研究生: |
徐蕙君 Shyu, Huey-Jiun |
---|---|
論文名稱: |
電腦鍵盤輸入適性練習策略之研究 An Adaptive Strategy on Keyboarding |
指導教授: |
何榮桂
Ho, Rong-Guey |
學位類別: |
碩士 Master |
系所名稱: |
資訊教育研究所 Graduate Institute of Information and Computer Education |
論文出版年: | 1996 |
畢業學年度: | 84 |
語文別: | 中文 |
論文頁數: | 56 |
中文關鍵詞: | 適性練習 、鍵盤輸入 、項目反應理論 |
英文關鍵詞: | Adaptive Drill System, keyboarding, Item Response Theory |
論文種類: | 學術論文 |
相關次數: | 點閱:220 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
鍵盤輸入的練習軟體是一種典型的反覆練習式(drill-and-practice) CAI
。然而,一般的鍵盤輸入練習CAI 並無自動調適之能力,無法依據使用者
個別能力的變化與進步情形,選擇適合使用者能力的難度題目作適性的練
習。本研究提出一種診斷鍵盤輸入難度的方法,並以項目反應理論(item
response theory; IRT)的三參數模式(three-parameter model),估計不
同按鍵組合的難度(difficulty)、鑑別度(discrimination)以及猜測參
數(pseudo-guess parameter),並建立電腦鍵盤輸入的練習題庫。另一方
面,練習時間的長短,也經常只是經由系統預先設定的練習時間來控制,
無法依據練習者個別的練習狀況,來調整練習時間的長度。本研究歸納練
習者能力值變化的資料畫出練習曲線,並觀察練習曲線的趨向作分類,依
據 Ho 與 Wang (1995) 所提出之適性練習系統 (adaptive drill
system; ADS)雛型,輸入 ADS 中以類神經網路(artificial neural
network)技術所製作的精熟練習決策器 (mastery decision processor;
MDP),預估練習者未來能力值,以預測的未來能力值中,相連能力值之差
值和作為精熟程度的判斷標準,差值和小於零代表已達精熟程度,即停止
練習。依據ADS的系統架構,配合鍵盤輸入練習題庫與適性練習策略,建
立電腦鍵盤輸入適性練習系統,就可依據個別練習者能力選取適合的題目
練習,且能依據練習狀況來決定練習中止的時間。經由實驗評估,電腦鍵
盤輸入適性練習系統的練習效率明顯優於一般的反覆練習式CAI,此對於
廣大的電腦鍵盤初學者與使用者,可提供適性且更有效率的練習方式。
There are so many keyboarding software packages we can buy in a software market. The practice method they used is to randomly select items and then provide practicing. In other words, they could not automatic adaptation according to the ability variations of the individual user, selecting proper items to users. This study proposes a diagnostic method about the keyboarding and applies the three-parameter model of the item response theory(IRT) to estimate the difficulty, discrimination, and pseudo-guessing parameter of keys on the keyboard. We could, therefore, create an item bank of keyboarding for practicing.On the other hand, practicing time of the keyboarding is usually controlled by the system. It could not adjust by the individual situation while a user practicing. Because of the shortcomings of traditional keyboarding software, therefore, we apply the adaptive drill system(ADS) model, proposed by Ho & Wang(1995), in our research. First of all we collect practicing curves of the abilities that users made while they are practicing, classify trends of curves, and then input those curves to the mastery decision processor(MDP), one of the most important unit in the ADS, that created by the neural network approach. After the MDP learns, it could be used to predict future abilities. By the differences of these output abilities that the MDP made, we classify whether a user reaches a mastery level, while sum of the differences less than 0, or not.The adaptive keyboarding system that combined by a calibrated keyboarding item bank and the adaptive drill system architecture. It could select proper items to individual user, provide them for practicing, and decide when the practicing process should be terminated. By the experimental evaluation, the practicing efficiency of the adaptive keyboarding system is obviously better than traditional drill and practice keyboarding CAI. It offers a new, adaptive, and efficient way for users who are practicing computer keyboard.