时间序列下模式挖掘模型设计-计算机工程与应用.PDFVIP

146 2015 ，51（19） Computer Engineering and Applications 计算机工程与应用 时间序列下模式挖掘模型设计 1 1 2 3 张可佳 ，李春生 ，姜海英 ，赵 森 1 1 2 3 ZHANG Kejia , LI Chunsheng , JIANG Haiying , ZHAO Sen 1.东北石油大学 现代教育技术中心，黑龙江 大庆 163318 2.大庆油田有限责任公司 第二采油厂地质大队，黑龙江 大庆 163000 3.大庆油田有限责任公司 矿区服务事业部，黑龙江 大庆 163000 1.College of Computer and Information Technology, Northeast Petroleum University, Daqing, Heilongjiang 163318, China 2.The Second Oil Plant Geological Brigade, Daqing Oil Field Co., Daqing, Heilongjiang 163000, China 3.Services Department, Daqing Oil Field Co., Daqing, Heilongjiang 163000, China ZHANG Kejia, LI Chunsheng, JIANG Haiying, et al. Design of pattern-mining based on time series. Computer Engineering and Applications, 2015, 51 （19）：146-151. Abstract ：In the process of pattern mining applied in intelligent method, in order to improve accuracy and reliability in data change model, this paper makes logic transformation in field impact factor which is defined by experts, and it is based on FC closure model. It uses mean square error algorithm based on time series of processing the original data, and uses huge intensity to judge the function then reject invalid element, and reduces the dimension of data to complete data preparation. Appropriate and practical mathematics model is selected to do the temporal data fitting. The thinking of classi- fication analysis method is drawed on, bringing in CCM-ECM model to describe the final mining results, finishing pattern mining model （TODM ）based on sequential design. At the same time, this paper proposes a set of more scientific calcula- tion methods for the confidence level calculation and adaptive adjustment of the model, to achieve the purpose that excavates potential law in the data deeply, and increases the degree of high precision detailed description on the data changed model. Final