A novel and proposed triad machine learning-based framework for the prognosis of Huntington’s disease

Autor: Deepak, Shashikant, Ojha, Ananta, Acharjya, Kalyan, Mishra, Reshmi, Gantra, Amit, Kalaiarasan, C., Upadhyay, Ramakant, Walke, Suchita
Zdroj: International Journal of Information Technology; April 2024, Vol. 16 Issue: 4 p2603-2610, 8p
Abstrakt: A hereditary condition known as Huntington's Disease (HD) results in the gradual degeneration of brain nerve cells, impairing a person's capacity for speech, walking, and reasoning. Owing to its severity, novel strategies are crucial for the creation of techniques that support the accurate categorization of this illness. This research proposes a tripartite machine learning-based framework that uses gait dynamics and temporal parameter extraction to predict the prognosis of Huntington's illness. Phases including data pre-processing, feature extraction, classification, and projected output are included in the suggested triad method. The proposed framework uses a triad combination of classifiers for classification: Support vector machines (SVM), K-nearest neighbor (KNN), and Naïve Bayes (NB). The suggested framework's effectiveness is assessed by comparing it to current machine learning classifiers using parameters from the public repository on the dynamics of gait in neurodegenerative diseases. The experimental results show that the triad classifiers employed in the proposed approach are able to achieve higher accuracy (85.45%), higher sensitivity (78.37%) and higher specificity (76.55%) than the existing classifiers used in the recent studies.
Databáze: Supplemental Index
Externí odkaz: