Integrating Spatial Heterogeneity to Enhance Spatial Temporal Crop Yield Predictions
A Comparative Assessment of Machine Learning Models Incorporating Spatial Heterogeneity for Enhanced Crop Yield Predictions
Author:
Francis Michael Msangi
Advisors:
Prof.Dr.Edzer Pebesma
Prof.Dr. Ana Cristina Costa
Dr. Francis Kamau Muthoni
Author:
Francis Michael Msangi
Advisors:
Prof.Dr.Edzer Pebesma
Prof.Dr. Ana Cristina Costa
Dr. Francis Kamau Muthoni
Abstract:
This project focused on improving crop yield prediction by integrating spatial heterogeneity into machine learning models using the Geographically Weighted Random Forest (GWRF) approach. Using maize yield data from Zambia and Malawi, the study compared GWRF with the standard Random Forest (RF) model for farms practicing conservation agriculture (CA) and conventional practices (CP). Results showed that GWRF outperformed RF in both scenarios, demonstrating its potential for enhancing spatial-temporal crop yield predictions and identifying the area of model applicability.Full project available here.
Please feel free to contact me at my email address or through my LinkedIn Account below.
francismich196@gmail.com