Explanatory and Predictive Modelling in the Study of Overweight and Obesity: The Example of Health Behaviour Among Estonian Adult Population

Toomas Gross
The use of machine learning models has become increasingly popular in the study overweight and obesity, complementing research on these topics by means of “traditional” statistical methods. Motivated by this methodological shift as well as the idiosyncrasies of the Estonian context, this thesis has three aims. Building on the data from the 2020 Health Behaviour Among Estonian Adult Population Survey (n = 1,737), it firstly scrutinises the possible associations between being overweight (BMI ≥ 25.0) or obese (BMI ≥ 30.0) and various socio-demographic and behavioural variables through explanatory modelling, using binary logistic regression analysis. Secondly, it compares the performance of various commonly used machine learning algorithms for classification problems when predicting overweight and obesity, respectively. And thirdly, the thesis discusses the advantages and limitations of explanatory and predictive modelling more generally and in the study of overweight and obesity more specifically, entering into dialogue with various other studies that have used these two approaches to the topic.
Graduation Thesis language
Graduation Thesis type
Master - Conversion Master in IT
Rajesh Sharma
Defence year