Forecasting future adoption rates of agroecological innovations using machine learning: perspectives from Burkina Faso’s 45 provinces from 2010 to 2020

Publication Type

Journal Article

Name of Author

• Theodore Nikiema, Institute of Mathematics and Physics Sciences, University of Abomey Calavi, Cotonou, BeninFollow
• Pamela Giselle Katic, Natural Resources Institute, University of Greenwich, Chatham, UK
• Issaka Abdou Razakou Kiribou, Africa Centre of Excellence for Climate Smart Agricultural and Biodiversity Conservation Haramaya University, Dire Dawa, Ethiopia
• Sylvain Kpenavoun Chogou, Faculty of Agronomic Sciences, University of Abomey Calavi, Cotonou, Benin
• Eugene C. Ezin, Institute of Training and Research in Computer Science, University of Abomey Calavi, Cotonou, Benin

Publication Date (Issue Year)

2026

Journal Name

International Journal of Agricultural Sustainability

Abstract

Implementation of agroecological innovations tends to be long-term processes, with the practices of one year often linked to those of previous years. However, previous studies have focused on understanding drivers of adoption at farm level, with adoption measured at a point in time. In this study, we use a decade of panel data from the Permanent Agricultural Survey in Burkina Faso from 2010 to 2020 and machine learning approaches, to model adoption rates of agroecological innovations at the provincial level as an autoregressive process. This modeling approach allows us to exploit the time series nature of our dataset to forecast future adoption rates. Our results showcase the potential of machine learning algorithms to improve the forecasting of agroecology adoption rates and provide a model that can be used as a base for proposing interventions to support the adoption of agroecological innovations. The LSTM model reached a R² of 75% compared to 27% for the ARIMA family baseline model. The framework we proposed allows the identification of priority areas for targeted interventions and provides a foundation on which future studies can be built to predict and track agroecology adoption rates over time.

Keywords

Agroecology, sustainable agriculture, climate change, time series, prediction, recurrent artificial neural network

Rsif Scholar Name

Theodore Nikiema

Rsif Scholar Nationality

Mozambique

Cohort

Cohort 4

Thematic Area

ICTs Including Big Data and Artificial Intelligence

Africa Host University (AHU)

Université d'Abomey-Calavi, Benin

Recommended Citation

Nikiema, T., Katic, P. G., Kiribou, I. A., Chogou, S. K., & Ezin, E. C. (2026). Forecasting future adoption rates of agroecological innovations using machine learning: perspectives from Burkina Faso’s 45 provinces from 2010 to 2020. International Journal of Agricultural Sustainability https://doi.org/10.1080/14735903.2026.2674472