Abstract

The inability of the current techniques to identify patients in the preclinical stage of Alzheimer's disease, which can persist up to ten years before clinical symptoms appear, makes early detection of the condition difficult. Several studies have shown the potential of the cerebrospinal fluid biomarkers amyloid beta 1-42, T-tau, and P-tau in the early stages of Alzheimer's disease. Based on the levels of these cerebrospinal fluid biomarkers, we employed machine learning models in this study to categorize various phases of Alzheimer's disease. The National Alzheimer's Coordination Centre database of 537 patients' electronic health records was examined, and the patients were separated into groups based on their mini-mental state scores and cognitive dementia ratings. To find significant differences between the Alzheimer's stages, statistical and correlation studies were conducted. Then, to categorize the stages of Alzheimer's disease, machine learning classifiers such as KNearest Neighbor, Ensemble Boosted Tree, Support Vector Machine, Logistic Regression, and Naive Byes classifiers were used. According to the results, Ensemble Boosted Tree has the highest accuracy for binary classification, while Ensemble Bagged Tree offers superior accuracy for multiclassification. The results of this study should assist doctors in making an informed choice regarding the early diagnosis of Alzheimer's disease based solely on cerebrospinal fluid biomarkers, monitoring the illness's course, and putting the right intervention measures in place.

Date of publication

Spring 4-25-2023

Document Type

Thesis

Language

english

Persistent identifier

http://hdl.handle.net/10950/4185

Committee members

Dr. Shawana Tabassum, Dr. Premananda Indic, Dr. Regan Beckham

Degree

Masters in Electrical Engineering

Available for download on Thursday, April 24, 2025

Share

COinS