應用模糊理論估計適性測驗之受試者能力

Autor: Lee Min-Feng, 李銘峰
Rok vydání: 2004
Druh dokumentu: 學位論文 ; thesis
Popis: 92
Computer Adaptive Test (C.A.T.) is a popular test method in nowadays. For example, GRE, TOEFL, etc. have adopted C.A.T. for holding examinations. It is not only an improvement of technology but also a trend in the future that spending less time, testing no matter where you are as long as you have computers, and giving examinee suitable questions based on their abilities. In this research, it proposes a fuzzy adaptive test model that testing examinee''s ability. It uses description of uncertain elements, which are given in fuzzy mathematics, and blurs it during a examinee''s solving process becomes a kind of fuzzy numbers. In this model, we take examinee''s solving condition, and use fuzzy ability to calculate fuzzy regression in order to estimate examinee''s ability. The aim of this research is to put fuzzy theory into Item Response Theory, and proposes a new model to be used on the estimate examinee ability. In the new model, we use fuzzy regression as a index of ability-estimation to help estimate examinee''s ability. In this research, we also compare it to the traditional method, MLE. Under the circumstances of fewer questions, it can show that new model has higher precision when it compares to MLE. Besides, with the increasing of examinee''s questions, the ability that are estimated by new model and MLE, errors are almost the same, moreover new model''s error is fewer than MLE. Key Words: Item Response Theory, adaptive test, fuzzy theory, fuzzy regression, and estimate ability.
Databáze: Networked Digital Library of Theses & Dissertations