Compression Processing Estimation Method for Time Series Big Data

The rapid development of computer science has caused an explosion of mining interest in the time series big data domain. Thus the data processing architecture has been proposed to meet the demand for optimizing the performance of systems. This paper presents an implementation of data processing methods for uncertain time series big data with noise. The Kalman filter, an estimation technique to extract high dimension characteristics of states in the target tracking field, is adaptive and can guarantee tracking target states with certain measurement range. Thanks to the Kalman filter, we can compress a datasets by irregularly sampling observation data, which is called the compression processing estimation method (CPEM). The simulation results and its comparisons to the mean value method (MVM) show that we can quickly, accurately extract important information of time series and get a good compression result.