R/summary.altf2.R

In fDMA: Dynamic Model Averaging and Dynamic Model Selection for Continuous Outcomes

summary.altf2 <- function(object, ...)
  {
   x <- object
   
   nmods <- length(x$coeff.)
   
   cat("Mean coefficients: ")
   cat("\n")
   
   c <- matrix(,ncol=ncol(x$coeff.[[1]]),nrow=nmods)
   for (i in 1:nmods)
    {
      c[i,] <- colMeans(x$coeff.[[i]],na.rm=TRUE)
    }
   colnames(c) <- colnames(x$coeff.[[1]])
   rownames(c) <- names(x$coeff.)
   c <- round(c,digits=4)
   print(c,quote=FALSE)
   cat("\n")

   
  for (j in 1:nmods)
    {
      probs <- round(colMeans(x$rel.var.imp.[[j]],na.rm=TRUE),digits=2)
      min.probs <- round(apply(x$rel.var.imp.[[j]],2,min, na.rm=TRUE),digits=2)
      max.probs <- round(apply(x$rel.var.imp.[[j]],2,max, na.rm=TRUE),digits=2)
     
      inc <- vector()

      for (i in 1:length(probs))
        {
          rvi1 <- as.vector(x$rel.var.imp.[[j]][,i])
          rvi2 <- na.exclude(rvi1)
          rvi2 <- rvi2[rvi2>0.5]
          inc[i] <- round(length(rvi2)/length(rvi1),digits=2)
        }
      rm(i)
     
      s <- cbind(min.probs,probs,max.probs,as.matrix(inc))
      colnames(s) <- c("min rvi", "mean rvi", "max rvi", "inc")
      rownames(s) <- colnames(x$coeff.[[1]])
      temp <- names(x$rel.var.imp.)[j]
      names(temp) <- c("")
      print(temp,quote=FALSE)
      cat("\n")
      print(s.quote=FALSE)
      cat("\n")
    }
   cat("\n")
   cat("* rvi - relative variable importance")
   cat("\n")
   cat("* inc - frequency when relative variable importance > 1/2")
   cat("\n")
   cat("\n")

   if (! all(names(x$coeff.)==c("av. TVP")))
    {
      cat("Frequency when p-values for t-test are less than: ")
    }
   cat("\n")
   
   for (k in 1:nmods)
    {
      if (! names(x$p.val.)[k]=="av. TVP")
       {
         vv <- rep.int(NA,3)
         for (i in 1:ncol(x$p.val.[[1]]))
          {
             v <- as.vector(na.exclude(x$p.val.[[k]][,i,drop=FALSE]))
             v1 <- v[v<0.01]
             v2 <- v[v<0.05]
             v3 <- v[v<0.10]
             v1 <- length(v1) / length(v)
             v2 <- length(v2) / length(v)
             v3 <- length(v3) / length(v)
             v <- c(v1,v2,v3)
             vv <- rbind(vv,v)
          }
         vv <- vv[-1,,drop=FALSE]
         vv <- t(vv)
         vv <- round(vv,digits=2)
         colnames(vv) <- colnames(x$coeff.[[1]])
         rownames(vv) <- c("0.01","0.05","0.10")
         temp <- names(x$p.val.)[k]
         names(temp) <- c("")
         print(temp,quote=FALSE)
         cat("\n")
         print(vv,quote=FALSE)
         cat("\n")
       }
    }
   
   cat("\n")
   cat("Forecast quality measures: ")
   cat("\n")
   print(x$summary,quote=FALSE)
   cat("\n")
  }