I have a dataset on wildfires that I fitted to a Gumbel distribution with a set of covariates (using the gevrFit function in the eva package in R). The result of the model returned the coefficients for each of the covariates, but there are two, one for the location parameter and one for the scale parameter. However, I'm having trouble wrapping my head around what the results mean. Specifically, since they seem to be used to fit the model parameters, I'm having trouble visualizing what each of the coefficient mean for the dependent variable (fire size in this particular case).
In addition, I'm also having a hard time understanding how the formula for the location and scale variables should be formed. It is not intuitive to me to try and think about the relationship between covariates with the model distribution parameters rather than for the dependent variable.
I don't have a easily reproducible data set for people, but I have attached both the model and the R model summary here.
model.gum.temp <- gevrFit(test.data$logsize, method = "mle", locvars = coef.all1,
locform = ~tmax + tmin + prcp, scalevars = coef.all1, scaleform = ~tmax +tmin, gumbel = TRUE)
Summary of fit:
Estimate Std. Error z value Pr(>|z|)
Location (Intercept) -0.96941900 0.0715180 -13.55489 7.4123e-42 ***
Location tmax 0.00434970 0.0017680 2.46023 1.3885e-02 *
Location tmin -0.00103350 0.0019255 -0.53675 5.9144e-01
Location prcp -0.14147062 0.0870534 -1.62510 1.0414e-01
Scale (Intercept) 0.57545500 0.0485126 11.86196 1.8655e-32 ***
Scale tmax 0.00157053 0.0011341 1.38478 1.6612e-01
Scale tmin 0.00078513 0.0012853 0.61086 5.4129e-01
---
Signif. codes: 0 '***' 0.001 '*' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Thanks everyone!
