Time sequence real-time classification method based on Gram summation angle field imaging and Shortcut-CNN

Abstract

The invention discloses a time series data real-time classification method based on Gram summation angle field imaging and Shortcut-CNN. The method comprises the following steps: 1) collecting time series data; 2) normalizing the acquired time series data to obtain data; 3) representing the data in polar coordinates; (4) converting the data processed in the step (3) into a Gram matrix with the shape of M * M, and storing the Gram matrix as a gray level image, namely an EEG image; and 5) inputting the EEG image obtained in the step 4) into a Shortcut-CNN model to obtain a classification result.According to the method, the technical problem that a common classification method cannot directly use a two-dimensional convolutional neural network to perform high-performance real-time classification on time series data is solved, and the Shortcut-CNN model in the method has better classification performance compared with VGG16 and shallow CNN.

Description

technical field [0001] The invention relates to the technical field of data classification, in particular to a real-time classification method for time series data. Background technique [0002] Time series data is everywhere. Human activities and the natural world generate time series data all the time. Electroencephalogram (Electroencephalogram, EEG) data, weather data, data such as monitoring heartbeat and blood pressure, and data generated during the working process of sensors are all time-series data. A major task in time series data processing is classification. The classification of time series data can be widely applied to fields such as finance, industry, and medical treatment, and has important social and economic value. In recent years, due to the increasing availability of time series data, the demand for classification of time series data has also exploded. [0003] As people's research on time series data continues to deepen, the performance of time series ...

AI Technical Summary

Problems solved by technology

[0005] In view of this, the purpose of the present invention is to provide a real-time classification method for time series data based on Gram summation angle field imaging and Shortcut-CNN, to solve the problem that common classification methods cannot directly use two-dimensional convolutional neural network to classify time series data. Technical issues of high-performance classification of data, and technical problems of difficult real-time classification

Images