Beygelzimer, Alina and Langford, John and Lifshits, Yuri and Sorkin, Gregory B. and Strehl, Alex
(2009)
*Conditional probability tree estimation analysis and algorithms.*
In: Uncertainty in artificial intelligence, 18-21 June 2009, Montreal, QC, Canada.

## Abstract

We consider the problem of estimating the conditional probability of a label in time $O(\log n)$, where $n$ is the number of possible labels. We analyze a natural reduction of this problem to a set of binary regression problems organized in a tree structure, proving a regret bound that scales with the depth of the tree. Motivated by this analysis, we propose the first online algorithm which provably constructs a logarithmic depth tree on the set of labels to solve this problem. We test the algorithm empirically, showing that it works succesfully on a dataset with roughly $10^6$ labels.

Item Type: | Conference or Workshop Item (Paper) |
---|---|

Official URL: | http://uai.sis.pitt.edu/displayArticles.jsp?mmnu=1... |

Additional Information: | © 2009 the Authors |

Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science |

Sets: | Research centres and groups > Management Science Group Departments > Management |

Date Deposited: | 13 May 2011 10:37 |

Last Modified: | 06 Jun 2012 17:16 |

URI: | http://eprints.lse.ac.uk/id/eprint/35637 |

### Actions (login required)

View Item |