initial values -> check for convergence of fit

NAMESPACE

@@ -28,6 +28,7 @@ importFrom(ismev,gev.fit)
 importFrom(pbapply,pbsapply)
 importFrom(stats,lm)
 importFrom(stats,make.link)
+importFrom(stats,median)
 importFrom(stats,optim)
 importFrom(zoo,as.zoo)
 importFrom(zoo,rollapplyr)

R/gevdfit.R

@@ -272,6 +272,7 @@ gev.d.fit<-
 #' @param link list of 5, link functions for parameters, created with \code{\link{make.link}}
 #' @return list of initail values for mu_tilde, sigma_0, xi, eta
 #' @importFrom stats lm 
+#' @importFrom stats median 
 #' @importFrom ismev gev.fit
 #' @keywords internal 
 
@@ -279,9 +280,10 @@ gev.d.init <- function(xdat,ds,link){
   durs <- unique(ds)
   mles <- matrix(NA, nrow=length(durs), ncol= 3)
   for(i in 1:length(durs)){
-    test <- try(mles[i,] <- gev.fit(xdat[ds==durs[i]],show = FALSE)$mle,silent = TRUE)
-    if("try-error" %in% class(test)){mles[i,] <- rep(NA,3)}
+    test <- try(fit <- gev.fit(xdat[ds==durs[i]],show = FALSE),silent = TRUE)
+    if("try-error" %in% class(test) | fit$conv!=0){mles[i,] <- rep(NA,3)}else{mles[i,] <- fit$mle}
   }
+  if(all(is.na(mles))){stop('Initial values could not be computed for this dataset.')}
   # get values for sig0 and eta (also mu_0) from linear model in log-log scale
   lmsig <- lm(log(mles[,2])~log(durs))
   lmmu <- lm(log(mles[,1])~log(durs))
@@ -289,13 +291,13 @@ gev.d.init <- function(xdat,ds,link){
   # sig0 <- exp Intercept
   siginit <- link$siglink$linkfun(exp(lmsig$coefficients[[1]]))
   # eta <- mean of negativ slopes 
-  etainit <- link$etalink$linkfun(mean(c(-lmsig$coefficients[[2]],-lmmu$coefficients[[2]])))
+  etainit <- link$etalink$linkfun(median(c(-lmsig$coefficients[[2]],-lmmu$coefficients[[2]])))
   # mean of mu_d/sig_d 
   # could try:
   # mu0/sig0 = exp(lmmu$coefficients[[1]])/exp(lmsig$coefficients[[1]])
-  muinit <- link$mulink$linkfun(mean(c(mles[,1]/mles[,2]),na.rm = TRUE))
+  muinit <- link$mulink$linkfun(median(c(mles[,1]/mles[,2]),na.rm = TRUE))
   # mean of shape parameters 
-  shinit <- link$shlink$linkfun(mean(mles[,3],na.rm = TRUE))
+  shinit <- link$shlink$linkfun(median(mles[,3],na.rm = TRUE))
   thetainit <- link$thetalink$linkfun(0)
   
   return(list(mu=muinit,sigma=siginit,xi=shinit,theta=thetainit,eta=etainit))