基于全方位优化算法的改进马田系统分类方法

Classification using improved Mahalanobis-Taguchi system based on omni-optimizer

针对马田系统的若干不足,提出一种改进的马田系统优化模型,其核心思想是根据分类问题的目的和特点提出若干优化目标,采用优化模型替代正交表和信噪比筛选关键变量.针对模型的特点,采用了一种全方位优化算法进行求解.通过对4个UCI数据集的算例分析表明,该方法不仅有较好的分类精度,且能筛选关键变量,降维效果明显.最后对一个实际生产案例进行了研究,结果表明该方法在保持高分类效率的情况下,能够显著减少质量检测变量,降低成本,提高生产效率.

For the inadequacies of Mahalanobis-Taguchi system（MTS）,the authors propose an improved MTS optimization model（MTSO）.The core idea is that a number of optimization objectives are proposed based on the purpose and characteristics of the data classification problem and optimization model is used for screening important variables instead of orthogonal arrays and signal-noise-ratio.The authors also propose an omni-optimizer method to solve the optimization problem according to characteristics of the model.Results of studying 4 UCI data sets indicate that the proposed method is very effective not only in classification but also in dimensionality reduction.Finally,MTSO is employed to analyze the practical quality inspection process of notebook computer manufacturing.Implementation results show that the inspection attributes are significantly reduced and that the inspection process can also maintain high inspection accuracy so that the production costs can be reduced and enhance the productivity.