Title: Crop yields prediction using ensembles of machine learning algorithms

Authors: Mariam Temitope Anifowose; Joseph Damilola Akinyemi; Adejoke O. Olamiti

Addresses: Department of Computer Science, University of Ibadan, Ibadan, Oyo State, Nigeria ' Department of Computer Science, University of Ibadan, Ibadan, Oyo State, Nigeria ' Department of Computer Science, University of Ibadan, Ibadan, Oyo State, Nigeria

Abstract: Crop yield is crucial for food security and is often predicted using farmers' knowledge and experience. However, recent research has shown that machine learning (ML) can produce better predictions. This research developed an efficient predictive model for crop yield by creating ensembles of several ML algorithms, including extreme gradient boosting (XGBoost), decision tree regressor (DT), K-nearest neighbour regressor (KNN), and random forest regressor (RF). Historical crop yield data from the Food and Agriculture Organization (FAO) data repository and climate data from the Climate Change Portal were integrated for analysis and prediction. The mean-ensemble and stacking approaches were used to create ensembles of the ML algorithms. The experimental results showed that stacked-KNN and stacked-DT had the best performance, with an R²-score of 97% and 96% respectively. The ability to accurately forecast crop yields using historical data could enhance agricultural productivity, inform policymakers' decisions and ensure food security in Nigeria.

Keywords: crop yield prediction; machine learning; ensemble machine learning; stacking.

DOI: 10.1504/IJSSS.2024.144737

International Journal of Society Systems Science, 2024 Vol.15 No.2, pp.147 - 169

Received: 28 Jul 2023
Accepted: 06 Dec 2024
Published online: 28 Feb 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article