Early Detection of Alzheimer's Disease Using Various Machine Learning Techniques: A Comparative Study
Aakash Shah, Dhruvi Lalakiya, Shekha Desai, Shreya Shreya, Vibha Patel
Abstract
Alzheimer's disease (AD) is an incurable, progressive neurodegenerative disease, which leads to the loss of memory. Various psychological tests used by doctors to diagnose patients, albeit, there is no well-defined procedure to detect AD in its early stage. This paper comprises of exhaustive analysis and accuracy of various machine learning techniques on a combination of different biomarkers associated with the disease. Since no prior research has suggested the use of Voting Classifier Algorithm for early detection of AD, the present study attempts to crystallize the intrusion of the averaging factor through Soft Voting Classifier, which influences the accuracy of the output in a way, that it removes the possibility of inaccuracies in the result by averaging the probable outcomes of the classifiers in the previous stage, giving the best possible result and making it more precise. A performance gain of 86% is obtained using Soft Voting classifier, wherein 437 cases are analyzed in our dataset, eliminating all the ineffective entries.