基于矩阵的双论域模糊概率粗糙集增量更新算法  

Incremental updating of fuzzy probability rough sets over two universes based on matrix method

作　　者：刘丹 李敬伟 LIU Dan;LI Jing-wei(College of Computer Science&Technology,Henan Institute of Technology,Xinxiang 453003,China;Big Data Engineering Research Center of Henan for Production&Manufacturing IoTs,Xinxiang 453003,China)

机构地区：[1]河南工学院计算机科学与技术学院,河南新乡453003 [2]河南省生产制造物联大数据工程技术研究中心,河南新乡453003

出　　处：《控制与决策》2021年第3期553-564,共12页Control and Decision

基　　金：河南省高等学校重点科研项目(19B520005)。

摘　　要：双论域模糊概率粗糙集是针对双论域信息系统的一种新的数据挖掘模型,现实应用中的双论域信息系统总是处于动态更新中,针对该问题提出一种基于矩阵的双论域模糊概率粗糙集增量式更新方法.首先,通过矩阵方法重新对双论域模糊概率粗糙集进行表示;然后,在矩阵表示模型的基础上,分别研究双论域信息系统两个论域中对象增加和减少时模型的增量式更新机制;最后,基于该增量式更新提出相应的增量式更新算法.实验分析表明:相比较于非增量式更新算法,所提出的增量式更新算法可以在很短的时间内完成模型的动态更新,从而验证算法的有效性;同时,与其他相关算法相比,所提出算法具有一定的优越性.The fuzzy probability rough set over two universes is a new data mining model for the two universes information system. However, the two universes information system in real application is always in dynamic updating. In order to solve this problem, a fuzzy probability rough set over two universes incremental updating method based on the matrix is proposed. Firstly, the fuzzy probability rough sets over two universes are expressed by matrix method. Then, on the basis of matrix expression model, the incremental updating mechanism of the model is studied when the objects increase and decrease in two universes of the two universes information system, respectively. And based on this incremental updating, the corresponding incremental updating algorithm is finally proposed. The experimental analysis shows that the proposed incremental updating algorithm can complete dynamic updating of the model in a very short time compared to the non-incremental updating algorithm, which verifies the effectiveness of the proposed algorithm. At the same time,compared with other related algorithms, the proposed algorithm has certain advantages.

关 键 词：粗糙集 双论域信息系统 模糊关系 概率 增量式学习 矩阵 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]