Alzheimer’s Disease Diagnosis Using Machine Learning Approach

Alzheimer’s disease is the most common form of neurodegenerative disease found in the world today and it has shown no signs of curability to this day. Not only that, the causes of this disease are fairly unclear which makes the diagnosis and treatment to Alzheimer’s disease a very difficult task to accomplish. The aim of the research is to assist medical professionals in the early diagnosis of Alzheimer’s disease before it has fully metastasized and medical practices become useless. In the research a total of nine machine learning models were used which include standalone models as well as ensemble machine learning models to automate the process of diagnosis of this illness and compare the efficiency of each model. Each model uses the best parameters to make predictions which revealed that the employed classification model using random forest performed the best among all the other models. The best parameters for each model were automatically set by employing loops and conditional statements. The results revealed that the accuracy of the random forest classification model was same as AdaBoost ensemble model however, its overall performance was better than all the other models employed, with the highest accuracy percentage of 84.2105% and an AUC score of 84.4444%.