Facilitating health care in remote and rural area using mobile phone infrastructure.
Prediction of disease based on data sends by patient using machine learning (Support Vector Machine).
Characterise the proposed model with different type of kernels and parameters to find optimum classification.
Mobile phone applications have been widely used in various fields, including health care. Generally, this technology is used to overcome problems in health care by utilising mobile phone features for facilitating basic needs in health services. This study proposes an intelligent mobile health monitoring system that can be used in rural and remote areas where health services are still lacking. The system was made based on client/server architecture. Nine symptoms of typhoid, cough and diarrhoea from 30 patients were gathered from a hospital. Based on this data, a machine learning model using Support Vector Machine (SVM) was performed to distinguish these diseases. To find the best model parameters of the SVM, three different kernels (linear, polynomial, and Radial Basis Function (RBF)) were analysed. The result showed that RBF with degree 2 provided the best result in this particular application. The system was designed to receive input from patients about symptoms of the disease they have. The mobile phone application sends the data of the symptoms using Short Message Service (SMS) to the server. Furthermore, a machine algorithm module in the server identifies to which disease it belongs to based on the machine learning model created before. The prediction result is accessible to the doctor and the nearest Community Health Center (CHC). Based on the result, the doctor proposes a treatment plan for the patient to be recorded and sent to the patient by CHC. The proposed mobile health monitoring system has run properly and is ready to be evaluated in a real situation.Le texte complet de cet article est disponible en PDF.
Keywords : Telehealth, ICT, Health care, Kernels