近年來，心電圖應用的領域範圍相當廣泛，但是對於接收數據後沒有建立起一套篩選標準，導致研究者需要先判斷心電圖訊號品質才能繼續後續分析。本研究為了改善以往只能使用肉眼判斷訊號品質，或自訂閥值之評估方法，故提出統計指標配合決策樹篩選。本研究實驗數據使用Physionet 生理訊號資料庫共 200 筆，以及感測器與模擬器收集之訊號共 238 筆，在本研究所獲得的準確率分別達95%、93.24%，而先前研究所使用的數據庫，透過本研究方法獲得的準確率為83%。本研究不受感測儀器以及受測者之個體差異，皆能保有良好的準確度，透過高準確度的訊號品質評估，可有效地減少心電圖訊號品質評估人員的負擔。;Electrocardiogram is likely to find very extensive use in many application areas in recent years, however, there is no standard electrocardiographic criteria in screening after collecting the ECG data, which leads researchers have to proceed quality estimation of the ECG signal before interpreting the analysis. For improving the traditional method of signal quality estimation by the naked eye or the improved threshold, are used as statistical indicators to identify with decision tree analysis screening. A total of 200 retrieved data through Physionet database, and a total of 238 collected signals by ECG sensors and simulators are analyzed in this study. A comparison between this study and previous studies’ database in the accuracy was 95% and 93.24%, and 83%, respectively. Therefore, this study result is not influenced by the devices and individual differences and both can maintain good accuracy. Through the high level of signal accuracy quality assessment, the burden for the ECG researchers can be efficiently reduced.
