摘要
在日常生活中广泛存在着各种时间序列数据 ,发现时间序列知识、对时间序列进行预测正成为数据挖掘与知识发现的重要内容。首先提出了基于云模型的时间序列预测机制 ,该机制以云理论为知识表示的理论基础 ,提出了两种预测知识 :准周期变化规律和当前趋势 ,并综合两种不同粒度的预测知识实现了时间序列的预测。然后着重于运用云模型进行知识表达、定量数值与定性知识的转换以及综合不同时间粒度的知识进行时间序列预测。
In our daily life, there are various kinds of time series data, and time series prediction becomes one of the important aspects of Data Mining and Knowledge Discovery (DMKD). This paper presents a new mechanism of time series prediction based on cloud models. Cloud theory is introduced as the theoretical basis of the mechanism. Two kinds of predictive knowledge, quasi periodical regularity and current tendency are proposed which are represented by different granularities as predictive linguistic rules and current cloud respectively, and summed up by synthesized cloud. Soft inference is realized to get the predictive results in several forms. We focus this paper on the be application of cloud theory to transform between quantitative and qualitative knowledge, synthesize different kinds of knowledge and realize the soft inference.
出处
《解放军理工大学学报（自然科学版）》
EI
2000年第5期13-18,共6页
Journal of PLA University of Science and Technology(Natural Science Edition)
基金
国家高技术研究发展计划 (863计划 )资助项目 !(863 -3 0 6-ZT0 6-0 7-2 )
国家重点基础研究发展规划 (973计划 )资助项目 !(G19980 3 0
关键词
数据挖掘
时间序列预测
云模型
预测语言规则
综合云
data mining
time series prediction
cloud models
predictive linguistic rule
synthesized cloud
