應用資料探勘技術於教師教學評量之研究

A Study of Teacher's Teaching Assessment by Using Data Mining Technology

王智立(Chih-Li Wang) ； 黃婷(Ting Huang)

《智慧科技與應用統計學報》 12卷2期 (2014/12) Pp. 109-126

摘要

在這個知識爆炸的時代，隨著資訊科技的快速發展，大量資料的累積已是隨處可見的，但要從大量資料中尋找出有用的知識或資訊卻不是件容易的事情，如何有效的去運用這些資料，挖掘出有價值的資訊給決策者，讓他們可以做出最佳的策略，這是值得去探討的。資料探勘是近年來逐漸蓬勃發展的一項技術，是一種結合數個不同問題領域（problem domain）的專業技術（technologies），並且可找出資料中資訊的一個流程（procedure）（廖述賢、溫志皓，2012）。本文將使用類神經網路（neural network）、決策樹（decision tree）等兩種資料探勘技術，並以實際的教學評量資料為例，找出影響教師教學評量分數高低的因素，並應用於改善教學方式，提高教學品質，最後本文比較此兩種資料探勘技術的優劣，可發現決策樹的準確度較高，但修剪後則小幅下降，類神經網路修剪前後的準確度則差異不大；在教學特性中兩種探勘技術的準確度相差不大。

關鍵字

資料探勘 ； 教師教學評量 ； 類神經網路 ； 決策樹

並列摘要

With vigorous development of the Internet in this era of knowledge explosion, it's simple to see lots of data increased, but it's not easy to find useful knowledge or information from large data. Therefore how to use these data to find valuable information to decision maker to do best decision that it's good to discuss. Recent year data mining is a vigorous development technology that collects several professional technologies of different problem domains, and it can to find a procedure of information in data (Liao, Wen, 2012). In this paper, we use two data mining technologies of neural network and decision tree and use teaching assessment data. We want to find what the fact is to affect teaching assessment score that can use to improve teaching ways and to raise teaching quality. Finally we compare two data mining technologies to find merits and drawbacks. In teacher' characteristic that we can see decision tree's accuracy is high. But it decreases when we cut its branch. Neural network's accuracy didn't have many differences from beginning to end. And in teaching' quality didn’t have many differences of accuracy on two data mining technologies.

並列關鍵字

Data mining ； teaching assessment ； neural network ； decision tree