基于Apriori的关联规则算法及其在电厂中的应用  

Apriori-Based Association Rule Algorithm and Its Application in Power Plant

作　　者：黄文成[1] 贾立[1] 彭道刚 李望 

机构地区：[1]上海大学机电与自动化工程学院,上海200072 [2]上海电力学院自动化工程学院,上海200090

出　　处：《系统仿真学报》2018年第1期266-271,共6页Journal of System Simulation

基　　金：国家自然科学基金(61374044;61773251);上海市科委国际合作项目(15510722100);上海市科委创新行动计划(16111106300;17511109400)

摘　　要：针对目前电厂历史运行数据的挖掘技术对于高维度,存量大的样本数据存在挖掘效率低下,算法运行过程中部分参数设置无理论指导以及目标参数值确定不合理等问题,提出一种基于Apriori算法框架的改进量化关联规则挖掘算法。它以电厂运行经济性为目标,利用目标制导对样本空间进行样本维约束和数量压缩,进一步提高解决该类问题算法的挖掘效率和参数目标值确定的合理性。通过对某300MW机组运行数据的分析表明:改进量化关联规则算法能够提高数据挖掘效率并完成电厂参数目标值的合理确定。The data mining technology for historical data of power plant has the problem of low efficiency as the data dimension is high and data size is large. Some parameters are set without theoretical guidance and the objective parameter value is not reasonably determined in the corresponding algorithm of data mining. A mining algorithm with improved quantitative association rule based on Apriori is proposed. Aiming at the economical operation of power plant, target guidance is used to constrain the dimension and compress the quantity in sample space, which improves the mining efficiency and determines the parameter＇s target value reasonably. The operation data of a 300MW unit is analyzed and its results show that the improved quantitative association rule algorithm can improve the efficiency of data mining and determine the parameter value more accurately.

关 键 词：量化关联规则 参数目标值 经济性指标 目标制导 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]