Istraživanje čimbenika rizika za nastanak alergijskih dišnih bolesti s pomoću Bayesovih mreža temeljenih na podacima iz više izvora [Exploring the risk factors for the development of allergic respiratory diseases by Bayesian networks built on multiple-source data]
Year: 2017
Type: dissertation
Abstract: It is suggested that multi-source data analysis could help to elucidate the aetiology \nof non-communicable, chronic diseases. However, statistical methods for \nmulti-source data analysis with partially overlapping data sources are not developed \nenough. \nThe study proposes new method for meta-analysis of such data, G–MA, applying \nit to Bayesian network of multi-source data. In simulation experiments, G-MA \nshowed lower mean squared error and narrower confidence intervals in parameter \nestimation, when compared to single data source analyses. \nBayesian network with G-MA method (Bayes–G–MA) was applied to analyse the \nfactors related to the presence of allergic respiratory diseases (ARD) in four Croatian \nepidemiological sources. Compared to single source analyses, Bayes–G–MA \nrecognised lower number of direct relationships with ARD (allergic diseases in childhood, \nallergic diseases in parents, body mass index), while others became indirect \n(gender, physical activity) or unrecognised (smoking). Bayes–G–MA suggested two \nnew relationships (with depression and alcohol consumption) between variables not \ntested together in any single source. These results could reflect true relationships, \nbut also lack of statistical power or intrinsic property of the method to under-report \npresent relationships. \nIn conclusion, the main contribution of proposed method is in the assessment of parameters, with expert help needed in evaluation of causal relationships.
Author: Kovačić, Jelena
Institution: University of Zagreb
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2963431971
