The data processing and decision support in healthcare systems

Abstract

The article describes the characteristics of diagnostic and management tasks solved in healthcare information systems. We propose an approach based on the use of ensemble models (variant modeling) to solve problems of decision support. The variant modeling is a method that applies the set of models for solving one task. The models can have different structures and use different input features. Also, we propose the method called “Modeleteka” for structure store and automatically select models to solve tasks in healthcare information systems. This method allows append new models and train old from ensemble in the maintenance of information system.