基于J2EE平台应用时序挖掘算法对纺织品出口风险预测的研究

贸易数据是按时间记录下的、不断更新中的海量数据。首先引入时间序列模式的概念,分析了时间序列的本质问题;其次改进了AprioriAll算法挖掘贸易序列数据库的有用序列模式;然后使用离散傅里叶变换子序列相似性查找的方法,将现有序列与挖掘到的感兴趣的序列模式进行子序列匹配,得到有用的知识;最后结合实际情况,合理搭建系统平台,将改进的算法应用在该平台之下得到满意的效果。

J2EE Based Platform of Time Series Mining Algorithm Forecast the Risks of Textile''''s Exports

It is well known that the commerce data are collected by time sequence and the commerce datebases have a great many records which update incessantly. Firstly, in this paper, the concept Time Series is introduced and its intrinsic property is discussed. Secondly, a novel data mining algorithm base on AprioriAll is developed to find useful sequential pattern in commerce databases, then use discrete Fourier transform analysis the similarity sub- series searching method. Finally, the system platform is developed and obtain a satisfied results.