Dependence between cognitive impairment and metabolic syndrome applied to a Brazilian elderly dataset
作者：
Highlights：
• A Bayesian Network (BN) for modeling a medical dataset of Brazilian elderly is proposed.
• The learned BN structure is inspected for uncovering dependence relationships among focus variables.
• An indirect association between Metabolic Syndrome (MetS) and Cognitive Impairment (CI) is found.
• The probabilistic dependence between the focus nodes MetS and CI is conditional to both Body Mass Index (BMI) and Age.
• The role of MetS (and its components) in elderly cognition can non-trivially change over different Weight (or BMI) and Age ranges.
摘要
•A Bayesian Network (BN) for modeling a medical dataset of Brazilian elderly is proposed.•The learned BN structure is inspected for uncovering dependence relationships among focus variables.•An indirect association between Metabolic Syndrome (MetS) and Cognitive Impairment (CI) is found.•The probabilistic dependence between the focus nodes MetS and CI is conditional to both Body Mass Index (BMI) and Age.•The role of MetS (and its components) in elderly cognition can non-trivially change over different Weight (or BMI) and Age ranges.
论文关键词：Population aging,Cognitive impairment,Risk factors,Metabolic syndrome,Associations discovery,Bayesian network,Directed Acyclic Graph,D-separation
论文评审过程：Received 4 January 2018, Revised 12 July 2018, Accepted 22 July 2018, Available online 31 July 2018, Version of Record 23 August 2018.
论文官网地址：https://doi.org/10.1016/j.artmed.2018.07.003
