Improvement of the Personalized Mobile U-Health Service System

There are many problems with applying the machine learning technique, which is widely used in the conventional healthcare field, during the mobile u-health service analysis step. First, research on the mobile u-health service is just beginning, and there are very few cases where the existing techniques have been applied in the mobile u-health service environment. Second, since the machine learning technique requires a long learning period, it is not suitable for application in the mobile u-health service environment, which requires real-time disease management. Third, the various machine learning techniques that have been proposed until now do not include a way to assign the weight factors to the disease-related variables, and thus its use as a personalized disease prediction system is somewhat limited. This paper proposes PCADP, which is an ontology-based personalized disease prediction method, to solve such problems and to interpret the bio data analysis of the mobile u-health service system as a process. Moreover, the mobile u-health service ontology framework was modeled as a semantics type in order to meaningfully express the mobile u-health data and service statement based on PCADP. To validate the performance and efficiency of the PCADP technique proposed in this paper, the 5-cross validation method was used to measure the accuracy of the prediction. The validation of PCADP using a virtual disease group verified that the technique proposed in this paper shows much greater accuracy compared to existing methods. Moreover, the PCADP prediction method improved the flexibility and real-time attributes, which are the essential elements of any diagnosis technique in the mobile u-health environment, and showed efficiency in the continuous improvement of the monitoring and system of the diagnosis process.