國小數學領域整數四則單元電腦診斷測驗及補救教學系統研發

Computer diagnostic test and remediable teaching system development make use of in mathematics" The Four Basic of Integer" realm of elementary school.

林怡濱(lin i pin)

指導教授 ： 劉湘川

亞洲大學/資訊學院/資訊工程學系碩士班/碩士(2007年)

摘要

本研究旨在探討以貝氏網路為基礎，結合知識結構理念，並根據能力指標、子技能、錯誤類型，建立「整數四則」單元的專家知識結構及貝氏網路結構雛形，並以學生作答反應，針對紙本測驗樣本進行學生試題結構分析，以學生結構與貝氏網路的結合建立三種貝氏網路模型。目的在了解學生在學習「整數四則」的能力指標所對應之教材內容時，可能會產生的錯誤類型。並利用貝氏機率統計方法來進行資料的分析，並探討不同決斷值的設定對辨識率之影響以及在專家知識結構及學生試題結構下對整個貝氏網路辦識率的影響，最後建立補救教學系統並驗證其成效。研究結果發現： 1.以專家知識結構結合貝氏網路建構以概念性的評量架構為主的評量模式，可有效應用於診斷學生知錯誤類型與子技能。 2.欲建構出一個完整且有效的貝氏網路，首先需進行文獻探討建立貝氏網路，再利用實際資料進行分析以修正貝氏網路。以貝氏網路為推論工具的測驗診斷系統能有效的精準診斷，並達到評量出學童們常犯的錯誤或是迷思概念的診斷效果。 3.本研究中之診斷系統能同時完成施測及補救教學活動，達成同時同工之成效。 4.本研究研發之電腦適性學習系統，確實能作到預測率穩定、適性化、個別化的教學目的。 5.以貝氏網路推論電腦適性測驗，在適性和完整作答情況下，錯誤類型、子技能符合程度具有效度。

關鍵字

知識結構 ； 貝氏網路 ； 電腦適性測驗

並列摘要

This research aim inquire in to take Bayesian networks as foundation, combine Knowledge Structure Theory for the principle, Also according to capability beacon, subtechnical ability, false type, create"The Four Basic of Integer" unit of expert Theory and the structure embryo of the Bayesian networks of the Knowledge Structure, and make to answer reaction by student, aim at paper to test sample to carry on a student to try a structure analysis originally, create with the wedge bonding of student's structure and the Bayesian networks three kinds of Bayesian networks models.Purpose understand student at learn the capability beacon of"The Four Basic of Integer" to should it teaching material content, may generate of false type. The setting which also makes use of a Bayesian probability to statistics a method to carry on the analysis of data, and inquire into dissimilarity decision a value influences recognition rate and the Knowledge Structure and student try the influence that a structure down do to know a rate to the whole Bayesian networks at expert, finally create remediable teaching system and identify its result. The results：
1. Combine Bayesian networks a construction to take concept evaluation structure as main evaluation mode by Expert's Knowledge Structure, can be applied in diagnosis student to acknowledge wrong mistake type and sub- technical ability effectively.
2.Want to construct a complete and valid Bayesian networks, need to carry on a cultural heritage study to create first Bayesian networks, make use of again actual data to carry on analysis to revise Bayesian networks.Take Bayesian networks as to inference the test diagnosis of tool system ability valid precise diagnosis, and reach diagnosis result of evaluating the bug or enigma concept that the school-childrens often make.
3.The diagnosis system in this research can complete to test and rectify a teaching activity in the meantime, reaching in the meantime together the result of the work.
4.The computer adaptability of this research development learns system, really canning do an estimate rate stabilization, the adaptability turn and turn separately of teaching purpose.
5.Inference Computerized Adaptive Testing by Bayesian networks, under the situation that the adaptability and the integrity make to answer, the false type, sub- technical ability matches degree to have an effect.

並列關鍵字

Knowledge Structure Theory ； Bayesian networks ； Computerized Adaptive Testing

參考文獻

洪素敏(2004)。國小五年級學童分數迷思概念補救教學之研究。國立嘉義大學/數學教育研究所論文。

王怡群（2004）以知識結構為主符合SCORM 規格之適性化補救教學網頁製作-以高職氧化還原反應課程為例。臺中健康暨管理學院/資訊工程學系研究所論文。

沈俊達（2004）以知識結構為主的電腦適性診斷測驗系統之研究~以國民小學數學領域為例。臺中健康暨管理學院/資訊科技研究所論文

許嘉芬（2004）。九年ㄧ貫數學領域能力指標電腦化適性測驗及補救教學系統研究—以四年級數與量中的『有理數』為例。臺中健康暨管理學院/資訊科技研究所論文。

彭彥翔（2006）。以漸進提示法為基礎之國小數學科電腦化動態評量。亞洲大學/資訊工程學系研究所論文。