|
|||||
|
|
Performance analysis of a greedy algorithm for inferring Boolean functions |
|
| Authors: | Daiji Fukagawa Tatsuya Akutsu |
| Affiliation: | a Department of Intelligence Science and Technology, Graduate School of Informatics, Kyoto University, Kyoto 606-8501, Japan b Bioinformatics Center, Institute for Chemical Research, Kyoto University, Kyoto 611-0011, Japan |
| Abstract: | We analyzed average case performance of a known greedy algorithm for inference of a Boolean function from positive and negative examples, and gave a proof to an experimental conjecture that the greedy algorithm works optimally with high probability if both input data and the underlying function are generated uniformly at random. |
| Keywords: | Analysis of algorithms Unbalanced function Uniform distribution Machine learning |
| 本文献已被 ScienceDirect 等数据库收录！ | |