傳統的紙筆測驗僅能呈現出學生的分數水準,無法診斷出每一個學生在學習階段後所欠缺的技能及錯誤類型,老師們也只能挑出全體學生錯誤最多的題目,做整體性的補救教學。 貝氏網路是目前應用十分熱門的統計工具,在人工智慧領域及醫療方面應用也十分廣泛。它的判斷方法主要是以機率的方式來整合問題的不確定性,而許多心理計量的學者更以此統計方法應用在教育評量上。 本研究擬以國小六年級數學「量與實測」的能力指標為範圍,嘗試以機率推理為基礎的貝氏網路作為分析工具,來探討應用貝氏網路診斷學生錯誤類型的可行性。 本研究有以下四個目的: 一、 編製可鑑別出學生能力指標達成度且適用於電腦化診斷測驗系統的試題。 二、 以六年級數學科「量與實測」相關能力指標為例,建立一套以貝氏網路為基礎的電腦自動化診斷模式。 三、 探討貝氏網路模式,從樣本中對學生錯誤類型與能力指標判斷發生機率的辨識率高低。 四、 依據子技能及錯誤類型製作FLASH補救教學動畫元件,並探討電腦診斷測驗及補救教學是否達到預期的成效。
We only get score form traditional paper test which analysis details rarely such as weaknesses and mistaken types. According to the circumstance, teachers can only offer the remedy instructions based on the most mistaken types. In fact, variety weaknesses cause to different mistaken types. Bayesian Networks is a very popular statistics analysis tool. It applies perfectly in artificial intelligence and medical treatment. It judges and integrates the problem uncertainties by the probability method. And many philosophy scholars apply Bayesian Networks on Educational Rating. The main idea of the study is to research the ability of quantity and measurement index on Grade 6 and demonstrate the applicable of student mistaken types on the basis of Bayesian Networks which is a probability analysis method. Four purposes as below. 1. Designing the computerized quiz question and set up the reliable index. 2. Establishing cybernate diagnosis mode based on Bayesian Networks, which bounds the ability of quantity and measurement index on Grade 6. 3. Studying the ability of Mistaken Types by using Bayesian Networks. 4. Making remedy instructions flash and demonstrate results of the remedy instructions program.