Automatic Documentation and Mathematical Linguistics. Vol. 50, No. 2, Pages 70-78, 2016

A. S. Kleshchev ^a, *, M. V. Petryaeva ^a, **, S. V. Smagin ^b, **, and M. Yu. Chernyakhovskaya ^a

^a Intelligent Systems Laboratory, Institute of Automation and Control Processes, Far East Branch, Russian Academy of Sciences, Vladivostok, Russia

^b Department of Applied Mathematics, Mechanics, Control and Software, School of Natural Sciences, Far Eastern Federal University, Vladivostok, Russky Island, Russia

Correspondence to:
*e-mail: kleschev@iacp.dvo.ru
**e-mail: margaret@iacp.dvo.ru
**e-mail: smagin.sv@dvfu.ru

Abstract—This paper proposes a problem-oriented method for the objective formation of easily interpretable knowledge databases for intelligent systems. We describe the InForMedKB software complex, which is designed for the inductive formation of medical diagnostics knowledge databases; it was used to perform the proposed method. Expert analysis of the results of using the developed software complex, viz., the inductively formed Acute Appendicitis database of medical diagnostic knowledge for a mathematical dependence model with parameters, which is near real-life ontology of medical diagnostics, is given.

Keywords: inductive formation of knowledge, machine learning, dependence model with parameters, easily interpretable dependence model, ontology for medical diagnostics, database of medical diagnostic knowledge, classification problem, learning algorithm, training set, software complex

An Analysis of the Results of the Inductive Formation of Diagnostic Medical Knowledge Databases