( 1. 中南大学 冶金科学与工程学院， 长沙 410083;
2. 广西大学 数学系， 南宁 530004)
摘 要: 在事务拓扑空间的基础上，将灰色系统理论引入属性集的关联规则挖掘中， 提出了一种新的适用于铝电解工业控制现场的灰关联度挖掘框架， 并给出了基于该框架的灰关联规则挖掘算法， 即Gray_CTL挖掘算法。 将算法分解为两个小问题： 1）计算关于时间属性的灰关联度， 这是算法的核心； 2）挖掘灰关联规则。 以电解槽的热平衡数据挖掘为例，对某电解槽一个月的生产数据进行分析后， 获得的灰关联规则说明该段时间内分子比、 槽电压等因素对温度的影响程度较大。
关键字: 铝电解； 数据挖掘； 灰关联规则； 算子； Gray_CTL算法
in aluminum electrolysis control
（ 1. College of Metallurgical Science and Engineering,
Central South University, Changsha 410083, China;
2. Department of Mathematics, Guangxi University,
Nanning 530004, China）
Abstract: The theory of gray system was brought into the mining of association rule about attribute sets, which based on the transaction topology space. Moreover, a new framework of mining using gray association degree was brought forward, and it can be used in aluminum electrolysis process control. The mining algorithm of gray association rule, viz. the mining algorithm of Gray_CTL, was described under the new framework. The algorithm was divided into two parts: the first is to calculate the gray association degree about time attribute and is the core of the algorithm; the second is to mine the gray association rule. An example of mining the gray association rule in the thermal equilibrium data was provided. After the analysis of production data coming from some electrolysis cell within a month, the obtained gray association rule indicates that the molecular ratio and cell voltage have greater effects on temperature of electrolyte than any other factors.
Key words: aluminum electrolysis; data mining; gray association rule; operator; Gray_CTL algorithms
