基于样本分析的试题难度系数修正算法

针对当前题库试题难度系数的确定和评价标准不明确,提出一种以考试结果为样本获得难度系数并自我修正的算法:将样本数据序列{dk}构造到希尔伯特空间上,利用空间的完备性,从理论上确保序列{dk}收敛于常量D。基于该算法构造机器学习模型,它具有快速收敛和对历史数据进行学习和修正的特点。

Correction algorithm of item difficulty coefficient based on sample analysis

Item difficulty coefficient is an important index in producing an exam system and measuring the quality of exam.But it is not clear and easy on how to be sure and to evaluate the coefficient.This paper determines the degree of the difficulty on exam sample data.The degree is divided into theory and sample difficulty which is obtained from the factors of samples.The sequence of sample data is constructed into Hilbert space and convergence in a constant that is theory difficulty coefficient.According to the sample statistical learning theory,a machine learning model is constructed,which has rapid convergence and keeps learning and correcting the historical data.