论文部分内容阅读
解决不平衡数据分类问题,在现实中有着深远的意义。马田系统利用单一的正常类别构建基准空间和测量基准尺度,并由此建立数据分类模型,十分适合不平衡数据分类问题的处理。本文以传统马田系统方法为基础,结合信噪比及F-value、G-mean等分类精度,建立了基于遗传算法的基准空间优化模型,同时运用Bagging集成化算法,构造了改进马田系统模型算法GBMTS。通过对不同分类方法及相关数据集的实验分析,表明:GBMTS算法较其他分类算法,更能够有效的处理不平衡数据的分类问题。
To solve the problem of unbalanced data classification has far-reaching significance in reality. The Martin system uses a single normal category to construct a reference space and a measurement reference scale, and a data classification model is thus established that is well suited to the processing of unbalanced data classification problems. In this paper, based on the traditional Martin system, combined with the classification accuracy of signal-noise ratio, F-value and G-mean, a reference space optimization model based on genetic algorithm is established. At the same time, Bagging integration algorithm is used to construct the improved Martin system Model Algorithm GBMTS. Through the experimental analysis of different classification methods and related datasets, it shows that the GBMTS algorithm can deal with the classification of unbalanced data more effectively than other classification algorithms.