Evaluating the performance of different diagnostic adaptive testing algorithm combining knowledge structure

• Main Authors: Shu-Yu Cho, 卓淑瑜
• Other Authors: Bor-Chen Kuo
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/56908298434672248577

碩士 === 國立臺中教育大學 === 教育測驗統計研究所 === 99 === This study propose a novel cognitive diagnosis computerized adaptive testing algorithm, knowledge structure based item selection strategy, which provides ancillary information by knowledge structure to improve the diagnosis accuracy at the beginning of administrating cognitive diagnosis computerized adaptive test. To investigate the performance of different cognitive diagnosis computerized adaptive testing algorithms with different types of Q matrix, a simulation study is implemented. There are some results as follow: 1.The diagnosis accuracy decreases as the number of attributes measured per item in average increase. 2.With different types of Q matrix, the diagnosis accuracies of random rule, KL, PWKL and HKL decrease as the number of attributes measured per item in average increase. Nevertheless, the diagnosis accuracies of SHE do not affected by using different types of Q matrix. 3.Under different item selection algorithms, PWKL and HKL have the best performance. 4.The performance of knowledge structure based PWKL and HKL are better than PWKL and HKL.