Publicou 12 edições por ano
ISSN Imprimir: 1064-2315
ISSN On-line: 2163-9337
Indexed in
Reconstruction of the Model of Probabilistic Dependences by Statistical Data. Tools and Algorithm
RESUMO
The tools and an algorithm for reconstruction of probability models of dependencies in the class of monoflow structures (a subclass of Bayesian networks), are developed. "Proliferator-D" algorithm is computationally efficient (subcubic complexity) and performs a small number of tests of conditional independence only of the first rank. The correctness of the algorithm is justified by simple assumptions, which are empirically robust with respect to the size of a data sample. When the generative model goes beyond monoflow structures, the algorithm gradually degrades to the known Kruskal algorithm and produces the cover (approximation) of the model by a tree. The proposed algorithm can be easily modified to improve the quality of reduction (approximation) of Bayesian networks.
-
Balabanov O. S., Induced Dependence, Factor Interaction, and Discriminating Between Causal Structures, Cybernetics and Systems Analysis, 52, 1, 2016. Crossref
-
Balabanov O.S., On the classes of causal networks, identifiable by simple independence tests, PROBLEMS IN PROGRAMMING, 2-3, 2018. Crossref
-
Balabanov O.S., Principles and analytical tools for reconstruction of probabilistic dependency structures in special class, PROBLEMS IN PROGRAMMING, 1, 2017. Crossref