以資料探勘技術預測影響大學教師評價學校的因素

Data Mining Technique for Constructing the Prediction Model of University Teachers' Evaluation of Schools

蔡欣倫(Hsin-Luen Tsai) ； 魏慧美(Hui-Mei Wei) ； 郭隆興(Lung-Hsing Kuo)

《科學與人文研究》 9卷2期 (2022/02) Pp. 24-38

https://doi.org/10.6535/JSH.202202_9(2).0002

摘要

大學教師的工作包含教學、研究和服務，此外，有些學校更要求教師，尤其是新進教師擔任行政工作，對於學校的政策和要求，因為有教師評鑑的壓力，老師們總是儘可能配合。然而，大學教師對任教學校的評價，也是改善學校，讓學校永續經營的重要因素。本研究將透過台灣高等教育整合資料庫（TIPED）中大專校院專任教師調查問卷與結果進行探索，預計探索教師所處教育環境中，所承擔的教學任務，行政工作，研究工作，以及在校內資源中，各種與教師教學、研究和服務相關的資料，透過統計和資料探勘的技術，探索這些觀察變項，預測哪些因素影響大學教師對於學校評價。研究結果顯示，Logistic和BayesNet兩分類器預測的模型是最佳選擇。學校組織文化影響最大，其次是學校法規、人力資源、設備，第三是學校類型。J48決策樹說明組織文化中等以上的學校，教師給予學校的評價越好；私立大學教師對學校的評價以中等和以上為主。除了組織文化之外，學校法規、人力資源、設備也扮演重要角色，即使組織文化不佳，但其他條件能提供，教師對於學校也能有中等以上的評價。結論亦提出高等教育行政決策主管如何經營學校組織文化的建議，以期創造友善校園，營造永續的高等教育環境。

關鍵字

資料探勘 ； 大學教師 ； 學校組織文化 ； 類神經網路 ； 決策樹

並列摘要

The mandated tasks of the university teachers include teaching research and service. However, some university teachers, especially new faculties, are required to serve as part-time administrators. With regard to school policies and requirements, teachers always work as hard as they can under the pressure of teacher appraisal. In fact, the university teachers' evaluation on schools is also an element for improving the school to make it development sustainable. This research explored the predictable factors affecting full-time university teachers' evaluation of their schools based on the analysis of the questionnaires and results from teachers of universities in the Taiwan Integrated Postsecondary Education Database (TIPED). The influential factors may be related to the observational variables, such as the educational environment, the teaching tasks, administrative work, research work, and the resources in the school. Results indicated that BayesNet and Logistic classifiers form the best models. In addition, school organizational culture has the greatest impact on the teachers' evaluation of the school, followed school regulations, human resources, facilities, and school types. The J48decision tree illustrates most teacher's evaluation of school organization culture are above the middle; private university teachers tend to rate middle and above. In addition, human resources and facilities also play a vital role. The teachers' evaluations might be subject to many factors; therefore, teachers can also have evaluated a middle or higher level of the school if other conditions are satisfied even if the organizational culture is not good. The conclusion also puts forward suggestions on how higher education administrators should manage the school's organizational culture in order to create a friendly campus and sustainable higher education environment.

並列關鍵字

Data Mining technique ； University teachers ； School culture ； Artificial Neural Network ； Decision Tree

參考文獻

Wong, T.-T., & Yeh,P. -Y. (2020).Reliable Accuracy Estimates from k-Fold Cross Validation. IEEE Transactions on Knowledge and Data Engineering, 32(8), 1586-1594.

Witten, I. H., Frank, E., Hall, M.A., & Pal, C.J. (2017). Data Mining: Practical Machine Learning Tools and Techniques, Fourth Edition, Elsevier: Morgan Kaufmann.

Quinlan, J. R. (1993). Combining instance-based and model-based learning. In Proceedings of the tenth international conference on machine learning, 236-243.

彭森明（2005）。台灣高等教育資料系統之建置與運用(2/3): 大專院校專任教師（D00086）【原始數據】。取自中央研究院人文社會科學研究中心調查研究專題中心學術調查研究資料庫https://srda.sinica.edu.tw。doi:10.6141/TW-SRDA-D00086。

胡航，杜爽，梁佳柔，康忠琳. (2021). 學習績效預測模型構建：源于學習行為大資料分析[J]. 中國遠端教育（4）：8-20.