Design And Implementation Of Content Server Software Based On Big Data Processing

With the popularity of wearable medical devices, more and more people begin to pay attention to personal health problems. These devices work every time and collect a large number of physiological data. As time goes by, more and more data bring a severe challenge to the data storage and query system. At the same time, people are eager to get more than just the personal physiological data collected by wearable devices, but also need to get the really valuable information hidden behind these data. This thesis designs and implements a content server software for large data processing, which is designed to mine users’ physiological data, and the mining results will be transformed into health medical information in the form the user can understand.After fully studying the relative theories of the system. Combined with the actual situation of healthcare system, universal data acquisition interface is designed according to the actual situation of the healthcare system, database tables are established and setup the relationship between tables according to data types. This thesis designs a user physiological data management system, through the system user can view real-time physiological condition information. Then according to the authoritative medical guidance, designs a physiological data mining system, the system uses Hadoop as a distributed computing platform and uses Mahout as machine learning tool. The association rule mining algorithm FP-Growth helps to find user physiological characteristic according to the variation of physiological data with time. Finally according to the need of personalized healthcare, relying solely on single measurement data cannot provide useful information for the user. At last, according to concrete results of user physiological data mining, this thesis designs a healthcare information content recommendation system, the thought of content recommendation system derives from search engine, content recommendation system consists several parts, such as the content grasping part, the content preprocession part and the content query part. Content recommendation system provides users with pertinent health information.In this thesis, physiological data is transformed into healthcare information user can understand easily. After the system is further completed, it can be used in the aspect of physiological information monitoring, healthcare of sub-healthy population and rehabilitation of patients. thus it has an important significance to the improvement of physical fitness and disease prevention.