一类周期时态关联规则的知识发现问题  

Knowledge Discovery for a Kind of Periodic Temporal Associated Rules

作　　者：田媛[1] 李向军[2] 孟志青[3] 

机构地区：[1]湖南工程学院计算机科学系,湖南湘潭411101 [2]西安文理学院计算机系,陕西西安710065 [3]浙江工业大学经贸管理学院,浙江杭州310032

出　　处：《西安文理学院学报（自然科学版）》2005年第2期64-68,共5页Journal of Xi’an University(Natural Science Edition)

基　　金：西安文理学院自然科学基金资助项目(KY2004027)

摘　　要：探讨了一类基于某种时态约束的周期时态关联规则的有关问题.首先,引入了时态型的定义以及相关的概念.然后,给出了一种时态事件模型,用以描述基于时态型的不同属性的各种状态的事件,同时定义了一类事件间的周期时态关联规则,它能在诸如股票波动、天气预报、商品销售等领域提供短期的预测和决策.最后,给出了发现此类周期时态关联规则算法的主要步骤,并且给出了对股票数据挖掘的实验结果分析.This paper discusses a kind of periodic temporal associated rules, based on temporal constraint. First, a definition of temporal type and the interrelated conceptions are introduced. Then, based on the different attributes with the temporal type, a type of temporal event describing various states of events is defined; the periodic temporal associate rule between two events, which can provide the prediction and decision in the shorter time for fields such as stock fluctuation, weather forecast, commodity sale and so on is also explained. Finally, the main idea of an algorithm to discover the periodic temporal associated rules is given. Meanwhile, the analysis of experiment with the data mining of stocks is made.

关 键 词：时态型 时间粒度 周期时态关联规则 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]