應用文字探勘探索痛經婦女之疾病經驗

Using text mining techniques to explore the illness experience among women with dysmenorrhea

病人經驗在醫療和護理過程當中都是相當重要的依據，若能善用病人口語描述自己的症狀及治療，有助於提供更為適切的護理介入。本研究目的為探討文字探勘技術分析痛經婦女疾病經驗之應用，發掘高頻詞彙、相關詞彙及區辨詞彙來呈現病人經驗，研究對象為40位接受中醫治療的痛經婦女，以疾病經驗文本為分析資料，使用R軟體為程式撰寫平台，利用文字探勘技術找出痛經婦女高頻詞彙及相關詞彙，並以「邏輯斯迴歸」、「支持向量機」、「分類迴歸樹」技術，測試是否能利用口語詞彙來區分婦女之症狀嚴重度高低。應用文字探勘技術可探索病人的口述疾病與治療經驗，有別於常見的問卷調查，也提出處理質性資料的另一種可能。 Patient experience is a crucial reference for both the medical and nursing care processes. Utilization of patients' oral descriptions of their own symptoms and treatment enables nursing professionals to provide adequate nursing intervention. To Examine the application of text mining on analyzing women with dysmenorrhea, and present patient experiences through vocabularies used by the patients, namely, frequently-used, related, and discriminative phrases. Forty women with dysmenorrhea who were treated with traditional Chinese medicine were recruited as participants. Using R statistical software, their illness experiences in textual form were analyzed using text mining to determine related and frequently used vocabularies of these women. In addition, logistic regression, support vector machine, and classification and regression tree were conducted to determine whether these women's illness severity can be differentiated from the vocabularies they used. The application of text mining for examining the verbal descriptions of patients' illness and treatment experiences is different from the commonly used questionnaire method. Text mining provides another option for qualitative data analysis.