A Machine Learning Approach for Predicting Deterioration in Alzheimer's Disease

- This paper explores deterioration in In lieu of effective treatment for Alzheimer's disease [5], Alzheimer's Disease using Machine Learning. Subjects research has turned towards the possibility of early were split into two datasets based on baseline diagnosis detection of Alzheimer's biomarkers before the onset of (Cognitively Normal, Mild Cognitive Impairment), symptoms [6]. Work in this area has noted that accurate with outcome of deterioration at final visit (a binomial prediction of the risk of developing Alzheimer's disease essentially yes/no categorisation) using data from the and subsequent early interventions aimed at delaying the Alzheimer's Disease Neuroimaging Initiative onset of symptoms would ease the burden of suffering as (demographics, genetics, CSF, imaging, and well as financial costs of care for the patient and their neuropsychological testing etc). Indeed, research has suggested that the 100 iterations. We were able to demonstrate good structural brain changes that precipitate Alzheimer's predictive ability using CART predicting which of those symptoms may begin several years before the onset of in the cognitively normal group deteriorated and notable symptoms [6], and this may provide an opportunity received a worse diagnosis (AUC = 0.88).