應用文字探勘技術探討戒菸學生之經驗

Abstract

目的：為探討有吸菸的高中職在學學生關於其戒菸經驗，並進一步瞭解是否有其他隱含的資訊尚未被發掘，透過應用文字探勘技術進行探討戒菸經驗，並比較高、低成癮組之高頻詞彙的異同。方法：本研究先藉由吸菸學生描述個人戒菸的經驗、感受及想法，透過訪談調查瞭解其戒菸經驗，而後運用文字探勘進行探討，共有60份文本資料，將其分為高成癮及低成癮組後，分別進行高頻詞彙、文字雲及其相關詞彙的探討。結果：尼古丁成癮群集分組方面，研究受訪60位有戒菸經驗的學生，在吸菸多久_(月)、吸菸量_(支/天)與尼古丁成癮總分方面高成癮都高於低成癮組，達統計上的顯著差異，在社會支持性總分、工具性、資訊性及評價性方面，低成癮組都高於高成癮組，達明顯統計上的顯著差異；在戒菸經驗的文字雲部分，取詞頻前30個繪圖成文字雲，並將其高頻詞彙進行屬性歸納，而高成癮組獨有與健康方面有關的高頻詞彙居多，如「咳嗽」、「呼吸」及「尼古丁」等，其他方面如「忍耐」、「誘惑」、「意志力」等，情緒以負面為主如「壓力」等，低成癮組則以金錢詞彙居多，如「省錢」等；高、低成癮組之相關詞彙方面，高成癮組的相關詞彙仍與健康方面較為常見，如「很喘」、「二手菸」及「呼吸困難」等，其他方面則有「習慣」、「下意識」、「聚在一起」等詞彙，情緒方面仍以「壓力」及「暴躁」為主，而低成癮組仍以金錢相關詞彙為主，如「省錢」等，在健康方面則有「跑步」等詞彙，結論：本研究將質性訪談文本，以文字探勘量化的方式處理巨量資料的優勢，來探討戒菸經驗及行為，發現不管是在高頻詞彙或相關詞彙，高成癮組皆以健康方面的詞彙居多，低成癮組都是金錢方面的詞彙居多，兩組的情緒方面也都是負面的情緒詞彙，藉由本研究結果，期望能提供給研究戒菸經驗相關領域者，在文本資料處理上多一種選擇。Objective: This explores the smoking cessation experience of high school students with smoking and learns more about other hidden information that has not yet been discovered. It explores the experience of smoking cessation through the application of text search technology, and compares the similarities and differences of high-frequency vocabulary between high and low addiction groups. Methods: This study first described the experience, feelings and thoughts of individual smoking cessation by smoking students, and learned about their smoking cessation experience through interviews. Then they used text exploration to explore 60 copies of text data, which were divided into high addiction and low addiction, and then discussed high-frequency vocabulary, word cloud and related vocabulary. Results: In the nicotine addiction cluster group, 60 students with cessation experience were surveyed, and they had high addiction to smoking _ (months), smoking _ (branch/day). In terms of the total score of nicotine addiction, high addiction group is higher than the low addiction group, which is statistically significant.In the total social support score, instrumentality, information and evaluation, the low addiction group was higher than the high addiction group, which is statistically significant; In the word cloud part of the smoking cessation experience, the top 30 words are drawn into a word cloud, and their high-frequency vocabulary is attributed. The high-addiction group has many high-frequency vocabulary related to health, such as "cough". "breathing" and "nicotine", etc., other aspects such as "bearing", "temptation", "willpower", etc.,and mainly negative, such as "stress", etc. for emotion.The low-addiction group is dominated by money vocabulary, such as "saving money"; in terms of the vocabulary of the high and low addiction groups, the vocabulary of the high-addiction group is still more common with health, such as "very asthma", "second-hand smoke" and "dyspnea". Other aspects include the words "habit", "sub-consciousness" and "getting together".Emotional aspects are still dominated by "stress" and "irritability", while low-addiction groups are still dominated by money-related vocabulary, such as "saving money." In terms of health, there are words such as "running". Conclusion: This study will explore the advantages of massive data in qualitative interviews and quantitative methods to explore smoking cessation experience and behavior.It is found that in high-frequency vocabulary or related vocabulary, the high-addiction group is mostly in health vocabulary, the low-addiction group is mostly in money vocabulary, and the emotional aspects of both groups are negative emotional vocabulary.With the results of this study, it is expected that it can be provided to those involved in the study of smoking cessation experience, and there is one more choice in text processing.