Utilization of Dark Data from Electronic Health Records for the Early Detection of Alzheimer’s Disease

Abstract

Information management and the presence of dark data is the most alarming and vulnerable topic that needs to be addressed amidst the rapid growth of technology. The unused, untapped, or unstructured data in our archives can be called as dark data. Under the HIPPA regulations, the clinical data needs to be stored and secured for years, so the amount of dark data in healthcare databases consumes a large amount of storage space. The paper shows the comparative performance of the random forest algorithm with and without dark data from the patient’s health record for the early detection of Alzheimer’s disease. The model executes in two ways, (1) considering only the Alzheimer’s disease parameters and (2) including diabetes disease parameters which are considered as dark data for the Alzheimer’s disease. The results of the research work clearly say that utilizing the health dark data has increased the accuracy by 16.3% and hence helps in making better decisions.