【摘要】 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。更多还原

【Abstract】 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.更多还原