건강검진 데이터와 마코프 체인을 이용한 만성질환 상태전이 과정 분석

Abstract

This study analyzes what state a specific group of people exist at with respect to chronic diseases and their risk factors, and the movement between states. Moreover, it also predicts what state people in the group will belong to in the future for the chronic disease. Types of chronic diseases to be analyzed are hepatitis and hypertension. For this purpose, a discrete time Markov chain model is constructed using the medical checkup data related to hepatitis and hypertension. Through the medical checkup data, the states composed of the observed chronic diseases and related effective variables are defined, and the state space is configured. The transition probability matrix is obtained, and the Markov property and time homogeneity are verified to show the validity of the modeled Markov chain. The analysis of the steady state is performed for the Markov chain where time homogeneity is applied, and subsequently, through the analysis of transition probability, which state of people are most likely to suffer from the symptoms of chronic diseases, and which state of people are to escape from chronic diseases are analyzed.