A Completion Method for Labeled Time Series Data

Abstract

The present invention discloses a method for completing tagged time series data, which is mainly used to solve the problem of missing a continuous segment of time series data common in real scenes. The core idea of ​​the data completion method includes two aspects: first, using Hankel matrix technology organizes low-dimensional time series into high-dimensional forms, introduces high-order time dependencies, and uses matrix decomposition to complete missing data on this basis, thereby effectively overcoming the problem of data loss; second, in the overall algorithm The tag information is modeled in the framework, and the tag information is used to support the data completion process, so that the completed data is more in line with the real scene. By rationally using the above two ideas, the method proposed by the present invention can achieve a better completion effect in the real time series data missing scene; at the same time, the method has strong interpretability, and can also be based on the method More expansions are carried out so that it can be effectively used in various real scenarios.

Description

technical field [0001] The invention relates to a computer application method for data collection and transmission of time series, in particular to a method for completing time series data with labels. Background technique [0002] With the continuous development of computer intelligent perception technology, computing power and storage technology, a huge amount of data can be obtained every day, and there is a lot of knowledge in these data that is worth mining. Time-series data is a collection of observations made in chronological order, which is widely used in many different kinds of applications, such as: behavior capture, sensor networks, weather forecasting, financial market modeling, etc. For time series data, common analysis and processing tasks include prediction / regression, outlier detection, pattern recognition, etc., but these tasks are usually based on complete data. [0003] However, in real scenarios, due to the inevitable data loss caused by equipment perfor...

AI Technical Summary

Problems solved by technology

However, this method does not work well in the case of sparse data, because the initial values ​​​​filled with simple methods may be misleading to the subsequent optimization process.

In addition, for related work based on matrix decomposition, due to the loss of the entire column of the original data, this will cause the decomposed matrix to be unable to learn in the corresponding column.

[0005] 2) The influence of external factors on the time series cannot be expressed

However, in real scenarios, due to the influence of external factors, time series data is often uncertain. Therefore, when dealing with data loss in a continuous period, especially when the data loss spans a long period of time, related work cannot be calculated. get the actual value

[0006] It is very common in real-world scenarios that time series data is missing for a continuous period of time. However, existing related methods cannot achieve good results in dealing with this problem.

Images