Detection of Abnormal Behaviors in Patients with Dementia and Preliminary Symptoms in Smart Home

The number of elderly people who need help in their daily routines is increasing rapidly. Dementia is one of the most important causes of disability in elderly people and its outbreak has been a major burden on human societies. The purpose of this research was using intelligent home technology to monitor elderly behaviors, identify abnormal behaviors, and discover the initial signs of dementia before the onset of the disease. Early diagnosis of dementia at an early stage can lead to a high improvement in its treatment and delay the disease. In this applied, descriptive-analytic study, the abnormal behavior and early symptoms of dementia were identified using machine learning techniques.  The kmedoide algorithm was used to analyze abnormal behaviors and to assess the quality of sleep as the primary symptoms of dementia, the valid PSQI questionnaire was used. Matlab 2012 was used for implementation. The results in the abnormal behavioral section indicated that clustering algorithms have high efficacy in detecting abnormal behavior in smart home, and also results in early symptom examinations led to poor sleep recognition in the PSQI as a primary symptom of dementia. The behavior of the elderly, their abnormal behavior and early signs of diseases such as dementia can be recognized using the technology of the system under the supervision of the smart home.