Detection of Dementia using Machine Learning Algorithms

In recent times, dementia, particularly Alzheimer's disease, has become a major global health issue. The World Health Organisation reports an alarming 55 million individuals living with dementia worldwide, with an additional 10 million new cases identified annually. Dementia, a group of cognitive disorders affecting fundamental abilities like thinking, remembering, and reasoning, is primarily attributed to the destruction of brain cells, exerting a profound impact on brain function. Alzheimer's disease, representing 60-80 percent of all dementia cases, is a progressive neurological condition, while vascular dementia, the second most prevalent form, results from microscopic brain hemorrhages and artery blockages. The progression of Alzheimer's disease disrupts memory and cognitive functions. This study aims to contribute to the ongoing discourse by comparing the efficacy of machine learning algorithms in forecasting dementia incidence over two years against established models. The objective is to identify the most effective analytical techniques and determine the optimal number of predictive components for such forecasts. It is noteworthy that dementia-like symptoms may arise from treatable conditions, such as thyroid issues and vitamin shortages, underscoring the critical importance of early diagnosis and intervention. The significant challenges posed by dementia, particularly for elderly individuals struggling to manage their emotions, highlight the urgency for effective treatments. In this investigation, the accuracy of various machine-learning methods is assessed using a Kaggle dataset dedicated to dementia research. This analysis aims to contribute to the ongoing efforts aimed at enhancing dementia prediction and early diagnosis. The ultimate goal is to improve the well-being of individuals grappling with this debilitating affliction. The study reveals promising accuracy levels with Random Forest, GaussianNb, SVM, and AdaBoost, all demonstrating a high accuracy of 0.99, reinforcing the potential of machine learning in advancing our understanding and management of dementia.