Incremental Learning for Alzheimer's Disease on Medical Cloud Service Environment

Abstract

According to previous survey data, there estimated about 150,000 people suffering from Alzheimer's disease in Taiwan, in which over 65 years old population’s prevalence rate is 4 to 5% and the prevalence rate doubled every increasing in the age of 5. For population of 85 years of age and older, half to one-third of them have the chance of suffering from Alzheimer's disease. If the disease can be accurately predicted before it occurred, or if it can be correctly classified after the occurrence of disease then it will be of great help on the treatment or prevention of this disease. In this paper, the cloud care environment for incremental learning will be considered. The experimental results show that in the case of dynamic learning, the ensemble model with multiple classifiers can achieve more than 80% accuracy, and the results of this study will be applied to hospitals for further modeling for other interested target disease.