Logical aspects of experimental design and analysis
Предложена схема формализации задач активной идентификации объекта с использованием аппарата теории моделей - современного раздела математической логики. Теория моделей позволяет погрузить предмет “планирование и анализ эксперимента” в контекст семантического анализа. Семантический анализ понимается как установление соответствия между миром и его формальным представлением. С этой точки зрения представления об исследуемом объекте выражаются в некоторой прикладной теории. Предложен вывод модели для данной теории как процесс интерпретации, в котором ключевая роль отводится “экспериментатору”. Полученные результаты могут быть использованы при проектировании архитектур интеллектуальных систем для экспериментальных исследований, для построения онтологии эксперимента, создания баз знаний Purpose. The purpose of this work is to formalize the tasks of active object identification based on the apparatus of model theory - a modern section of mathematical logic. Model theory allows putting the subject “planning and analysis of an experiment” in the context of semantic analysis. Semantic analysis is understood as establishing a correspondence between the world and its formal representation. From this point of view, the concept of the object under study is expressed in some applied theory, which allows applying formal methods of model theory to it. Methods. It is assumed that the model is derived for this theory as an interpretation process, in which the key role is assigned to the experimenter. As a research method, it is proposed to use commutative diagrams that reflect the process of interpretation and extension of communication diagrams for the so-called equipped theories of planning and analysis of experiments. Results. The properties of the proposed models are proved and examples for planning a regression experiment are presented as an illustration. It is proved that for linear models it is possible to construct a finitely axiomatization capable theory. Findings, originality. The obtained results can be used in the design of architectures for an intelligent system in experimental research, building an experiment ontology and creation of knowledge bases. These studies will allow using logical programming to implement images of the presented commutative diagrams for equipped theories as applied systems for planning and interpreting the experiment