Identification of Character Traits Associated with Successful Weight Loss
Name
Eerik Sven Puudist
Abstract
This study examines the relationship of personality traits with changes in body mass index (BMI) using a more detailed personality measurement questionnaire and a larger sample (N > 45000) compared to previous studies.
Using regression models (linear models, random forest, XGBoost), Gaussian mixture models, and factor analysis, we examine the relationships of personality traits with a person's largest historical BMI, largest BMI decline, and BMI regain. The goal is to find which character traits influence those outcomes the most to assist in developing personalized weight loss treatment plans.
We point out that BMI is related to almost all aspects of one's personality. Those relationships seem to be linear. However, the factors predicting the change in BMI are different between genders, and therefore both genders should be addressed separately in the development of weight loss methodologies.
Using regression models (linear models, random forest, XGBoost), Gaussian mixture models, and factor analysis, we examine the relationships of personality traits with a person's largest historical BMI, largest BMI decline, and BMI regain. The goal is to find which character traits influence those outcomes the most to assist in developing personalized weight loss treatment plans.
We point out that BMI is related to almost all aspects of one's personality. Those relationships seem to be linear. However, the factors predicting the change in BMI are different between genders, and therefore both genders should be addressed separately in the development of weight loss methodologies.
Graduation Thesis language
Estonian
Graduation Thesis type
Master - Data Science
Supervisor(s)
Uku Vainik, Raivo Kolde, Kadri Arumäe
Defence year
2024