Corrected a major bug in SSMcycle

Basically SSMcycle gave wrong system matrix T for all models, as
nondiagonal entries contained - sign in wrong place... No idea how this
has been unnoticed for so long.

ChangeLog

@@ -7,6 +7,7 @@ Changes from Version 1.0.3 to 1.0.4:
  * Corrected a bug in approximation algorithm which caused the approximation to fail for seemingly 
    random models.
  * Fixed a bug in SSMcycle which caused error with common components.
+ * Fixed bug in SSMcycle which resulted erroneus system matrix T in all cases.
  * Fixed a bug in SSMseasonal which caused error in SSModel when using common components.
  * SSMseasonal with trigonometric seasonal now works properly when period is odd.  
  * Fixed a bug in coef.KFS which caused function to return smoothed states even with argument 

R/SSMcycle.R

@@ -15,16 +15,16 @@ SSMcycle <- function(period, type, Q, index, a1, P1, P1inf, n, ynames) {
             stop("type must be 'distinct' or 'common'.")
     }
 
-    if (!(length(period) == 1 & period > 1 & abs(period - round(period)) == 0)) 
-        stop("period of the cycle component must be integer larger than 1. ")
+    if (!(length(period) == 1 & period > 0)) 
+        stop("Period of the cycle component must be larger than 0. ")
 
 
     lambda <- 2 * pi/period
     m <- 2 * ((p - 1) * (type == 1) + 1)
     Z <- matrix(0, p, m)
     T <- matrix(0, m, m)
     Z_univariate <- matrix(c(1,0), 1, 2)
-    T_univariate <- matrix(c(cos(lambda), sin(lambda), -sin(lambda), cos(lambda)), 2, 2)
+    T_univariate <- matrix(c(cos(lambda), -sin(lambda), sin(lambda), cos(lambda)), 2, 2)
     if (type != 2) {
       for (i in 1:p) {
         Z[i, ((i - 1) * 2 + 1):(i * 2)] <- Z_univariate

R/predict.SSModel.R

@@ -138,7 +138,7 @@ predict.SSModel <- function(object, newdata, n.ahead, interval = c("none", "conf
     }
 
   } else {
-    if(!is.null(n.ahead) || !missing(n.ahead)){
+    if(!missing(n.ahead) && !is.null(n.ahead)){
 
       tv  <- logical(5)
       tv[1] <- dim(object$Z)[3] > 1
@@ -203,7 +203,8 @@ predict.SSModel <- function(object, newdata, n.ahead, interval = c("none", "conf
       }
       out <- KFS(model = object, smoothing = "signal",maxiter=maxiter)
       for (i in 1:p) {
-        pred[[i]] <- cbind(fit=out$thetahat[timespan, i]+(if(object$distribution[i]=="poisson") log(object$u[timespan, i]) else 0), 
+        pred[[i]] <- cbind(fit=out$thetahat[timespan, i]+
+                             (if(object$distribution[i]=="poisson") log(object$u[timespan, i]) else 0), 
                            switch(interval, none = NULL, out$thetahat[timespan, i] +
                                     (if(object$distribution[i]=="poisson") log(object$u[timespan, i]) else 0) +
                                     qnorm((1 - level)/2) * sqrt(out$V_theta[i, i, timespan]) %o% c(1, -1)),