M. I. Zabezhailo, “To the computational complexity of hypotheses generation in JSM-method. Part II”, Artificial Intelligence and Decision Making, 2015, Issue 2,Pages <nobr>3

Methods of reasoning and knowledge representation

To the computational complexity of hypotheses generation in JSM-method. Part II

M. I. Zabezhailo

Applied Research Center for Computer Networks

Abstract: Some new possibilities to organize intelligent data analysis of different data types (sets, graphs, numerical data) by means of JSM-method of automated hypotheses generation are discussed. JSM-method is presented as a platform for intelligent data analysis of different types of data.
Combinatorial properties (computational complexity) of causal dependencies reconstruction from empirical data by JSM-tools are discussed. Stability of NP-completeness and PC-completeness of some JSM-combinatorial problems for different data types is demonstrated. Detailed analysis for computational complexity of numerical data analysis in the style of J.S. Mill’s method of concomitant variations is presented.

Keywords: JSM-method of automated dependencies reconstruction from empirical data, computational complexity and search optimization, causal dependencies in numerical data, approximate JSM-method.