基于岭估计和AMOGA的马田系统分类方法
发布时间：2018-11-18 09:50
【摘要】：马田系统是多变量数据挖掘中模式识别方法的新进展,变量间的复共线性会通过马氏距离影响马田系统变量筛选的效果和判别的准确率。为了克服复共线性对马田系统的负面影响,提出了基于岭估计新的测量尺度—岭马氏距离,通过变量敏感性和条件数绘制三条岭迹来确定岭参数,并设计了自适应多目标遗传算法进行基准空间优化,使得马田系统分类方法对病态数据具有更好的耐受性。通过案例验证了岭马氏距离可以很好的克服复共线性,并提高马田系统分类方法的判别准确率。
[Abstract]:Multi field system is a new development of pattern recognition method in multivariate data mining. The complex collinearity between variables will affect the effect of variable selection and the accuracy of discriminating through Markov distance. In order to overcome the negative influence of complex collinearity on the Matton system, a new measurement scale of Ridge estimation, called Ridge distance, is proposed. The ridge parameters are determined by mapping three ridge traces by variable sensitivity and conditional number. The adaptive multi-objective genetic algorithm is designed to optimize the reference space, which makes the classification method of Matian system more tolerant to ill-conditioned data. It is proved by a case study that Lingmao distance can overcome the complex collinearity and improve the accuracy of the classification method.
【作者单位】： 南京理工大学经济管理学院;江苏科技大学经济管理学院;国立台北商业大学管理学院;
【基金】：国家自然科学基金资助项目(71271114) 教育部人文社会科学基金资助项目(14YJA910004)
【分类号】：TP18;TP311.13
本文编号：2339707
[Abstract]:Multi field system is a new development of pattern recognition method in multivariate data mining. The complex collinearity between variables will affect the effect of variable selection and the accuracy of discriminating through Markov distance. In order to overcome the negative influence of complex collinearity on the Matton system, a new measurement scale of Ridge estimation, called Ridge distance, is proposed. The ridge parameters are determined by mapping three ridge traces by variable sensitivity and conditional number. The adaptive multi-objective genetic algorithm is designed to optimize the reference space, which makes the classification method of Matian system more tolerant to ill-conditioned data. It is proved by a case study that Lingmao distance can overcome the complex collinearity and improve the accuracy of the classification method.
【作者单位】： 南京理工大学经济管理学院;江苏科技大学经济管理学院;国立台北商业大学管理学院;
【基金】：国家自然科学基金资助项目(71271114) 教育部人文社会科学基金资助项目(14YJA910004)
【分类号】：TP18;TP311.13
【相似文献】
相关期刊论文 前2条
1 归庆明;m个相依回归方程系统回归参数的预检验岭估计[J];解放军测绘学院学报;1998年02期
2 李崇贵,石强,赵宪文,斯林;用岭估计研究以RS和GIS为基础的森林蓄积预报[J];遥感技术与应用;2001年03期
相关硕士学位论文 前1条
1 陶建波;基于岭估计的马田系统分类方法研究及其应用[D];南京理工大学;2016年
，本文编号：2339707
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2339707.html
