@article{fdi:010067232,
  title = {{N}eural approach to inverting complex system : application to ocean salinity profile estimation from surface parameters},
  author = {{G}ueye, {M}.{B}. and {N}iang, {A}. and {A}rnault, {S}abine and {T}hiria, {S}. and {C}r{\'e}pon, {M}.},
  editor = {},
  language = {{ENG}},
  abstract = {{A} neural network model is proposed for reconstructing ocean salinity profiles from sea surface parameters only. {T}he method is applied to the tropical {A}tlantic. {P}rior data mining on a complete dataset shows that latitude and sea surface salinity are the most relevant surface parameters in the prediction of salinity profiles. {A} classification using a self-organizing map learned on a large multivariate dataset is able to retrieve the most probable vertical salinity profiles from the surface parameters only. {B}oth in situ and modelled oceanic data are used to evaluate the results. {T}he reconstruction misses some salinity features in areas with high time-space variability in which the limited available dataset was unable to provide the complete variability ranges during the learning process. {H}owever, apart from these restricted areas, the salinity profiles are reproduced with correlations greater than 0.95 for most of the profiles of the test set.},
  keywords = {{ATLANTIQUE} ; {OCEAN} {INDIEN}},
  booktitle = {},
  journal = {{C}omputers and {G}eosciences},
  volume = {72},
  numero = {},
  pages = {201--209},
  ISSN = {0098-3004},
  year = {2014},
  DOI = {10.1016/j.cageo.2014.07.012},
  URL = {https://www.documentation.ird.fr/hor/fdi:010067232},
}