基于理想度的故障样本集评估优选方法研究

Evaluation Method for Failure Sample Set Based on Ideality

针对测试性验证试验中多次抽样得到的样本集存在随机误差的问题,提出基于理想度的故障样本集评估优选方法。首先确定了样本集评估指标体系,利用信息熵法确定了各个指标的权重,然后通过建立优化理想度模型来描述故障样本集的代表性。应用表明,该方法可以有效地评估故障样本集,指导故障样本的选取,结论分析表明该方法相比已有方法更加合理,工程适用性更高。

An evaluation method for failure sample set based on ideality relation analysis is proposed to solve the problem of random error sample set in the testability verification test.First,the evaluation requirement index of sample set is determined,the weight of each factor is obtained by using information entropy.Then the ideality model is founded to describe the typical fault sample set.Finally,the application shows that this method can effectively evaluate the failure sample set and direct the selection of failure sample set.The result indicates that this method is more rational and more practical in project compared with other method.