Author, Institution: Andrius Lauraitis, Kaunas University of Technology
Science area, field of science: Natural Sciences, Informatics N009
Scientific Supervisor: Prof. Dr. Rytis Maskeliūnas (Kaunas University of Technology, Technological Sciences, Informatics Engineering, T007 and Natural Sciences, Informatics, N009).
Dissertation Defence Board of Informatics Science Field:
Prof. Dr. Habil. Rimantas Barauskas (Kaunas University of Technology, Natural Sciences, Informatics, N009), chairman,
Prof. Dr. Habil. Gintautas Dzemyda (Vilnius University, Natural Sciences, Informatics, N009),
Prof. Dr. Vacius Jusas (Kaunas University of Technology, Natural Sciences, Informatics, N009),
Dr. Zbigniew Marszalek (Silesian University of Technology, Poland, Natural Sciences, Informatics, N009),
Prof. Dr. Alfonsas Misevičius (Kaunas University of Technology, Natural Sciences, Informatics, N009).
The doctoral dissertation is available on the internet and and at the libraries of Kaunas University of Technology (K. Donelaičio g. 20, Kaunas), Vytautas Magnus University (K. Donelaičio g. 52, Kaunas) and Vilnius Gediminas Technical University (Saulėtekio al. 14, 10223 Vilnius).
Dissertation addresses problematics related to the analysis and diagnosis of central nervous system disorders (CNSD) using mobile smart devices and Self-Administered Gerocognitive Exam (SAGE) testing methodology. Based on a priori scientific knowledge, an extension of SAGE methodology for detecting tremor, cognitive, speech and energy expenditure impairments for CNSD patients is proposed in this work. The practical significance of this dissertation has been evaluated by the implementation of Neural Impairment Test Suite (NITS) Android mobile application, which provides feedback on the patient’s health status and makes predictions on disease progression. Early stage Huntington’s, Parkinson’s, dementia, cerebral palsy CNSD patients and healthy test subjects not at risk group were involved in experimental research. Supervised learning classifiers were integrated in this dissertation and hybrid model was developed for combining single classifiers into an ensemble. Experiments were carried out for solving sick vs. healthy binary classification problem. Implemented classifier ensemble (hybrid model) results in ~2% increased accuracy, as compared to standalone models, i.e. 96.12% as best-fit combination based on the highest accuracy value and minimum time expenditure (0.59 seconds) for model training.