Marine weather prediction method and system based on space-time hybrid convolution attention

The invention discloses a multivariate space-time prediction method and system for marine weather forecast. The method comprises the following steps: S1, making a sea surface weather data set; s2, constructing an StHCFFormer model, training the StHCFFormer model by using the sea surface weather data...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: ZHAO HANQING, YU HANGYI, WANG JUNKAI, LIN LIANLEI, ZHANG ZONGWEI, GAO SHENG
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator ZHAO HANQING
YU HANGYI
WANG JUNKAI
LIN LIANLEI
ZHANG ZONGWEI
GAO SHENG
description The invention discloses a multivariate space-time prediction method and system for marine weather forecast. The method comprises the following steps: S1, making a sea surface weather data set; s2, constructing an StHCFFormer model, training the StHCFFormer model by using the sea surface weather data set as a sample, and carrying out training and optimization; and S3, based on the trained StHCFFormer model, inputting sea surface weather data to carry out sea surface weather prediction. According to the method, global representation and local features of a space are deeply explored by using a convolution-space self-attention mixing mechanism, a time dependency relationship of time sequence data is captured by using time sequence self-attention, and channel-level scaling is performed through a local feature extraction module to capture dynamic correlation of multiple variables; and meanwhile, uncertainty related loss is introduced to dynamically adjust the multi-task weight, so that the globally optimal solution
format Patent
fullrecord <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_CN118296453A</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>CN118296453A</sourcerecordid><originalsourceid>FETCH-epo_espacenet_CN118296453A3</originalsourceid><addsrcrecordid>eNqNjEEKwjAURLNxIeodvgfoolZFl1IUN7pyKZTfZEoCbVKSr9Lbq8UDuJqBN2-m6n7h6DzoBRaLSH2EcVpc8NRBbDDE3lAakqCjmhMMfVDqWSMT14HsUEdnSAf_DO1jFFkE_tvmatJwm7D45UwtT8dbec7QhwrjiYdU5TXPd6v9dr0pDsU_mzfoJDxj</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Marine weather prediction method and system based on space-time hybrid convolution attention</title><source>esp@cenet</source><creator>ZHAO HANQING ; YU HANGYI ; WANG JUNKAI ; LIN LIANLEI ; ZHANG ZONGWEI ; GAO SHENG</creator><creatorcontrib>ZHAO HANQING ; YU HANGYI ; WANG JUNKAI ; LIN LIANLEI ; ZHANG ZONGWEI ; GAO SHENG</creatorcontrib><description>The invention discloses a multivariate space-time prediction method and system for marine weather forecast. The method comprises the following steps: S1, making a sea surface weather data set; s2, constructing an StHCFFormer model, training the StHCFFormer model by using the sea surface weather data set as a sample, and carrying out training and optimization; and S3, based on the trained StHCFFormer model, inputting sea surface weather data to carry out sea surface weather prediction. According to the method, global representation and local features of a space are deeply explored by using a convolution-space self-attention mixing mechanism, a time dependency relationship of time sequence data is captured by using time sequence self-attention, and channel-level scaling is performed through a local feature extraction module to capture dynamic correlation of multiple variables; and meanwhile, uncertainty related loss is introduced to dynamically adjust the multi-task weight, so that the globally optimal solution</description><language>chi ; eng</language><subject>CALCULATING ; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS ; COMPUTING ; COUNTING ; ELECTRIC DIGITAL DATA PROCESSING ; MEASURING ; METEOROLOGY ; PHYSICS ; TESTING</subject><creationdate>2024</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20240705&amp;DB=EPODOC&amp;CC=CN&amp;NR=118296453A$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25564,76547</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20240705&amp;DB=EPODOC&amp;CC=CN&amp;NR=118296453A$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>ZHAO HANQING</creatorcontrib><creatorcontrib>YU HANGYI</creatorcontrib><creatorcontrib>WANG JUNKAI</creatorcontrib><creatorcontrib>LIN LIANLEI</creatorcontrib><creatorcontrib>ZHANG ZONGWEI</creatorcontrib><creatorcontrib>GAO SHENG</creatorcontrib><title>Marine weather prediction method and system based on space-time hybrid convolution attention</title><description>The invention discloses a multivariate space-time prediction method and system for marine weather forecast. The method comprises the following steps: S1, making a sea surface weather data set; s2, constructing an StHCFFormer model, training the StHCFFormer model by using the sea surface weather data set as a sample, and carrying out training and optimization; and S3, based on the trained StHCFFormer model, inputting sea surface weather data to carry out sea surface weather prediction. According to the method, global representation and local features of a space are deeply explored by using a convolution-space self-attention mixing mechanism, a time dependency relationship of time sequence data is captured by using time sequence self-attention, and channel-level scaling is performed through a local feature extraction module to capture dynamic correlation of multiple variables; and meanwhile, uncertainty related loss is introduced to dynamically adjust the multi-task weight, so that the globally optimal solution</description><subject>CALCULATING</subject><subject>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>ELECTRIC DIGITAL DATA PROCESSING</subject><subject>MEASURING</subject><subject>METEOROLOGY</subject><subject>PHYSICS</subject><subject>TESTING</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2024</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNqNjEEKwjAURLNxIeodvgfoolZFl1IUN7pyKZTfZEoCbVKSr9Lbq8UDuJqBN2-m6n7h6DzoBRaLSH2EcVpc8NRBbDDE3lAakqCjmhMMfVDqWSMT14HsUEdnSAf_DO1jFFkE_tvmatJwm7D45UwtT8dbec7QhwrjiYdU5TXPd6v9dr0pDsU_mzfoJDxj</recordid><startdate>20240705</startdate><enddate>20240705</enddate><creator>ZHAO HANQING</creator><creator>YU HANGYI</creator><creator>WANG JUNKAI</creator><creator>LIN LIANLEI</creator><creator>ZHANG ZONGWEI</creator><creator>GAO SHENG</creator><scope>EVB</scope></search><sort><creationdate>20240705</creationdate><title>Marine weather prediction method and system based on space-time hybrid convolution attention</title><author>ZHAO HANQING ; YU HANGYI ; WANG JUNKAI ; LIN LIANLEI ; ZHANG ZONGWEI ; GAO SHENG</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_CN118296453A3</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>chi ; eng</language><creationdate>2024</creationdate><topic>CALCULATING</topic><topic>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>ELECTRIC DIGITAL DATA PROCESSING</topic><topic>MEASURING</topic><topic>METEOROLOGY</topic><topic>PHYSICS</topic><topic>TESTING</topic><toplevel>online_resources</toplevel><creatorcontrib>ZHAO HANQING</creatorcontrib><creatorcontrib>YU HANGYI</creatorcontrib><creatorcontrib>WANG JUNKAI</creatorcontrib><creatorcontrib>LIN LIANLEI</creatorcontrib><creatorcontrib>ZHANG ZONGWEI</creatorcontrib><creatorcontrib>GAO SHENG</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>ZHAO HANQING</au><au>YU HANGYI</au><au>WANG JUNKAI</au><au>LIN LIANLEI</au><au>ZHANG ZONGWEI</au><au>GAO SHENG</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Marine weather prediction method and system based on space-time hybrid convolution attention</title><date>2024-07-05</date><risdate>2024</risdate><abstract>The invention discloses a multivariate space-time prediction method and system for marine weather forecast. The method comprises the following steps: S1, making a sea surface weather data set; s2, constructing an StHCFFormer model, training the StHCFFormer model by using the sea surface weather data set as a sample, and carrying out training and optimization; and S3, based on the trained StHCFFormer model, inputting sea surface weather data to carry out sea surface weather prediction. According to the method, global representation and local features of a space are deeply explored by using a convolution-space self-attention mixing mechanism, a time dependency relationship of time sequence data is captured by using time sequence self-attention, and channel-level scaling is performed through a local feature extraction module to capture dynamic correlation of multiple variables; and meanwhile, uncertainty related loss is introduced to dynamically adjust the multi-task weight, so that the globally optimal solution</abstract><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier
ispartof
issn
language chi ; eng
recordid cdi_epo_espacenet_CN118296453A
source esp@cenet
subjects CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
MEASURING
METEOROLOGY
PHYSICS
TESTING
title Marine weather prediction method and system based on space-time hybrid convolution attention
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-23T16%3A10%3A28IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=ZHAO%20HANQING&rft.date=2024-07-05&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3ECN118296453A%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true