Informatics for dementia detection using statistical inference

Dementia is the most common degenerative disease, majorly affecting the elderly persons. The early diagnosis can prevent or delay its progression. This chronic and degenerative disease affects millions of people worldwide. This severely marked degenerative illness is frequently linked to dementia. The common symptoms of dementia are mood swings, difficulty concentrating, memory loss and confusion. In this paper, statistical based technique is used to identify essential features from the benchmark dataset i.e. open access series of imaging studies. This analysis reveals that the essential clinical features of dementia are mini-mental state examination, clinical dementia rating. Whereas, anatomical structure like estimated intracranial volume, normalized whole volume and atlas scaling factor. These features follows a normal distribution. In addition to that, these features are pass to different machine learning techniques, which are used to detect dementia. The performance is measured in terms true positive rate, false positive rate, precision, recall and F-measures has been improved and obtained the accuracy of 89.30%. The attained results seem to be more robust and precise, as compared to existing techniques.