Predicting High Loss, Low Frequency Agricultural Drought Events: An Example Using Machine Learning for Cereal Crops in Eth Conference Paper uri icon