Disease Classification and Recommendation System using CNN and Geo-demographic Data

This research proposes a CNN-enabled framework, that combines demographic and geographic information in a system to improve disease classification and ultimately provide personalized health recommendations. To classify diseases such as dengue, malaria, typhoid, and cholera, the model utilizes age, gender, the location represented in latitude and longitude, and seasonal aspects of disease. Further, the platform categorizes disease, as well as represents a disease prediction of risk in different geography based on time by capturing the temporal and spatial patterns of the disease. Part of the framework design is a recommendation engine that provides preventive and therapeutic recommendations based on the user's health, and community context. The platform is deployed as a web application and was designed to be scalable, quick to use, and friendly for public use, and intended for healthcare practitioners, and health policy-makers, as well as members of the public. Overall, this framework can also support proactive public health and planning by providing early warnings, planning and allocating limited resource effectively, and continued surveillance of disease, which would support preparedness planning for future events.