Title :
Cross-validation with active pattern selection for neural-network classifiers
Author :
Leisch, Friedrich ; Jain, Lakhmi C. ; Hornik, Kurt
Author_Institution :
Inst. fur Stat. und Wahrscheinlichkeitstheor., Tech. Univ. Wien, Austria
fDate :
1/1/1998 12:00:00 AM
Abstract :
We propose a new approach for leave-one-out cross-validation of neural-network classifiers called “cross-validation with active pattern selection” (CV/APS). In CV/APS, the contribution of the training patterns to network learning is estimated and this information is used for active selection of CV patterns. On the tested examples, the computational cost of CV can be drastically reduced with only small or no errors
Keywords :
learning (artificial intelligence); neural nets; optimisation; pattern classification; probability; active pattern selection; cross-validation; network learning; neural-network; optimisation; pattern classifiers; probability; Artificial neural networks; Australia; Computational efficiency; Costs; Knowledge engineering; Predictive models; Sampling methods; Scholarships; Systems engineering and theory; Testing;
Journal_Title :
Neural Networks, IEEE Transactions on
