一种优化覆盖中心的三支决策模型

发布时间：2018-08-02 20:59

【摘要】：三支决策理论是传统二支决策上的拓展,具有三种决策规则,即接受、拒绝和不承诺。三支决策广泛适用于不确定或不完整信息的处理。基于覆盖算法的三支决策模型能够自动确定三个域,但是,传统覆盖算法的覆盖中心选取是个不可控的随机过程,单次实验的精度无法保证。因此,本文提出了一种优化覆盖算法中心的三支决策模型(optimal center in constructive covering algorithm,简称OCCCA)。该模型结合最近均值思想,在获取覆盖中心时,先求取数据集同类样本的均值,然后选取与均值最近的样本作为覆盖中心,从而实现优化覆盖算法中心的三支决策模型。实验表明,OCCCA比传统覆盖算法在三支决策模型分类准确率上有平均5%的提高。
[Abstract]:The three-branch decision theory is an extension of the traditional two-branch decision, which has three kinds of decision rules, namely, acceptance, rejection and non-commitment. Three decisions are widely used in the processing of uncertain or incomplete information. The three-branch decision model based on the coverage algorithm can automatically determine the three domains. However, the traditional coverage center selection is an uncontrollable stochastic process, and the accuracy of a single experiment cannot be guaranteed. Therefore, in this paper, a three-branch decision model, (optimal center in constructive covering algorithm, for OCCCA)., is proposed to optimize the coverage algorithm center. This model combines the idea of the nearest mean value, when obtaining the covering center, the mean value of the same kind of sample in the data set is first obtained, and then the sample closest to the mean value is selected as the covering center, so as to realize the three-branch decision model of optimizing the covering algorithm center. Experiments show that OCCCA can improve the classification accuracy of three-branch decision models by an average of 5% compared with the traditional covering algorithm.
【作者单位】：安徽大学计算机科学与技术学院;安徽大学计算机智能与信号处理教育部重点实验室;
【基金】：国家自然科学基金资助项目(61175046);国家自然科学基金青年资助项目(61402006) 安徽大学研究生学术创新项目(E14101002)
【分类号】：O225

【相似文献】