Time sequence data missing value interpolation method based on attention mechanism

Abstract

The invention discloses an attention mechanism-based time series data missing value interpolation method. The method comprises the following steps of: obtaining time series data with missing values; inputting the time sequence data with missing values into the trained generator to obtain interpolated time sequence data; wherein the training of the generator comprises the following steps: inputting time sequence data with missing values into the generator, and obtaining complete time sequence data based on an attention mechanism; and inputting the time sequence data with missing values and the complete time sequence data into the discriminator, and carrying out adversarial training on the discriminator and the generator based on a loss function. According to the invention, new time series data conforming to original data set distribution can be generated. By trying on the attention mechanism, the expression ability of important features in the features can be enhanced, the expression ability of unimportant features can be reduced, and meanwhile, the processing efficiency can be improved. By means of the method, the accuracy of time sequence missing value interpolation can be improved, and the interpolation efficiency can be improved.

Description

technical field [0001] The invention relates to an attention mechanism-based time series data missing value interpolation method, which belongs to the technical field of computer science. Background technique [0002] In recent years, with the development of artificial intelligence technology, time-series data appears more and more frequently in human life. Time-series data is a sequence in which the values ​​of the same statistical index are arranged in chronological order, reflecting the state changes and development laws of things and behaviors over time. Common time series include some medical data, such as changes in the blood sugar level of diabetics over time in a day, as well as changes in website visits, road traffic, etc. at different times. [0003] Due to the instability or interference of the data acquisition equipment, the collected data is often missing. The lack of time series data will cause certain difficulties in the analysis, modeling and practical appl...

AI Technical Summary

Problems solved by technology

The missing part of the data set often implies the historical change law of the data. Due to the lack of this part, the model cannot be fully driven, and the parameters obtained in the final training will also be quite different from the optimal parameters.

Images