為了校務研究與掌握教學品質，台灣的大專院校已全面實施教學評鑑，然而在教學評鑑資料中，學生回饋的意見往往很多，教學管理單位與教師需花費大量時間逐筆查看，方能了解每位學生的意見，且難以統整。為解決此問題，本研究以103-106年中原大學資訊管理學系必修課程之教學評鑑資料進行分析，透過文字探勘技術分析學生回饋的質化意見，藉由詞頻分析、文字雲分析、情緒分析針對不同屬性的課程（管理類、技術類、實作應用類、理論類、商用類）與低評鑑分數課程進行分析，以了解不同屬性的課程中學生的評價、意見及重視的教學元素。主要研究結論如下：（一）本研究提出之文字探勘分析模式可以有效運用在教學評鑑質化資料之分析，提供老師瞭解學生所重視的教學元素；（二）不同屬性的課程，學生所重視的教學元素有所差異，不同屬性課程的老師可以參考這些教學元素，豐富課程的準備與執行；（三）低評價課程的老師必須根據學生重視的教學元素，更有效地面對與改進教學方式。本研究分析結果將可供學校教學管理單位、系上、教師作為瞭解學生的課程學習意見、學校規劃課程以及老師改善教學內容之參考，以針對學生之學習需要，提高課程教學品質、提升學生的學習成效。

The universities in Taiwan have fully implemented the teaching evaluation for institution research and teaching quality. However, there are too many qualitative data in teaching evaluation from students, the administration and teachers must spend a lot of time and effort on reading each comment to understand students' opinions. It is hard to efficiently explore these opinions and get feedback for teaching. In order to solve this problem, this study adopted the text-mining technology to analyze the teaching evaluation data. Based on the teaching evaluation data of the compulsory courses of Information Management Department, Chung Yuan Christian University from the academic year 2014 to 2017, we use term-frequency method, word-cloud presentation, and sentiment-analysis to analyze the data of 5 types courses (including management, technical, practical, theoretical and business courses) and low evaluation-score courses to understand the critical teaching elements in these courses. The main results are as follows: (1) Teachers can understand the teaching elements which students want and care by the text-mining methodology applied to the qualitative analysis in teaching evaluation; (2) Students focus on different teaching elements by different features of course; (3) Teachers with low evaluation score must change their teaching method according students’ opinions. According the findings of this study, the administration and teachers can effectively understand students opinions about the courses and improve teachers’ teaching quality and students' learning performance.