A Survey of Machine Learning Techniques for Prediction of Alzheimer’s Disease

The impact on daily life caused by Alzheimer's Disease cognitive functions is of significant importance. It is projected that by the year 2060, there may be a surge in the number of individuals affected by Alzheimer's Disease, with estimates indicating that it could reach 14 million, whereas, in 2020, approx 5.8 million people in the United States were already affected. Typically, symptoms manifest after the age of 60 and become more severe as one age. Early prediction and detection of Alzheimer's is crucial for both care and prevention purposes. Various Machine Learning methods, such as Support Vector Machine, K-Nearest Neighbour, and Random Forest, contribute to the detection of Alzheimer's by utilizing neuroimaging biomarkers obtained from Positron Emission Tomography and Magnetic Resonance Imaging. Nevertheless, the handling of high-dimensional data is still a challenge for traditional machine learning algorithms. The strength of Machine learning lies in its ability to analyze multi-dimensional datasets without requiring extensive pre-processing, thus providing a promising avenue for Alzheimer's detection. The review of existing literature underscores the significance of early detection of Alzheimer's. Particularly, studies utilizing Diffusion Tensor Imaging, Positron Emission Tomography, and Fluid-Attenuated Inversion Recovery techniques show potential in distinguishing different stages of the disease. Methods for extracting useful features demonstrate effectiveness, particularly in the context of binary classification. This paper underlines the previous work done while utilizing Machine Learning techniques for the precise categorization and prediction of Alzheimer's disease. The approaches exhibit superior accuracy and require smaller datasets compared to manual feature extraction methods. A good value of accuracy is achieved by using different techniques applied to different datasets. The objective is to analyze and compare the existing techniques and methods used for the precision of Alzheimer's disease prediction.