Software - Probabilistic Forecast Scoring

This page contains software for Matlab for scoring probabilistic time series forecasts, as described in [1]. If you use this code, please cite [2]. Download the following code:

pscrps.m	Calculate the Continuous-Ranked Probability Score (CRPS) for a given set of forecasts and observations.
pspit.m	Calculate the Probability Integral Transform (PIT) z-series for a given set of forecasts and observations. Also calculates the confidence interval for each bit of the histogram of the z-series.
psmae.m	Calculate the Mean Absolute Error (MAE) for a given set of forecasts and observations, issuing the median of each forecast distribution as a point forecast.
psmse.m	Calculate the Mean Square Error (MSE) for a given set of forecasts and observations, issuing the mean of each forecast distribution as a point forecast.
pstest.m	Demonstrates how to use the above routines.

Typing 'help (function name)' shows instructions for each routine.

[1] T. Gneiting, F. Balabdaoui, A.E. Raftery (2007), "Probabilistic Forecasts, Calibration and Sharpness", Journal of the Royal Statistical Society, 69(2):243-268

[2] M.A. Little et al. (2008), "Parsimonious Modeling of UK Daily Rainfall for Density Forecasting", in Geophysical Research Abstracts, EGU General Assembly, Vienna 2008, Volume 10.