The one-inclusion graph algorithm is near-optimal for the prediction model of learning
| Title | The one-inclusion graph algorithm is near-optimal for the prediction model of learning |
| Publication Type | Journal Articles |
| Year of Publication | 2001 |
| Authors | Li Y, Long PM, Srinivasan A |
| Journal | IEEE Transactions on Information Theory |
| Volume | 47 |
| Issue | 3 |
| Pagination | 1257 - 1261 |
| Date Published | 2001/03// |
| ISBN Number | 0018-9448 |
| Keywords | Computer science, concept class, general-purpose algorithm, graph theory, learning prediction model, learning systems, near-optimal algorithm, Neural networks, one-inclusion graph algorithm, optimisation, Prediction algorithms, prediction theory, Predictive models, probability, Probability distribution, Upper bound, Vapnik-Chervonenkis dimension, Virtual colonoscopy |
| Abstract | Haussler, Littlestone and Warmuth (1994) described a general-purpose algorithm for learning according to the prediction model, and proved an upper bound on the probability that their algorithm makes a mistake in terms of the number of examples seen and the Vapnik-Chervonenkis (VC) dimension of the concept class being learned. We show that their bound is within a factor of 1+o(1) of the best possible such bound for any algorithm |
| DOI | 10.1109/18.915700 |