標題: 電腦輔助教學專家系統中知識管理之研究
A Study of Knowledge Management of Computer-Assisted Instruction Expert System
作者: 蔡昌均
Chang-Jiun Tsai
曾憲雄
Shian-Shyong Tseng
資訊科學與工程研究所
關鍵字: 電腦輔助教學;規則式專家系統;適性化教學;知識管理;知識表達;知識擷取;知識分析;Computer-Assisted Instruction;Rule-based Expert System;Adaptive Instruction;Knowledge Management;Knowledge Representation;Knowledge Acquisition;Knowledge Analysis
公開日期: 2001
摘要: 近年來,電腦輔助教學系統隨著網際網路的普及,愈來愈受重視,相關的應用技術及研究也相繼被提出,目的是希望讓電腦輔助教學系統可以如教師實際傳授課業給學生,導引學生進行學習;也希望利用電腦處理大量資訊及運算的能力,根據學生不同的學習狀況,提供適性化的教材,以補教師在傳統教學環境中,因學生人數眾多,無法因材施教的不足。在電腦輔助教學系統中,存在大量的教學內容與教學策略相關資訊,我們可統稱為電腦輔助教學系統中的知識。如何管理與存取大量的教材與教學策略,以讓教師與學生使用系統更為順利;如何擷取與儲存教師的教學經驗及策略,俾利於電腦輔助教學;如何分析學生的學習過程資訊,以藉由分析結果來更新原始教材與教學策略的設計;如何縮減系統管理所需的人力資源,以確實達到使用電腦系統的好處;以上這些問題可歸納成如何在電腦輔助教學系統中做好知識管理的工作。 在此論文中,我們提出電腦輔助教學專家系統(Computer-Assisted Instruction Expert System, CAI-ES),利用規則式專家系統的技術,推論儲存在知識庫中教師的教學策略與教學時的學習地圖,以依據學生的學習狀況,提供適當之教材予學生來進行學習;也就是說,每位學生的學習路徑與所學習的教材將依據本身的狀況而不同,以達到適性化教學的目標。 針對在電腦輔助教學專家系統中的知識管理,我們也提出相對應的知識管理機制,包括知識表達方法、知識擷取方法、知識管理器及知識分析器。在知識表達法中利用物件化特性,建構物件化教材與物件化規則集合,其中物件化規則集合即是儲存教師的教學策略,並搭配教學本體論的建構,可視為學習地圖,以達到適性化教學的目標。在知識擷取方法中,提出兩階段知識擷取,先建構學習地圖,再透過階層式表格法進行教學策略的知識擷取動作,最後建構出整體系統的教學知識庫。在知識管理器中,建構一個教材目錄管理器,透過如目錄服務的特性,讓教師及學生可以分享網路上的教材,這個教材目錄管理器並有半自動自我調整目錄結構的功能,以節省人力管理資源。最後,在知識分析器中,學生經過學習後,所留下的學習紀錄,可以透過所提出的兩階段模糊資料探勘及學習演算法進行分析後,提供教師及教材編輯者一些教材及學習地圖修改的建議資訊,以進行教學知識庫中相關知識的更新及修正動作。
With the fast growing and globally accepted of e-learning technology, Computer-Assisted Instruction (CAI) systems become a matter of great importance. The related applied technologies and studies are proposed to achieve the goal of guiding the learning process of students similar to teachers guide students to learn in traditional classroom. Moreover, based upon the ability of processing dramatic information, CAI system can provide different teaching materials for different students to learn according to the student’s learning aptitudes and records. In addition, there exist huge amount of teaching materials and teaching strategies in CAI system, called knowledge of CAI system. How to manage and access contents and teaching strategies of teachers for easily using e-learning system, how to extract teaching strategies from teachers for guiding students in learning process, how to analyze learning records of students for refining the design of original teaching materials and teaching strategies, and how to reduce manpower for maintaining e-learning system can be concluded as the problem, how to manage knowledge in e-learning system, and become an important issue. In this thesis, we propose a Computer-Assisted Instruction Expert System (CAI-ES), which applies rule-based expert system technology to infer the teaching strategies and learning map of knowledge base, and then provide different teaching materials for student to learn according to different learning aptitudes and records. In other words, each student has the different learning path and different teaching materials according to the learning aptitudes to achieve the goal of learning individually. For knowledge management in CAI-ES, we propose the corresponding knowledge management mechanisms, including Knowledge Representation, Knowledge Acquisition, Knowledge Organizer, and Knowledge Miner, to manage contents and knowledge. For Knowledge Representation, Object-Oriented Course Model (OOCM) and Object-Oriented Rule Model (OORM) are proposed to represent the teaching objects and the teaching strategies based upon object-oriented concepts. By applying the concept of ontology considered as learning map, the Teaching Strategy Knowledge Base (TS-KB) can be used to infer the suited teaching materials according to the learning aptitudes of students. For Knowledge Acquisition, Two-phase Knowledge Acquisition (Tp-KA) is proposed to acquire the learning map first and then to extract teaching strategies based upon Hierarchical Repertory Grid Analysis from teachers. For Knowledge Organizer, Course Directory Manager (CDM), which is a kind of web directory service, is proposed to manage and share Internet teaching resources. CDM can adjust the directory hierarchy semi-automatically to reduce the managing load. Finally, for Knowledge Miner, Two-phase Fuzzy Mining and Learning Algorithm (Tp-FML) is proposed to analyze the historical learning records of students and provide the suggestion about how to refine the teaching materials and learning map.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT900394007
http://hdl.handle.net/11536/68529
顯示於類別:畢業論文