Modeling of a directed acyclic graph for causal inference

DOI: 10.31673/2518-7678.2023.020202

Abstract

The paper proposes an algorithm for causal inference based on a set of input data with compliance with the proposed restrictions. This article is presents assumptions about the data set that affect the choice and accuracy of the causal method. The algorithm for constructing the structure of a directed acyclic graph is given. The adequacy of the algorithm was checked on a test mathematical model, which allowed the analysis to be carried out without a randomized experiment. The proposed algorithm allows extrapolation to reveal a causal model with specified assumptions and the possibility of stricter restrictions on the input data set.

Keywords: causal inference, causal graph, causal relationships, data set, modeling causal inference.