An Extended Cencov-Campbell Characterization of Conditional Information Geometry
暂无分享,去创建一个
We formulate and prove an axiomatic characterization of conditional information geometry, for both the normalized and the non-normalized cases. This characterization extends the axiomatic derivation of the Fisher geometry by Cencov and Campbell to the cone of positive conditional models, and as a special case to the manifold of conditional distributions. Due to the close connection between the conditional I-divergence and the product Fisher information metric the characterization provides a new axiomatic interpretation of the primal problems underlying logistic regression and AdaBoost.
[1] Shun-ichi Amari,et al. Methods of information geometry , 2000 .
[2] L. Campbell. An extended Čencov characterization of the information metric , 1986 .
[3] N. Čencov. Statistical Decision Rules and Optimal Inference , 2000 .
[4] John D. Lafferty,et al. Boosting and Maximum Likelihood for Exponential Models , 2001, NIPS.