引入内容平衡的最大信息量组块分层选题策略

The Maximum Information Stratification Method with Content Balancing in Computerized Adaptive Testing

在0-1计分下,为了解决最大信息量组块分层策略(MIS-B)中未考虑内容平衡的问题,通过加入改良多项式模型来平衡内容属性.计算机模拟试验显示:选题策略在保持MIS-B能力估计精准度这一前提下降低了项目重叠率,提高了题库使用均匀性和项目曝光率的均匀性.

In 0-1 scored CAT,a new item selection strategy is proposed to improve the MIS-B method by introducing the Modified Muhinomial Model. The results of Monte Carlo simulations show that compared with MIS-B, the ap- proach proposed in this paper can reducing item overexposure rate, balancing item usage within the item bank, and maintaining measurement precision.