動脈粥樣硬化疾病伴隨中風之評估研究

A Study on the Assessment of Atherosclerosis Associated with Stroke

吳崇碩(Chong-Shuo Wu)

指導教授 ： 張俊郎

國立虎尾科技大學/管理學院/工業管理系工業工程與管理碩士班/碩士(2017年)

摘要

現代社會的醫療進步和衛生環境改變，使台灣社會的人口結構呈現高齡化社會，心血管疾病是導致高齡人口死亡的重要因素之一，其中中風是較多人聽聞，也嚴重的威脅國人健康的疾病。
　　本研究以國內某醫療機構資料庫中動脈粥樣硬化疾病為研究對象，並透過相關的文獻和醫師訪談，運用人工智慧之粒子群最佳化演算法、基因邏輯斯迴歸演算法、交叉熵演算法計算每個因子權重，並分別結合支援向量機與類神經網路以建構六種預測模型與三種案例式評估系統，用以評估罹患動脈粥樣硬化疾病的患者是否會產生中風，本研究將對醫師的臨床輔助診斷有實質的幫助。
研究結果顯示，雖以粒子群最佳化演算法結合案例式推理評估系統準確率最佳，經K疊驗證後，平均準確率為92.75%，ROC曲線下面積為0.888，經傅利曼檢定為無顯著差異，表示模型之間並無差異，每種模型皆可做為評估系統權重值計算；六種預測模型中，雖以粒子群最佳化演算法結合倒傳遞類神經網路最佳，K疊驗證後平均準確率為90.67%，ROC曲線下面積為0.874，經傅利曼檢定為無顯著差異，表示模型間不存在顯著差異性，因此皆可做為預測之計算。本研究結果能提供給醫療機構或臨床工作者做為輔助診斷之參考依據，達成早期發現早期治療，便能減輕病患之疾病負擔把握黃金就醫時間。

關鍵字

動脈粥樣硬化 ； 中風 ； 粒子群最佳化演算法 ； 基因邏輯斯迴歸演算法 ； 交叉熵演算法 ； 案例式推理

並列摘要

In the modern society, medical advancement and changes in the health environment have led to Taiwan’s “aging society” type of population structure. Cardiovascular disease is one of the important factors leading to the death of the elderly, and “stroke” is a more frequently heard disease that seriously affects the health of the people.
In this study, patients with atherosclerotic disease in the database of an anonymous medical institution in Taiwan were adopted as research participants. Through relevant literatures and interviews with physicians, using the algorithm of artificial intelligence such as particle swarm optimization algorithm, genetic logistic regression algorithm, and cross-entropy algorithm, the weight of each factor was computed. In addition, the support vector machine and neural network were conjunctively used to construct six predictive models and three case-based assessment systems in order to evaluate whether patients with atherosclerotic disease will develop stroke. This study will offer substantial help to physicians during clinical aided diagnosis.
Research results show that although the PSO combined with the case-based reasoning system has the best accuracy, after the k-fold verification, the average accuracy, ACC was 92.75% and area under the ROC curve, AURC was 0.888. The Friedman’s test shows no significant difference, thus indicating no difference exists among the models. Therefore, the models are all suitable for calculating the evaluation system weights. Among the six predictive models, although the PSO combined with the BPN are the best, after the k-fold verification, the average ACC was 90.67% and AURC was 0.874. The Friedman’s test shows no significant difference, thus indicating no difference exists among the models and that they can all be used in predictive calculations. The research results shall serve as a reference for medical institutions or clinical workers during aided diagnosis, thereby achieving “early detection and early treatment”, relieving patients of the burden of disease, and ensuring timely treatment during the golden time.

並列關鍵字

atherosclerosis ； stroke ； particle swarm optimization algorithm ； genetic logistic regression algorithm ； cross-entropy algorithm ； case-based reasoning

參考文獻

78. 羅兆寶、林永煬、郭萬祐(2016)，「急性缺血性中風經動脈移除血栓治療之現況」，臨床醫學月刊，78卷，4期，頁185-189。

68. 劉鵬達、魏大森(2016)，「腦中風之跌倒預防」，血管醫學防治季刊，26期，頁14-16。

50. 陳品良、楊秉哲、谷圳(2016)，「基於大數據的跨境電子商務分析架構－以商品口碑分類系統為例」，電工通訊季刊，2期，頁48-57。

18. 李建邦、俞星辰(2014)，「以粒子群最佳化演算法結合分類器預測疾病及搜尋變數之研究」，數據分析，9卷，1期，頁57-74。

31. 邵曰仁、黃于庭、顏慈儀(2008)，「應用類神經網路與支援向量機以辨認製程多重品質特性出錯之研究」，Journal of Data Analysis，3卷，5期，頁69-81。