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

Time sequence prediction method and device based on Bayesian optimization and wavelet decomposition

A wavelet decomposition and time series prediction technology, applied in neural learning methods, based on specific mathematical models, special data processing applications, etc., can solve problems such as difficulty in ensuring prediction accuracy, and achieve high accuracy and improve accuracy.

Pending Publication Date: 2020-10-30
BEIJING TECHNOLOGY AND BUSINESS UNIVERSITY
View PDF3 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, after research, since most of the time series data are obtained from the real environment, the data often have strong volatility, randomness, and complexity, and it is difficult to guarantee predictions only by relying on deep neural networks to analyze and learn them. Accuracy

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
  • Time sequence prediction method and device based on Bayesian optimization and wavelet decomposition
  • Time sequence prediction method and device based on Bayesian optimization and wavelet decomposition
  • Time sequence prediction method and device based on Bayesian optimization and wavelet decomposition

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0057] like figure 1 As shown, this embodiment proposes a time series prediction method based on Bayesian optimization and wavelet decomposition, the method comprising:

[0058] S101. Optimizing model hyperparameters according to a Bayesian optimization method to obtain optimal hyperparameters.

[0059] Specifically, the hyperparameter selection of a deep learning model directly determines the performance of the model. In this embodiment, one of the Bayesian optimization methods: sequential model-based optimization (SMBO) is implemented through the python-based hyperopt library.

[0060] When using Bayesian optimization to determine model parameters, it is necessary to define an objective function and optimized hyperparameter space. Since the training process of deep learning is actually a black box, the root mean square error (RMSE) of the mixed model is used as the objective function of model hyperparameter optimization:

[0061]

[0062] Among them, m is the number of...

example 1

[0112] The method proposed in this example uses the Bayesian optimization method to determine the optimal hyperparameters. This example will demonstrate the use of the Bayesian optimization method and verify the results.

[0113] First, the data used in the experiment is explained. The research shows that the PM2.5 sequence has strong nonlinearity and strong randomness. Therefore, this experiment uses the PM2.5 data set from the US State Department, which records the data from 2013 to 2017. The average concentration of PM2.5 per hour in Beijing in the past five years, with a total of 37704 entries, and the unit of data is μg / m 3 . Set the model prediction cycle to 24 steps, that is, the function realized by the model is to predict the value of the next 24 hours for the historical data of the previous 24 hours, Figure 5 The overall structure of the model of this embodiment is shown. The test was carried out on the hourly PM2.5 content data in Beijing air from March 22, 2016 ...

example 2

[0141] In Example 1, the Bayesian optimization algorithm is implemented and the feasibility is verified. In this example, the advantages of the model proposed in this embodiment (WD-GRU) in terms of accuracy are demonstrated by comparing with other models.

[0142] First, the data used in the experiment will be explained. The data set and test set used in this experiment are the same as those in Example 1, and the prediction period is also set to 24 steps.

[0143] In this example, a comparison is made with five combined models that also include time-series data decomposition and deep networks. The combined models used include Composition-ARIMA-GRU-GRU, EMD_RNN (EMD and RNN combined), EMDCNN_GRU (EMD, CNN and GRU combined), WD-RNN (wavelet decomposition and RNN combination), WD-LSTM (wavelet decomposition and LSTM Combination) and WD-GRU (combination of wavelet decomposition and GRU) proposed in this embodiment.

[0144] Figure 7 The prediction results of these six models a...

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 provides a time sequence prediction method based on Bayesian optimization and wavelet decomposition, and the method comprises: optimizing a model hyper-parameter according to a Bayesianoptimization method, and obtaining an optimal hyper-parameter which comprises the number of wavelet decomposition layers, a mother wavelet function in wavelet decomposition, and a hyper-parameter of aGRU sub-predictor; acquiring acquired data, and performing wavelet decomposition on the acquired data according to the wavelet decomposition layer number obtained after optimization and a mother wavelet function in wavelet decomposition to obtain a decomposition result; building a GRU-based sub-predictor, and learning and predicting the decomposition result according to the hyper-parameters of the GRU-based sub-predictor obtained after optimization to obtain a training result; and obtaining a prediction result according to the training result. The Bayesian optimization algorithm is used for optimizing the hyper-parameters, and the method has very high accuracy in a long-term time sequence prediction task.

Description

technical field [0001] The present application relates to the field of time series forecasting, in particular to a time series forecasting method and device based on Bayesian optimization and wavelet decomposition. Background technique [0002] With the continuous advancement of industrialization and urbanization, the rapid development of information storage, sensor networks and computer technology, technologies such as the Internet have gradually played an important role in people's lives. A large amount of information comes from various interactive tasks on the Internet. Most of this information is a time series continuously generated at the same time interval, such as the temperature of meteorological monitoring stations, the concentration of atmospheric PM2.5 and other data. These data not only It is a simple record of historical events, and at the same time they store a lot of useful information, such as the temperature data of meteorological monitoring stations includi...

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): G06F17/14G06F17/16G06F17/18G06N3/04G06N3/08G06N7/00
CPCG06F17/148G06F17/16G06F17/18G06N3/08G06N7/01G06N3/045
Inventor 金学波张家辉苏婷立白玉廷孔建磊
Owner BEIJING TECHNOLOGY AND BUSINESS UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com