摘要
结合粗集的属性约简和支持向量机的分类机理,提出了一种融合算法。应用粗集理论的属性约简过程作为数据的预处理,把冗余的属性和冲突的对象从决策表中删去,却不损失任何有效信息;然后利用支持向量机进行分类识别。这样可以大大降低数据维数,减小支持向量机分类过程中的复杂度,在不同程度上避免了训练模型的过拟合现象,且分类性能更加优越。最后通过对雷达信号识别的仿真实例说明了上述方法的有效性。
A hybrid algorithm based on attribute reduction of rough set and classification principles of support vector machine(SVM) is presented.Firstly,the attribute reduction of rough set has been applied as preprocessor so that the redundant attributes and conflicting objects can be deleted from decision table but remaining efficient information lossless.Then,the classification and recognition based on SVM is realized.By this method,the dimension of data is reduced greatly,the complexity in process of SVM classification is decreased highly,and the over-fit of training model is prevented at some extent,so better classification performance can be obtained.Finally,the simulation experiment of radar signal recognition and its results show this combined method is effective.
出处
《航天电子对抗》
2010年第6期35-37,51,共4页
Aerospace Electronic Warfare
