Preventing overfitting in GP with canary functions

Authors:
Nate Foreman;Matthew Evett
Affiliations:
Altarum Institute, Ann Arbor, MI;Eastern Michigan University, Ypsilanti, MI
Venue:
GECCO '05 Proceedings of the 7th annual conference on Genetic and evolutionary computation
Year:
2005

Citing 3
Cited 6

Genetic programming: on the programming of computers by means of natural selection

Genetic programming: on the programming of computers by means of natural selection
Automatic early stopping using cross validation: quantifying the criteria

Neural Networks
Balancing accuracy and parsimony in genetic programming

Evolutionary Computation

Reducing overfitting in manufacturing process modeling using a backward elimination based genetic programming

Applied Soft Computing
Sequential metamodelling with genetic programming and particle swarms

Winter Simulation Conference
Validation sets for evolutionary curtailment with improved generalisation

ICHIT'11 Proceedings of the 5th international conference on Convergence and hybrid information technology
Improving the generalisation ability of genetic programming with semantic similarity based crossover

EuroGP'10 Proceedings of the 13th European conference on Genetic Programming
Random sampling technique for overfitting control in genetic programming

EuroGP'12 Proceedings of the 15th European conference on Genetic Programming
Where should we stop? an investigation on early stopping for GP learning

SEAL'12 Proceedings of the 9th international conference on Simulated Evolution and Learning

Quantified Score

Hi-index	0.00

Visualization

Abstract

Overfitting is a fundamental problem of most machine learning techniques, including genetic programming (GP). Canary functions have been introduced in the literature as a concept for preventing overfitting by automatically recognizing when it starts to occur. This paper presents a simple scheme for implementing canary functions using cross-validation. The effectiveness of this technique is demonstrated by applying it to the numeric regression problem. A list of conditions and criteria for applying this technique to other problem domains is also identified. Other strategies for dealing with overfitting in GP are discussed.