Linear Error in Probability Space (LEPS)
Calculates the linear error in probability spaces. This is the mean absolute difference between the forecast cumulative distribution value (cdf) and the observation. This function creates the empirical cdf function for the observations using the sample population. Linear interpretation is used to estimate the cdf values between observation values. Therefore; this may produce awkward results with small datasets.
leps(x, pred, plot = TRUE, ... )
x |
A vector of observations or a verification object with “cont.cont” properties. |
pred |
A vector of predictions. |
plot |
Logical to generate a plot or not. |
... |
Additional plotting options. |
If assigned to an object, the following values are reported.
leps.0 |
Negatively oriented score on the [0,1] scale, where 0 is a perfect score. |
leps.1 |
Positively oriented score proposed by Potts. |
Matt Pocernich
DeQue, Michel. (2003) “Continuous Variables” Chapter 5, Forecast Verification: A Practitioner's Guide in Atmospheric Science.
Potts, J. M., Folland, C.K., Jolliffe, I.T. and Secton, D. (1996) “Revised ‘LEPS’ scores fore assessing climate model simulations and long-range forecasts.” J. Climate, 9, pp. 34-54.
obs <- rnorm(100, mean = 1, sd = sqrt(50)) pred<- rnorm(100, mean = 10, sd = sqrt(500)) leps(obs, pred, main = "Sample Plot") ## values approximated OBS <- c(2.7, 2.9, 3.2, 3.3, 3.4, 3.4, 3.5, 3.8, 4, 4.2, 4.4, 4.4, 4.6, 5.8, 6.4) PRED <- c(2.05, 3.6, 3.05, 4.5, 3.5, 3.0, 3.9, 3.2, 2.4, 5.3, 2.5, 2.8, 3.2, 2.8, 7.5) a <- leps(OBS, PRED) a
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.