Machine Learning Approach for Prediction of Area Under Cultivation and Production for Vegetatively Propagated Crops
| cg.authorship.types | CGIAR and developing country institute | en_US |
| cg.contributor.initiative | Seed Equal | en_US |
| cg.creator.identifier | Kwame Ogero: 0000-0002-5141-6781 | en_US |
| cg.creator.identifier | NAMANDA SAM: 0000-0001-7822-0626 | en_US |
| cg.identifier.doi | https://doi.org/10.4160/cip.2025.01.020 | en_US |
| cg.subject.actionArea | Resilient Agrifood Systems | en_US |
| cg.subject.cip | CROP AND SYSTEMS SCIENCES CSS | en_US |
| cg.subject.cip | SEED SYSTEMS | en_US |
| cg.subject.cip | FOOD SECURITY | en_US |
| cg.subject.cip | BIGDATA | en_US |
| cg.subject.cip | SWEETPOTATOES | en_US |
| cg.subject.impactArea | Nutrition, health and food security | en_US |
| cg.subject.impactArea | Climate adaptation and mitigation | en_US |
| cg.subject.sdg | SDG 2 - Zero hunger | en_US |
| cg.subject.sdg | SDG 9 - Industry, innovation and infrastructure | en_US |
| cg.subject.sdg | SDG 12 - Responsible consumption and production | en_US |
| cg.subject.sdg | SDG 13 - Climate action | en_US |
| dc.contributor.author | Ahishakiye, E. | en_US |
| dc.contributor.author | Ogero, K. | en_US |
| dc.contributor.author | Namada, S. | en_US |
| dc.contributor.author | Rajendran, S. | en_US |
| dc.date.accessioned | 2025-02-01T02:34:44Z | en_US |
| dc.date.available | 2025-02-01T02:34:44Z | en_US |
| dc.identifier.uri | https://hdl.handle.net/10568/172714 | en_US |
| dc.title | Machine Learning Approach for Prediction of Area Under Cultivation and Production for Vegetatively Propagated Crops | en_US |
| dcterms.abstract | Vegetatively propagated crops (VPCs) such as cassava, sweet potatoes, and bananas, are a key component in ensuring food security for the low- and middle-income countries (LMICs). In agricultural planning and seed system management, it is essential to accurately predict the area under cultivation, production volumes, and yield rates of these crops. Traditional forecasting methods have fallen short in capturing the complexity of VPC production, as there are nonlinear relationships and dynamic environmental factors at play. This paper overcomes these shortcomings by using machine learning models to enhance the forecasting accuracy using data from the Seed Requirement Estimation (SRE) tool. We applied Random Forest, AdaBoost, and a Stacked Ensemble Model to forecast the area under cultivation and production volume in tons. After hyperparameter tuning, the Stacked Model performed better, yielding R² values of 0.8260 for area prediction and 0.7883 for production forecasting, outperforming the individual models. The results reflect the potential of the ensemble learning model to improve the accuracy of agricultural forecasts. The study emphasizes the role that advanced predictive models can play in enhancing agricultural policy decisions based on data, optimizing seed distribution, and ensuring food security in VPC-dependent regions. | en_US |
| dcterms.accessRights | Open Access | en_US |
| dcterms.audience | Academics | en_US |
| dcterms.audience | CGIAR | en_US |
| dcterms.audience | Development Practitioners | en_US |
| dcterms.audience | Donors | en_US |
| dcterms.audience | Extension | en_US |
| dcterms.audience | Farmers | en_US |
| dcterms.audience | General Public | en_US |
| dcterms.audience | NGOs | en_US |
| dcterms.audience | Policy Makers | en_US |
| dcterms.audience | Scientists | en_US |
| dcterms.bibliographicCitation | Ahishakiye, E.; Ogero, K.; Namada, S.; Rajendran, S. 2024. Machine Learning Approach for Prediction of Area Under Cultivation and Production for Vegetatively Propagated Crops. International Potato Center. 25 p. DOI: 10.4160/cip.2025.01.020 | en_US |
| dcterms.extent | 25 p. | en_US |
| dcterms.issued | 2024-12 | en_US |
| dcterms.language | en | en_US |
| dcterms.license | CC-BY-4.0 | en_US |
| dcterms.subject | machine learning | en_US |
| dcterms.subject | crop yield | en_US |
| dcterms.subject | seed systems | en_US |
| dcterms.subject | food security | en_US |
| dcterms.type | Working Paper | en_US |