Oversearching and Layered Search in Empirical Learning
Proceedings of the Fourteenth International Joint Conference on Artificial Intelligence. Learning. Pages 1019-1023.
Abstract
When learning classifiers, more extensive search for rules is shown to lead to lower predictive accuracy on many of the real-world domains investigated. This counter-intuitive result is particularly relevant to recent systematic search methods that use risk-free pruning to achieve the same outcome as exhaustive search. We propose an iterated search method that commences with greedy search, extending its scope at each iteration until a stopping criterion is satisfied. This layered search is often found to produce theories that are more accurate than those obtained with either greedy search or moderately extensive beam search.
@inproceedings{ijcai1995p132,
title = {Oversearching and Layered Search in Empirical Learning},
author = {Quinlan, J. Ross and Cameron-Jones, R. Mike},
booktitle = {Proceedings of the Fourteenth International Joint Conference on
Artificial Intelligence, {IJCAI-95}},
publisher = {International Joint Conferences on Artificial Intelligence},
editor = {Chris S. Mellish},
pages = {1019--1023},
year = {1995},
note = {Learning},
url = {https://ijcai.org/proceedings/1995/132},
}
TY - CPAPER TI - Oversearching and Layered Search in Empirical Learning AU - Quinlan, J. Ross AU - Cameron-Jones, R. Mike T2 - Proceedings of the Fourteenth International Joint Conference on Artificial Intelligence, IJCAI-95 PB - International Joint Conferences on Artificial Intelligence PY - 1995 SP - 1019 EP - 1023 UR - https://ijcai.org/proceedings/1995/132 ER -Copied
Machine-readable versions of this paper, for tools and AI agents:
