Published August 12, 2009 | Version v1

Machine Learning Models for Climate Prediction and Adaptation in Zimbabwean Environments

  • 1. Department of Artificial Intelligence, Midlands State University
  • 2. Midlands State University

Description

Zimbabwe is vulnerable to climate variability and change, necessitating robust prediction models for effective adaptation planning. A hybrid ensemble model combining Random Forest (RF) and XGBoost was employed to predict monthly rainfall. Model performance was evaluated using mean absolute error (MAE). The hybrid RF-XGBoost model achieved a MAE of 10.5 mm, indicating moderate accuracy in predicting rainfall. Machine learning models can be effectively utilised for climate prediction in Zimbabwean settings, supporting informed adaptation planning. Further research should explore the integration of socio-economic data to enhance model predictive capabilities and validate findings across different regions. machine learning, climate prediction, XGBoost, Random Forest, rainfall forecasting Model estimation used $\hat{\theta}=argmin_{\theta}\sum_i\ell(y_i,f_\theta(x_i))+\lambda\lVert\theta\rVert_2^2$, with performance evaluated using out-of-sample error.

Files

zenodo.18900896.pdf

Files (97.6 kB)

Name Size
md5:a6fe92d97b403f535c03446bfc005154
16.0 kB Download
md5:92195d8654fe4176b3f0f24d78ec1ac1
81.6 kB Preview Download