![]() | Only 14 pages are availabe for public view |
Abstract There is a big demand for healthcare services to monitor health issues and help people to detect diseases without a need for visiting a doctor every single day. A pandemic issue such as Corona Co-vid 19 pushed millions of people to visit healthcare providers to check up if they are infected or not. So a wearable device with multi health sensors with tracking the user physical parameters can help to decrease the pressure on the healthcare providers such as clinics, hospitals, etc. Using healthcare data which can be collected in a long time for every user as profile data from healthcare providers, can help in saving lives, times, and uncertainty. This thesis studies strategies for disease detection, and prediction. The thesis gives valuable knowledge also about relevant solutions. We present a disease detection and prediction system. The proposed system used a fusion of multi-sensors mounted on a wearable device to monitor the user’s medical streaming data to make on running time medical decisions. The proposed system used a biosensor fusion technique to increase the probability of diseased detection. The system used a decision level fusion to decrease false alarm of disease detection. It used a hybrid prediction algorithm to predict diseases based on user profile data, EHR data, and streaming data coming from wearable devices. We proposed an effective design for the wearable device based on sending alerts only on-demand to send only the detections of disorder of sensor readings to the cloud platform. We proposed an alerting system to help the user to be aware of potentialviii diseases. The proposed wearable device is light in order not to disturb the users. We proposed an effective method to detect dangerous situations which can lead to the death of the user like blood bleeding. The purpose of the thesis is to give an efficient solution for the prediction of the disease and as a helpful tool for the user to predict diseases on a cloud-based system. By using wearable devices that are capable of detecting disorder readings from sensors. The information systems concerned with the systems which collect the available information from different sources to support the final decision. So, the proposed system worked on the integration of multi-sensors data then decides whether there is a disease or not. The proposed system for smart homes using proposed speech recognition proposed an efficient solution with an accuracy rate of 97%. The results of the proposed system for heart disease prediction using with an accuracy of 90.6%. The proposed recommender system for breast cancer prediction using achieved an accuracy of 98.8% for the multi-class random forest, and 99.4% using multi-class neural network. |