Significance of Underestimation
Underestimation biases, as highlighted by Environmental Sciences, occur when models trained on low-severity events are used to predict high-severity events. This is a crucial consideration in environmental modeling. The practice of using low-severity data to predict high-severity events can lead to a significant underestimation of the potential impacts and risks associated with the high-severity events.
Synonyms: Belittling, Depreciation, Disregard, Minimization, Understatement, Miscalculation
The below excerpts are indicatory and do represent direct quotations or translations. It is your responsibility to fact check each reference.
The concept of Underestimation in scientific sources
Underestimation, according to regional sources, occurs when a model trained on low-severity events is tested against high-severity events. This bias is an important consideration when evaluating model performance.
From: Sustainability Journal (MDPI)
(1) It is a category of estimation error where the estimated value is lower than the actual value, and in the two-classification, the error label is set to 0 for underestimation.[1] (2) Underestimation biases occur when a training dataset of low-severity events is used to test a high-severity event, which is an important consideration.[2]