Information Synthesis Methods In Detection Of T-wave End Point

Information synthesis is the statistical methods to synthesize the different results from the researches with same purpose. Meta-Analysis is widely used in the field of clinical medicine, but it requires that these different results are independent, which limits its application. Reference [5] has used Bayesian methods to synthesize two inde-pendent results, this paper gives the Bayes method and EM method based on it. EM can combine any number of the results, and it is unnecessary to consider the correlation of the outcomes of different research.In clinical medicine, the detection of ECG T-wave end point is very important in study of heart disease. Now, there are many detection methods for it, so it will certainly improve the precision and reliability if these results can be integrated. This paper will use Meta-Analysis,Bayes and EM to detect the T-wave end point of ECG.The data of this paper is from QT database a public library. First we get the T-wave end points of Lead MLII and Lead V5based on slope change, then use these three information synthesis methods to integrate them to get the final result. Set the experts marking as comparison benchmark, we compare the integrated results with single-lead T-wave end point marks based on two different comparison principles. In addition, this paper improves information synthesis degree by adding the information from the T-wave end point detected by ECGpuwave software, then compare with the integrated result mentioned before to prove the flexibility and efficiency of these integrating multiple sources of information.