2026-05-16T04:05:54.316Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:11134 2025-02-06T16:53:18Z 3

Hybrid Model Optimization With Modified Metaheuristics for Parkinson’s Disease Detection 2024 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/11134 https://www.atlantis-press.com/proceedings/iciitb-24/126002407 V. Markovic A. Njegus L. Jovanovic T. Zivkovic D. Jovanovic D. Mladenovic Parkinson’s disease, a progressive neurological disorder primarily affecting elderly males, stems from dysregulation within the extrapyramidal tracts, notably the substantia nigra, lentiform nucleus, caudate nucleus, and ruber nucleus. This condition manifests as heightened cholinergic activity in the brain, correlating with cognitive decline, gait disturbances, sleep disorders, psychiatric symptoms, and olfactory dysfunction. Early detection is crucial for enhancing patient prognosis. Although neurological damage cannot be reversed, treatment can mitigate progression. However, patients often delay seeking treatment until symptoms significantly impair daily functioning, underscoring the importance of early detection. This study investigates the fusion of long short-term memory and extreme gradient boosting classifiers to develop an early detection system utilizing noninvasive shoe-mounted sensor data for observing patient gait. A tailored optimizer is introduced to enhance classification accuracy, achieving a notable accuracy of 0.896370, surpassing other contemporary optimizers in identical conditions. bookPart Atlantis Press 102 120 10.2991/978-94-6463-482-2_8 Proceedings of the 2nd International Conference on Innovation in Information Technology and Business (ICIITB 2024), Chapter in Advances in Computer Science Research