Efficient estimation and model selection in large graphical models |
| |
Authors: | Dag Wedelin |
| |
Institution: | (1) Department of Computing Science, Chalmers University of Technology, S-412 96 Gothenburg, Sweden |
| |
Abstract: | We develop a computationally efficient method to determine the interaction structure in a multidimensional binary sample. We use an interaction model based on orthogonal functions, and give a result on independence properties in this model. Using this result we develop an efficient approximation algorithm for estimating the parameters in a given undirected model. To find the best model, we use a heuristic search algorithm in which the structure is determined incrementally. We also give an algorithm for reconstructing the causal directions, if such exist. We demonstrate that together these algorithms are capable of discovering almost all of the true structure for a problem with 121 variables, including many of the directions. |
| |
Keywords: | Graphical models probabilistic expert systems machine learning Markov models causal structure |
本文献已被 SpringerLink 等数据库收录! |
|