Flood forecasting method based on stack self-encoder and support vector regression

A technology of support vector regression and stack auto-encoding, which is used in prediction, neural learning methods, instruments, etc., to achieve the effect of a small amount of dimension enhancement and good robustness

Inactive Publication Date: 2019-01-22
HOHAI UNIV
View PDF3 Cites 13 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

SAE also employs sparsity constraints, which are limited when doing the backward pass

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
  • Flood forecasting method based on stack self-encoder and support vector regression
  • Flood forecasting method based on stack self-encoder and support vector regression
  • Flood forecasting method based on stack self-encoder and support vector regression

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0038] Embodiments of the invention are described in detail below, examples of which are illustrated in the accompanying drawings. The embodiments described below by referring to the figures are exemplary only for explaining the present invention and should not be construed as limiting the present invention.

[0039] For large data sets, a single model often cannot achieve good prediction results. To solve this problem, SOM clustering can be used to divide the entire input space into several disjoint regions, that is, to classify the input data, and then predict the model for each Partitions for training and prediction. At the same time, in view of the limitations of the hidden layer of ANN on non-convex problems, it can be solved by combining the ability of SAE to deeply extract the characteristics of the data set and the strong prediction ability of SVR. Based on this idea, the present invention proposes a flood prediction method based on self-organizing network, stacked au...

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 flood forecasting method based on stack self-encoder and support vector regression. The method includes: firstly classifying the original hydrological data through SOM clustering, then training the network by layer-by-layer greedy training method, and training the first sparse encoder by each sample set after clustering, and obtaining the first layer feature, then training the second sparse automatic encoder by using the first layer features, and so on, until the SAE training of the N layer is completed, using the output of the deepest hidden layer of SAE as the inputof SVR to train the SVR model, comparing the error between the SVR output and the actual value to adjust the entire SAE-SVR model parameters, finally, clustering the test set by SOM and predicting byoptimized SAE-SVR model. The invention integrates the advantages of SAE in extracting data features and SVR in forecasting time series, and has good accuracy in forecasting flood.

Description

technical field [0001] The invention relates to a flood prediction method, in particular to a flood prediction method integrating a stacked self-encoder and a support vector regression, and belongs to the technical field of water conservancy disaster prevention and mitigation. Background technique [0002] Flood forecasting is one of the few feasible solutions for flood management, and it is also the basis for decision-making of flood control and emergency rescue. Its timeliness and accuracy are extremely important. In the past, hydrological models based on physics or concepts have been fully researched and developed; today, thanks to the improvement of hydrological data acquisition capabilities, the improvement of data management and modeling tools, and the development of intelligent computing and machine learning, how to It is an important research direction to effectively use existing resources to mine useful information contained in hydrological and meteorological data a...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06Q10/04G06N3/08G06K9/62
CPCG06N3/08G06Q10/04G06F18/214Y02A10/40
Inventor 刘凡杨丽洁毛莺池许峰
Owner HOHAI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap