
3	Improving the performance of generalizers via time-series-like pre-processing of the learning set. Los Alamos Laboratory Report LA-UR-91-350 - Wolpert - 1991
2	Hierarchical training of neural networks and prediction of chaotic time series, Phys. Lett. A 158 - Deppisch, Bauer, et al. - 1991
1	On the ability of neural networks to perform generalization by induction - Anshelevich - 1989