通用CAI课件的智能化评估模型

General intellectualized evaluation model for CAI courseware quality

对于CAI课件质量评估问题,提出基于正则化BP神经网络(RBPNN)的智能化评估模型。分析评估活动相关环节,进而给出解决方案,包括对评估数据的区间化处理、减少指标集参数冗余。然后通过学习训练确定RBPNN的网络权值及其它参数,对RBPNN输出评语集给出模糊处理。由此给出解决问题的算法,最终通过一个实例来说明算法的使用,实验结果表明,该评估模型比较实用,完全满足CAI课件质量评估的技术要求。

An Intellectual evaluation model based on regularized BP neural network （RBPNN） is developed for courseware quality eva!uation problem. Through analyzing a series of question on evaluation the solutions to them are presented respectively, including reducing unnecessary component part of index set, making datum interval, determining RBPNN＇s parameters and weights by means of samples training and further treating the output of the RBPNN as fuzzy set. And then the algorithm for their computation is given. At last, the use of the algorithm is described by an example, the experiment shows that the evaluation model is practical and satisfies completely technical requires of CAI courseware quality evaluation.