FEED: An ECG data mining algorithm in WE-CARE system

The cardiovascular disease (CVD) is a serious social problem. For solving this problem, we developed WE-CARE system. The use of 7 lead wires helps collecting sufficient ECG data to guarantee the detection accuracy without impairing the mobility of the system 910. While the traditional algorithms face efficiency challenges for the large data, complicated algoirthm and small capacity of wearable devices. In this paper, we propose a brand new ECG online analyzing algorithm to solve the efficiency problem, called the FEED (Fast Emergency ECG Detection) algorithm. Firstly, FEED algorithm will proeprocess the raw ECG data and do basic feature extramction for it. Secondly, it conducts abnormal signal calissification with the data uploads it to servers for deep analysis. Thirdly, it retrains the classifier and updates parameters for it. Through this kind of design, FEED can achieve high efficiency. The FEED solution is embedded organically in all three layers of the WE-CARE system to alleviate the critical ´3V´ challenges - variability, volatility and value. Therefore, FEED can fully optimize the performance for WE-CARE system.