Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Mixed time period mode multivariable time sequence prediction method based on neural network

A hybrid time and neural network technology, applied in the field of data forecasting based on deep learning, can solve the problem of not being able to solve multivariate forecasting tasks well, and achieve the effect of improving information utilization, promoting fusion, and alleviating forecast lag

Pending Publication Date: 2022-02-01
EAST CHINA NORMAL UNIVERSITY
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0010] In view of the fact that the existing forecasting method based on deep learning is limited to univariate time series forecasting and cannot solve the problem of multivariate forecasting tasks well, the present invention provides a mixed time period mode multivariate time series forecasting method based on neural network

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Mixed time period mode multivariable time sequence prediction method based on neural network
  • Mixed time period mode multivariable time sequence prediction method based on neural network
  • Mixed time period mode multivariable time sequence prediction method based on neural network

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0046] Specific implementation mode 1. Combination figure 1 and figure 2 As shown, the present invention provides a neural network-based mixed time-periodic mode multivariate time-series forecasting method, comprising,

[0047] The encoder is used to extract the features of the original input data of the mixed cycle, and after the feature extraction of multiple convolutional long-term short-term memory network units, the encoding fully connected layer and the vector merging unit included in the encoder, the length of the short cycle length + 3 is obtained. Future time series data characteristics;

[0048] Then the decoder processes the original input data of the mixed period and the characteristics of future time series data, and processes the data through multiple bidirectional long and short-term memory network units of the decoder, fusion attention mechanism, decoding fully connected layer, autoregressive model and comprehensive prediction unit After that, the final time...

specific Embodiment

[0164] The datasets that are usually faced are multivariate (multi-user) datasets, so it is necessary to consider that the scale of different variables may affect the quality of the evaluation. Use the following metrics to avoid this problem:

[0165] Empirical Correlation Coefficient (CORR):

[0166] ΔY it =Y it -mean(Y i ),

[0167]

[0168]

[0169] Root mean squared relative error (Root Relative Squared Error, RRSE):

[0170]

[0171] Relative Absolute Error (RAE):

[0172]

[0173] Ω Test represents the divided test set. RRSE and RAE are normalized versions of Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), respectively, which avoid the interference caused by scale size in multivariate evaluations. For RRSE and RAE, lower values ​​represent better predictions; for CORR, the opposite is true.

[0174] The forecast details and strategies are shown as follows:

[0175] According to the ratio of 0.6:0.2:0.2, the original data set is divided ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a mixed time period mode multivariable time sequence prediction method based on a neural network, and belongs to the technical field of data prediction based on deep learning. The problem that an existing prediction method based on deep learning is only limited to univariate time sequence prediction and cannot well solve a multivariate prediction task is solved. The method comprises the following steps: performing feature extraction on original input data of a mixed period by adopting an encoder, and after feature extraction of a plurality of convolutional long-short-term memory network units, a coding full-connection layer and a vector merging unit included in the encoder, obtaining future time sequence data features with the length of short period + 3; then enabling the decoder to process the mixed period original input data and future time series data features, and obtaining a final time series prediction value after data processing of a plurality of bidirectional long and short term memory network units, a fusion attention mechanism, a decoding full connection layer, an autoregression model and a comprehensive prediction unit of the decoder. The method is used for predicting the mixed periodic data time series.

Description

technical field [0001] The invention relates to a mixed time period mode multivariate time series prediction method based on a neural network, and belongs to the technical field of data prediction based on deep learning. Background technique [0002] In the real world, human activities and natural laws can generate a large number of multivariate time series data sets, which to a certain extent reflect and affect human behavior patterns and social operation mechanisms. Therefore, people usually want to use these historical observations to predict future trends and changes in order to better plan and make decisions about the development of things. For example, if traffic police can know ahead of time the occupancy rate of a city's roads in the coming hours, then they can make correct and appropriate traffic decisions based on these predicted data to avoid traffic jams. therefore. Multivariate time series forecasting has always been one of the focuses of machine learning. ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06N3/04G06N3/08G06K9/62
CPCG06N3/084G06N3/044G06N3/045G06F18/214G06F18/253
Inventor 王晟陈曦郑来文李治洪刘敏李庆利齐洪钢刘小平周共健
Owner EAST CHINA NORMAL UNIVERSITY
Features
  • Generate Ideas
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com