TY - GEN
T1 - Fast non-parametric Bayesian inference on infinite trees
AU - Hutter, Marcus
PY - 2005
Y1 - 2005
N2 - Given i.i.d. data from an unknown distribution, we consider the problem of predicting future items. An adaptive way to estimate the probability density is to recursively subdivide the domain to an appropriate data-dependent granularity. A Bayesian would assign a data-independent prior probability to "subdivide", which leads to a prior over infinite(ly many) trees. We derive an exact, fast, and simple inference algorithm for such a prior, for the data evidence, the predictive distribution, the effective model dimension, and other quantities.
AB - Given i.i.d. data from an unknown distribution, we consider the problem of predicting future items. An adaptive way to estimate the probability density is to recursively subdivide the domain to an appropriate data-dependent granularity. A Bayesian would assign a data-independent prior probability to "subdivide", which leads to a prior over infinite(ly many) trees. We derive an exact, fast, and simple inference algorithm for such a prior, for the data evidence, the predictive distribution, the effective model dimension, and other quantities.
UR - http://www.scopus.com/inward/record.url?scp=84862594932&partnerID=8YFLogxK
M3 - Conference contribution
SN - 097273581X
SN - 9780972735810
T3 - AISTATS 2005 - Proceedings of the 10th International Workshop on Artificial Intelligence and Statistics
SP - 144
EP - 151
BT - AISTATS 2005 - Proceedings of the 10th International Workshop on Artificial Intelligence and Statistics
T2 - 10th International Workshop on Artificial Intelligence and Statistics, AISTATS 2005
Y2 - 6 January 2005 through 8 January 2005
ER -