Random generation of hypotheses in the JSM method using simple Markov chains

Authors:
D. W. Vinogradov
Affiliations:
All-Russia Institute of Scientific and Technical Information, Russian Academy of Sciences, Moscow, Russia 125190
Venue:
Automatic Documentation and Mathematical Linguistics
Year:
2012

Citing 3
Cited 0

Exact sampling with coupled Markov chains and applications to statistical mechanics

Proceedings of the seventh international conference on Random structures and algorithms
Comparison of approximated paths in search of similarities for the JSM method

Automatic Documentation and Mathematical Linguistics
Formal description of adaptation of intelligent JSM-systems for clinical and laboratory data analysis

Automatic Documentation and Mathematical Linguistics

Quantified Score

Hi-index	0.00

Visualization

Abstract

A probability approach to hypothesis generation in plaussible reasoning of the JSM type using special Markov chains is described. This approach is similar the Monte Carlo methods, which are actively used to calculate combinatory configurations and the volumes of concave bodies. Simplest Markov chains are discussed, including nonmonotonous, monotonous, and "coupling in the future." Their properties are proven and coupling is proven with a probability 1 *.