Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data
The timely and reliable prediction of crop yields on a larger scale is crucial for ensuring a stable food supply and food security. In the last few years, many studies have demonstrated that deep learning can offer reliable solutions for crop yield prediction. However, a key challenge in applying de...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Multidisciplinary Digital Publishing Institute
2024
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/118396/ http://psasir.upm.edu.my/id/eprint/118396/1/118396.pdf |
| _version_ | 1848867507571523584 |
|---|---|
| author | Joshi, Abhasha Pradhan, Biswajeet Chakraborty, Subrata Varatharajoo, Renuganth Gite, Shilpa Alamri, Abdullah |
| author_facet | Joshi, Abhasha Pradhan, Biswajeet Chakraborty, Subrata Varatharajoo, Renuganth Gite, Shilpa Alamri, Abdullah |
| author_sort | Joshi, Abhasha |
| building | UPM Institutional Repository |
| collection | Online Access |
| description | The timely and reliable prediction of crop yields on a larger scale is crucial for ensuring a stable food supply and food security. In the last few years, many studies have demonstrated that deep learning can offer reliable solutions for crop yield prediction. However, a key challenge in applying deep-learning models to crop yield prediction is their reliance on extensive training data, which are often lacking in many parts of the world. To address this challenge, this study introduces TrAdaBoost.R2, along with fine-tuning and domain-adversarial neural network deep-transfer-learning strategies, for predicting the winter wheat yield across diverse climatic zones in the USA. All methods used the bidirectional LSTM (BiLSTM) architecture to leverage its sequential feature extraction capabilities. The proposed transfer-learning approaches outperformed the baseline deep-learning model, with mean absolute error reductions ranging from 9% to 28%, demonstrating the effectiveness of these methods. Furthermore, the results demonstrate that the semi-supervised transfer-learning approach using the two-stage version of TrAdaBoost.R2 and fine-tuning achieved a superior performance compared to the domain-adversarial neural network and standard TrAdaBoost.R2. Additionally, the study offers insights for improving the accuracy and generalizability of crop yield prediction models in diverse agricultural landscapes across different regions. |
| first_indexed | 2025-11-15T14:37:36Z |
| format | Article |
| id | upm-118396 |
| institution | Universiti Putra Malaysia |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-15T14:37:36Z |
| publishDate | 2024 |
| publisher | Multidisciplinary Digital Publishing Institute |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | upm-1183962025-07-09T04:27:35Z http://psasir.upm.edu.my/id/eprint/118396/ Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data Joshi, Abhasha Pradhan, Biswajeet Chakraborty, Subrata Varatharajoo, Renuganth Gite, Shilpa Alamri, Abdullah The timely and reliable prediction of crop yields on a larger scale is crucial for ensuring a stable food supply and food security. In the last few years, many studies have demonstrated that deep learning can offer reliable solutions for crop yield prediction. However, a key challenge in applying deep-learning models to crop yield prediction is their reliance on extensive training data, which are often lacking in many parts of the world. To address this challenge, this study introduces TrAdaBoost.R2, along with fine-tuning and domain-adversarial neural network deep-transfer-learning strategies, for predicting the winter wheat yield across diverse climatic zones in the USA. All methods used the bidirectional LSTM (BiLSTM) architecture to leverage its sequential feature extraction capabilities. The proposed transfer-learning approaches outperformed the baseline deep-learning model, with mean absolute error reductions ranging from 9% to 28%, demonstrating the effectiveness of these methods. Furthermore, the results demonstrate that the semi-supervised transfer-learning approach using the two-stage version of TrAdaBoost.R2 and fine-tuning achieved a superior performance compared to the domain-adversarial neural network and standard TrAdaBoost.R2. Additionally, the study offers insights for improving the accuracy and generalizability of crop yield prediction models in diverse agricultural landscapes across different regions. Multidisciplinary Digital Publishing Institute 2024-12-23 Article PeerReviewed text en cc_by_4 http://psasir.upm.edu.my/id/eprint/118396/1/118396.pdf Joshi, Abhasha and Pradhan, Biswajeet and Chakraborty, Subrata and Varatharajoo, Renuganth and Gite, Shilpa and Alamri, Abdullah (2024) Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data. Remote Sensing, 16 (24). art. no. 4804. pp. 1-15. ISSN 2072-4292 https://www.mdpi.com/2072-4292/16/24/4804 10.3390/rs16244804 |
| spellingShingle | Joshi, Abhasha Pradhan, Biswajeet Chakraborty, Subrata Varatharajoo, Renuganth Gite, Shilpa Alamri, Abdullah Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data |
| title | Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data |
| title_full | Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data |
| title_fullStr | Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data |
| title_full_unstemmed | Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data |
| title_short | Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data |
| title_sort | deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data |
| url | http://psasir.upm.edu.my/id/eprint/118396/ http://psasir.upm.edu.my/id/eprint/118396/ http://psasir.upm.edu.my/id/eprint/118396/ http://psasir.upm.edu.my/id/eprint/118396/1/118396.pdf |