Time sequence trend extraction and prediction method based on compressed sensing

Abstract

The invention discloses a time sequence trend extraction and prediction method based on compressed sensing, and belongs to the field of time sequence prediction methods. According to the method, random sub-sampling is carried out on an original time sequence to obtain an observation vector, and then trend extraction is carried out on the original time sequence through multiple times of reconstruction under the condition that the sparsity of a reconstructed signal is determined. An optimal trend sequence is found in the plurality of reconstruction trends by utilizing a similarity evaluation index based on the Euclidean distance sequence, and future trend development of the optimal trend is predicted by using a support vector regression prediction method. The trend information of the original time sequence can be well extracted only by utilizing the observation vector of random sub-sampling of the original time sequence, and the invention has a certain compression ratio and is more suitable for being used under the condition of big data mining processing. The sequence trend information is predicted, so that the problems of long-term prediction, noise interference and the like of thetime sequence trend can be effectively solved, and the result is more scientific and effective.

Description

technical field [0001] The invention relates to a time series trend extraction and prediction method based on compressed sensing, belonging to the field of time series prediction methods. Background technique [0002] Time series refers to the sequence of the values ​​of the same statistical index arranged in the order of their occurrence time. Time series data essentially reflects the trend of one or some random variables changing over time. The main purpose of time series analysis is to mine certain rules from data and use it to estimate future data. Time series analysis is to predict future data based on existing historical data. At present, the analysis and prediction of time series has been applied in many fields, such as housing price prediction and stock prediction in financial economy; temperature prediction and rainfall prediction in meteorology and hydrology; sensor data collection prediction in signal processing, etc. The application scenarios of time series fore...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the problems of long processing time and high complexity in the context of big data analysis and processing, existing time series trend extraction methods require full sampling of raw data, and existing time series

Images