利用概念結構比對設計輔助精靈系統

Designing a Conversation Agent Using Mapping of Concept Structure

陳俊佑(Chun-Yu Chen)

指導教授 ： 賀嘉生

中原大學/工學院/資訊工程研究所/碩士(2007年)

https://doi.org/10.6840/cycu200700776

摘要

網路是一種提供人們互動式溝通的科技工具，人類利用網路來交換彼此的訊息和意見，產生互動的關係，而藉由不受空間、時間限制的網路互動式學習環境，讓個性害羞的學生或學業成就較低的學生也較能夠表達自己的看法。
隨著網際網路的盛行，人們也開始藉由網路來交談、討論以及學習，在許多研究中發現透過合作學習網路的輔助教學可以增進學生的概念、提升學生的後段認知以及提出更多的問題解決策略。
為了讓學習者之間互動性增加，因此本篇論文中建立一個線上討論室系統，讓學習者在只要有網際網路的情況下，都可以交流意見、詢問和討論；並在系統中加入討論輔助精靈引導學生討論，不僅可以輔助老師(或助教)監督學習者討論，更可以帶領學習者往該學習的概念前進，提高學習成效。
在討論的過程中，系統會將所有學習者的對話內容儲存在資料庫中，討論輔助精靈會分析所有對話內容，針對與單元主題無關和與單元主題相關的分析結果，將重點放在與單元主題相關的分析上，在對話內容中與知識地圖的關鍵字作比對，並得到學習者在討論中所產生的概念，將這些討論到的概念去建立概念結構，最後，利用概念結構的比對來得到學習者欠缺的概念，針對這些欠缺的概念，討論輔助精靈會產生適當的引導話語，引導學習者往缺乏的概念方向去進行討論，不但可以減輕老師(或助教)的監督工作，更可以讓學習者根據引導話語去進行討論，學習概念和解題方法，以達到輔助學習的效用。

關鍵字

形式概念分析 ； 對話分析 ； 智慧型代理人 ； 線上學習 ； 概念點陣

並列摘要

Network is a technological tool for communication. People can use network to exchange messages and suggestions. Shy or low-grade students also can express their ideas by the interactive learning environment which don't bounded by space and time.

As internet grows to be more and more popular, people begin to talk, discuss, and learn by internet. In many researches, cooperative learning can promote students' concepts, further students' cognitions, and get more problem-solving policies.

In order to increase interactions for students, we provide a online discussion system which every learner can exchange ideas, query, and discuss on it. We will add conversation agent to guide learners, it will not only substitute for teachers supervision, but also show concepts which must be learned and promote learning effects.

During the discussion, the system will save the conversations of the learner into the database. The conversation agent will analyze the content of the conversation according to whether or not it is related to the topic of the section, focusing on analyzing the content that is related to the topic. After comparing the content of the dialogue to keywords in the knowledge map, acquiring the concept that using the concepts that have been acquired to build a structure of concepts. In the end, using the concept structure to compare and spot out what concepts the learner didn't gain, then focusing on those concepts the conversation agent will give proper words of guidance to guide the learner into discussing the missing concepts.

Not only does it take the work load off of supervising for the teachers(or teaching assistances), the guidance can also lead the learner into discussions of learning the concepts and problem solving and therefore be able to accomplish the goal of assisting in the learning process.

並列關鍵字

Concept Lattice ； Sentence Analysis ； Formal Concept Analysis ； E-Learning ； Intelligent Agent

參考文獻

[許瑛玿等2003] 許瑛玿&廖桂菁，情境式網路學習環境互動行為分析：以高中地球科學線上學習為例，師大書報：科學教育類，48(1)，pp.93-118，2003

[陳冠廷2006] 陳冠廷，利用言語行為理論設計問題本位學習中討論輔助精靈，未出版之碩士論文，中原大學資訊研究所，台灣，2006

[CHS 2005] C. Philipp, H. Andreas, and S. Steffen, Learning Concept Hierarchies from Text Corpora using Formal Concept Analysis, Journal of Artificial Intelligence Research, 24, pp.305-339, 2005

[HDMX 1979] C. D. Holley, D. F. Dansereau, B. A. McDonald, J. C. Garland, and K. W. Collins, Evaluation of a hierarchical mapping technique as an aid to prose processing, Contemporary Educational Psychology, 4, pp.227-237

[Sel 1994] T. Selker, Coach: A Teaching Agent that Learns, Communications of the ACM, Vol.37, No.7, pp.92-99.