Improving generalisability and transferability of machine-learning-based maize yield prediction model through domain adaptation
Improving generalisability and transferability of machine-learning-based maize yield prediction model through domain adaptation
Modern problems in agricultural management require non-traditional solutions, one of which is by utilizing domain adaptive machine learning models for crop yield prediction which are able to perform reliably in different temporal or spatial domains. However, most studies have focused on the application of domain adaptation to classification tasks such as crop type identification, while the application to regression tasks such as crop yield prediction have been limited. In this study, we explore the generalisability and transferability of ordinary Deep Neural Network (DNN) and domain adaptive neural network models created using three domain adaptation algorithms, namely Discriminative Adversarial Neural Network (DANN), Kullback-Leibler Importance Estimation Procedure (KLIEP), and Regular Transfer Neural Network (RTNN). These three algorithms represent feature-based, instance-based, and parameter-based domain adaptations, respectively. Maize yield records, weather variables, and remotely sensed features from 11 states in the US corn belt acquired in 2006–2020 were compiled and segregated into classes according to temporal (year) and spatial characteristics (annual growing degree days [GDD], vapor pressure deficit [VPD], soil organic content [SOC], and green chlorophyll vegetation index/GCI). We found that models trained using datasets from temperate regions with medium-high GDD and moderate VPD perform well whereas SOC does not significantly affect the generalisability. It is not advisable to train models with datasets constrained by GCI as this feature correlates significantly with the maize yield, and adaptation between two domains that rarely intercept will not work well. We also demonstrate that Kullback-Leibler divergence computed using features from source and target domains can be used to justify the feasibility of domain adaptation. Based on the divergence, a model trained in the US (or another region with sufficient data) is expected to work reliably in other regions through domain adaptation, especially feature-based adaptation.
Domain adaptation, Generalisability, Machine learning, Transferability, Yield prediction
Priyatikanto, Rhorom
c250c3ca-958c-46a2-969a-3ad689b8630b
Lu, Yang
6d9d9d4f-3177-4265-b03b-34d7129ec95c
Dash, Jadu
51468afb-3d56-4d3a-aace-736b63e9fac8
Sheffield, Justin
dd66575b-a4dc-4190-ad95-df2d6aaaaa6b
15 October 2023
Priyatikanto, Rhorom
c250c3ca-958c-46a2-969a-3ad689b8630b
Lu, Yang
6d9d9d4f-3177-4265-b03b-34d7129ec95c
Dash, Jadu
51468afb-3d56-4d3a-aace-736b63e9fac8
Sheffield, Justin
dd66575b-a4dc-4190-ad95-df2d6aaaaa6b
Priyatikanto, Rhorom, Lu, Yang, Dash, Jadu and Sheffield, Justin
(2023)
Improving generalisability and transferability of machine-learning-based maize yield prediction model through domain adaptation.
Agricultural and Forest Meteorology, 341, [109652].
(doi:10.1016/j.agrformet.2023.109652).
Abstract
Modern problems in agricultural management require non-traditional solutions, one of which is by utilizing domain adaptive machine learning models for crop yield prediction which are able to perform reliably in different temporal or spatial domains. However, most studies have focused on the application of domain adaptation to classification tasks such as crop type identification, while the application to regression tasks such as crop yield prediction have been limited. In this study, we explore the generalisability and transferability of ordinary Deep Neural Network (DNN) and domain adaptive neural network models created using three domain adaptation algorithms, namely Discriminative Adversarial Neural Network (DANN), Kullback-Leibler Importance Estimation Procedure (KLIEP), and Regular Transfer Neural Network (RTNN). These three algorithms represent feature-based, instance-based, and parameter-based domain adaptations, respectively. Maize yield records, weather variables, and remotely sensed features from 11 states in the US corn belt acquired in 2006–2020 were compiled and segregated into classes according to temporal (year) and spatial characteristics (annual growing degree days [GDD], vapor pressure deficit [VPD], soil organic content [SOC], and green chlorophyll vegetation index/GCI). We found that models trained using datasets from temperate regions with medium-high GDD and moderate VPD perform well whereas SOC does not significantly affect the generalisability. It is not advisable to train models with datasets constrained by GCI as this feature correlates significantly with the maize yield, and adaptation between two domains that rarely intercept will not work well. We also demonstrate that Kullback-Leibler divergence computed using features from source and target domains can be used to justify the feasibility of domain adaptation. Based on the divergence, a model trained in the US (or another region with sufficient data) is expected to work reliably in other regions through domain adaptation, especially feature-based adaptation.
Text
1-s2.0-S016819232300343X-main
- Version of Record
More information
Accepted/In Press date: 4 August 2023
e-pub ahead of print date: 9 August 2023
Published date: 15 October 2023
Additional Information:
Funding Information:
This work was funded through the “A new paradigm in precision agriculture assimilation of ultra-fine resolution data into a crop-yield forecasting model" project, supported by the King Abdullah University of Science and Technology , grant number OSR-2017-CRG6 , and through the “Building REsearch Capacity for sustainable water and food security In drylands of sub-Saharan Africa” (BRECcIA) project which is supported by UK Research and Innovation as part of the Global Challenges Research Fund , grant number NE/P021093/1 .
Publisher Copyright:
© 2023 The Authors
Keywords:
Domain adaptation, Generalisability, Machine learning, Transferability, Yield prediction
Identifiers
Local EPrints ID: 480970
URI: http://eprints.soton.ac.uk/id/eprint/480970
ISSN: 0168-1923
PURE UUID: 3740652c-6aa8-407f-8b93-8c1a68ad37b0
Catalogue record
Date deposited: 11 Aug 2023 17:33
Last modified: 18 Mar 2024 03:33
Export record
Altmetrics
Contributors
Author:
Rhorom Priyatikanto
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics