挖掘頻繁序列應用於自動課程編排

Applying Mining Frequent Sequential Patterns Algorithm to Curricula Auto Weaving

10.29495/CITE.200707.0150

郝維華；林丕靜；陳宏任；邱展逢

科技教育 ； 資料挖掘 ； 頻繁序列 ； 課程編排

科技教育課程改革與發展學術研討會論文集

2006期（2007 / 07 / 01）

150 - 155

繁體中文

面對推廣科技教育發展，如何選擇適當的教材與教學順序？如何達到適性化的理想？一向是很重要的議題。以往之研究多偏重於統計方法(Statistic)或是純經驗式(Heuristic)，然而統計方法需要大量的資料累積才能得到較精確的描述統計，而純經驗式方法則偏重於教學者之主觀經驗，並非學員學習狀態之客觀事實。本研究提出之挖掘頻繁序列之BFLA(Breadth First Lattice Algorithm)演算法，可快速挖掘教師對於不同類別學生實施之同步教學所採用之頻繁課程序列，提供教師可針對不同學員類別，自動編排教材之功能。科技教育核心課程可以藉由此演算法擷取教師同步教學記錄檔案之歷史性資料，挖掘教師針對不同學員所呈現之課程內容以及教學順序。例如,電學對於高中學生、高職學生、國中技藝班、在職訓練學員、電機學群以及非電機學群應該有不同之課程內容安排以及教學順序。

Selecting proper curriculum for diverse technical-education students is always an important issue to teachers and educational policy decision makers. How to select proper curriculum? How to arrange them as an adaptive curriculum to all kind of students automatically still is a pending problem. Most of previous studies are focus on statistic and heuristic methods and have reached excellent outcomes. As we all know that statistics methods needs large amount of data to conduct a more precise result, and heuristic method is totally relied on subjective knowledge of local teachers who might not had any experience of teaching students from different area. In this study we proposed a Data Mining method, BFLA (Breadth First Lattice Algorithm), to find out those most frequently used teaching sequence of each course organized by teachers. This can sort out the core materials of each lessons or each course for diverse students. For example, the course of electrical theory for different area students, for instance, vocational students, high school students, college students, vocational junior high students and vocational training program students, should adopt different materials and different learning sequences.