【摘要】 为了在高维不平衡质量特性数据集中提高关键质量特征识别效率,将EM(ExpectationMaximization)算法引入,通过逐步缩小比例较大数据集内的样本数量来抵消数据不平衡带来的负面影响。算例表明,该方法具有一定的合理性和可行性。更多还原

【Abstract】 In order to improve the efficiency of identifying the critical quality characteristic in data sets with high dimensional and imbalance,the Expectation Maximization Algorithm was introduced to offset the negative impact from the imbalanced data by reducing number of samples in larger class data set.The experimental results show that the method is sensible and feasible.更多还原