For early diagnosis of atrial fibrillation disease, the present invention proposes a smart device for capturing signal type of atrial fibrillation and diagnosis assistance, which is reliable and automatic, combined with evolutionary algorithms, machine learning, and computerized human-machine interface, and consists of a database, a classification model base and a user interface. Accordingly, by using the evolutionary algorithms to obtain three accurate characteristic parameters of P-wave signal type of Gaussian function of an electrocardiogram (ECG), using various different machine learning methods to establish a highly accurate ECG p-wave signal classification predicting model based on the collected P-wave signals on the ECG of patients with atrial fibrillation and P-wave signals on the ECG of persons with healthy and normal heart rhythm, using the computerized human-machine interface as a massage communication tool, the classification predicting model is used to subsequently provide a clinician with a reference for early diagnosis of the patient with atrial fibrillation, and further examinations and medications are provided to the patient so as to avoid occurrence of cardiovascular diseases or to suppress or reduce the deterioration of atrial fibrillation, thereby achieving the effects of improving the health care quality, promoting the therapeutic quality and lowering the expense of social cost.一種針對心房顫動疾病之早期診斷，提出一可靠且自動化並結合進化演算法、機器學習、及電腦化人機介面之心房顫動信號型態擷取及輔助診斷智能裝置，包括一資料庫、一分類模式庫、以及一使用介面所構成。藉此，利用進化演算法取得精確的心電圖之P波信號型態的高斯函數之三個特徵參數值，並將所蒐集之心房顫動病患心電圖與健康正常心律者心電圖之P波信號，應用各種不同機器學習方法，建立一高準確率之心電圖P波信號分類預測模型，再通過電腦人機介面作為訊息溝通工具，以此分類預測模型後續提供給臨床醫師早期診斷心房顫動病患之參考，針對此病患做更深入之檢查與用藥，使避免產生心血管疾病之發生，抑或降低心房顫動病患之惡化，達到增加健康照護品質、提升醫療質量，進而減少社會成本之支出。
