Back to search

University of Thessaly (UTH)

Applications of machine learning algorithms in quality of life

Abstract

dc:description

According to World Health Organization, a recent analysis showed that 1.71 billion people globally have musculoskeletal conditions. The societal impact in terms of direct healthcare costs and indirect (i.e., productivity loss) costs is enormous. Hence, it is vital to understand the pathophysiology of musculoskeletal diseases using artificial intelligence analytics tools with ultimate objective to develop techniques for their interpretation, diagnosis, prediction and rehabilitation. The aim of this thesis is to extend the current understanding of the contribution of the risk factors in the development of Knee Osteoarthritis and to uncover the rationale behind the biomechanical parameters from the anterior cruciate ligament post-surgery rehabilitation in order to avoid the outset of KOA. To achieve these goals, first of all we conducted a review about the machine learning techniques in knee osteoarthritis. Subsequently, we employed data from the osteoarthritis initiative (OAI) database (available on https://nda.nih.gov/oai/) and collected numerous biomechanical data from individuals who suffered from anterior cruciate ligament injury or not. This work led to five studies which are presented as different chapters of the current thesis. The review guided us to understand the literature gap and to develop machine learning techniques related to the prognosis and diagnosis of knee osteoarthritis as well as the interpretation of these models. In the second work, we used data from OAI and we worked on the prediction of KOA through the identification of risk factors that are relevant with KL progression. One of the main objectives of this work was to explore three different options with respect to the time period within which data should be considered in order to reliably predict KOA progression. The findings of this work were the input for the third work. So, the next step in the prediction task was to apply an evolutionary genetic algorithm (GA)-based wrapper technique, which leads to selected features that consistently work well at any possible data sample and, thus, have increased generalization capacity with respect to KOA progression. The impact of the selected risk factors on the prediction output was further investigated using SHapley Additive exPlanations (SHAP). The fourth work focused on the diagnosis task and interpretation of the model output. The objective of the present study was to provide a robust feature selection methodology based on fuzzy logic that could: (i) handle the multidimensional nature of the available datasets (OAI) and (ii) alleviate the defectiveness of existing feature selection techniques towards the identification of important risk factors which contribute to KOA diagnosis and interpretation. The fifth work has the aim to investigate the modification of the biomechanical parameters after an ACL injury, which is a risk factor for the onset of KOA. For this aim, a state-of-the-art explainability analysis based on SHAP and conventional statistical analysis attempted to uncover the rationale behind the decision-making mechanism of the best trained model and provide a holistic approach of quantifying the contribution of the input gait biomechanical parameters in the tasks of ACL injury diagnosis. The proposed AI methodologies may contribute to the development of new, efficient risk stratification strategies and identification of risk phenotypes of each KOA patient to enable appropriate interventions. Furthermore, features, that would have been neglected by the traditional statistical analysis, were identified as contributing parameters having significant impact on the ML model’s output for prediction of KOA progression, KOA diagnosis, ACL injury diagnosis during gait.

Degree

thesis:*
Grantor dc:publisher
University of Thessaly (UTH)
Year dc:date
2021

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Kokkotis, Christos
  • Κοκκότης, Χρήστος

Subjects

dc:subject × 20

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/50554

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Kokkotis, Christos; Κοκκότης, Χρήστος. Applications of machine learning algorithms in quality of life. University of Thessaly (UTH), 2021. http://hdl.handle.net/10442/hedi/50554