Skip to main navigation Skip to search Skip to main content

KNN regression model-based refinement of thermohaline data

  • Yu Gou
  • , Jun Liu
  • , Tong Zhang*
  • *Corresponding author for this work
  • Jilin University
  • CAS - Shenyang Institute of Automation

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper carries out a renement on the basis of existing data sets, whose level of granularity is not available for some experimental analysis such as thermocline research. The thermocline is sensitive to thermohaline data granularity for sudden sea temperature changes. We rened the data with the KNN regression method and managed to choose the optimal parameters for the construction of a prediction model. We also rened the temperature and salinity data in BOA_Argo using the regression forecast model. The original data, whose horizontal resolution is 1 °x 1 °and vertically divided into uneven 58 layers from the sea surface to 1,975 meters underwater, has been rened into a new set with the resolution of 1 °x 1 °horizontally and 1-meter interval vertically. At each point, we rened the previously uneven 58 temperature data samples into 1,976 evenly distributed data samples. The rened data sets can be used in experimental analysis, and the validity of this method has been veried by regional data.

Original languageEnglish
Title of host publicationProceedings of the 13th ACM International Conference on Underwater Networks and Systems, WUWNet 2018
PublisherAssociation for Computing Machinery
ISBN (Electronic)9781450361934
DOIs
StatePublished - 3 Dec 2018
Externally publishedYes
Event13th ACM International Conference on Underwater Networks and Systems, WUWNet 2018 - Shenzhen, China
Duration: 3 Dec 20185 Dec 2018

Publication series

NameACM International Conference Proceeding Series

Conference

Conference13th ACM International Conference on Underwater Networks and Systems, WUWNet 2018
Country/TerritoryChina
CityShenzhen
Period3/12/185/12/18

Keywords

  • BOA_ARGO
  • Granularity
  • KNN
  • Thermocline
  • Thermohaline data

Fingerprint

Dive into the research topics of 'KNN regression model-based refinement of thermohaline data'. Together they form a unique fingerprint.

Cite this