Forecast verification and scoring of expert probablistic opinions
Theme: Bayesian Statistics
Web: Permalink to this page
When an expert or a forecasting system makes a prediction involving the probabilities of future events, it is important to be able to evaluate the prediction after the events in question have happened or not happened. In the context of weather forecasts, this process is called "verification". Because, with probability forecasts, most predictions could be described as "partly correct", it is not always straightforward to carry out the evaluation. In straightforward contexts, the theory behind forecast verification and evaluation is well understood, but there are many practical issues that need to be taken forward.