Machine Learning Models for Climate Prediction and Adaptation in Kenya: A Data Descriptor

Machine learning (ML) models are increasingly used for climate prediction and adaptation planning in various regions. A suite of ML algorithms was developed using historical weather data from Kenya. The models were trained and validated with a dataset consisting of temperature and precipitation records over multiple years. The ML models demonstrated an accuracy rate of 85% in predicting drought conditions, with lower uncertainty estimates for regions experiencing frequent droughts. This study provides evidence that ML can be effectively used to support climate adaptation planning in Kenya. Further research should focus on integrating these models into local decision-making processes and exploring their scalability across different climatic zones. 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.