R/tvGarchKalmanLoglike.R

In tvGarchKF: Time-Varying Garch Models Through a State-Space Representation

#' @title Models tv-Garch Filter Kalman LogLikehood.
#' @description It is the function to use in the process to fit coefficients.
#' @param x Vector of coefficents to fit.
#' @param series Time series.
#' @param c Vector containing coefficents of c.
#' @param alpha Vector containing coefficents of alpha.
#' @param beta Vector containing coefficents of beta.
#' @param nsample Value of time series length.
#' @param type Vector of function type for c, alpha and beta.
#' @param exponentes Vector for exponenets in NoLineal.
#' @param trig Type of trigonometric function.
#' @param arg Value of argument for the trigonometric function.
#' @param predict Value for time to generate predict.
#' @param trace.log Variable to print names of coefficients.
#' @return Value of loglike in model.
#' @importFrom stats na.exclude
#' @export
tvGarchKalmanLoglike <- function(x, series, c, alpha, beta, nsample=length(series), type=c("polynomial","NoLineal","trigonometric"), exponentes, trig, arg, predict, trace.log=FALSE){

  pred<-rep(NA,predict)
  series<-c(series,pred)

  nsample <- length(series)
  u <- (1:nsample)/nsample

  aux1 <- length(c)
  aux2 <- length(alpha)
  aux3 <- length(beta)
  if (aux1<=1){
    c = x[1]
  }else{
    c = c()
    for(i in 1:aux1){
      c[i] = x[i]
    }
  }
  if (aux2<=1){
    alpha = x[1+aux1]
  }else{
    alpha = c()
    for(i in 1:aux2){
      alpha[i] = x[i+aux1]
    }
  }
  if (aux3<=1){
    beta = x[1+aux1+aux2]
  }else{
    beta = c()
    for(i in 1:aux3){
      beta[i] = x[i+aux2+aux1]
    }
  }

  auxA <- list(c=c,alpha=alpha,beta=beta)
  auxB <- list()
  for(i in 1:length(type)){
    if(type[i] == "polynomial"){
      auxC <- polynomial(nsample, coeficientes = unlist(auxA[i]))
    }
    if(type[i] == "NoLineal"){
      auxC <- NoLineal(nsample, coeficientes = unlist(auxA[i]), exponentes = exponentes)
    }
    if(type[i] == "trigonometric"){
      auxC <- trigonometric(nsample, coeficientes = unlist(auxA[i]), trig = trig, arg = arg)
    }
    auxC <- list(auxC)
    auxB <- c(auxB,auxC)
  }

  c <- unlist(auxB[1])
  a <- unlist(auxB[2])
  b <- unlist(auxB[3])

  for(i in 1:nsample){
    if(any(is.nan(a[i])) || any(is.nan(b[i])) || any(is.nan(c[i]))) {
      stop("Se produjo un error al estimar los parametros")
    }
  }

  for(i in 1:nsample){
    if(trace.log) {
      cat("a[i]:",a[i],"|b[i]:",b[i],"|c:",c,"\n")
    }
  }

  for(i in 1:nsample){
    if ((c[i]<=0) & (a[i]<=0) & (b[i]<=0) & ((b[i]+a[i])>=1) ) {
      return(100000000.)
    }
  }

  for (i in 1:nsample) {
    if ((b[i]+a[i])>=1){
      return(100000000.)
    }
  }

  y <- NULL

  for(i in 1:nsample){
    y[i] <- ifelse (is.na(series[i]), 100000000., series[i]^2-c[i]/(1-a[i]-b[i]))
  }

  P <-rep(0, nsample) # Varianza del error de estimaci?n
  X <-rep(0, nsample)   # Representa el estado (alfa)
  F.m <-rep(0, nsample)
  K <-rep(0, nsample)

  # Inicializo P_1
  P[1]<-(a[1]^2)/(1-(a[1]+b[1])^2)

  if (aux1 > 1){
    resultado <- .C("tvGarch1ab2", as.integer(nsample), as.double(a), as.double(b), as.double(c), as.double(P), as.double(X), as.double(F.m), as.double(K), as.double(y))
  }else{
    resultado <- .C("tvGarch1ab", as.integer(nsample), as.double(a), as.double(b), as.double(c), as.double(P), as.double(X), as.double(F.m), as.double(K), as.double(y))
  }

  X=resultado[[6]]
  F.m=resultado[[7]]
  K=resultado[[8]]
  ####
  e.sig2=X+(c/(1-a-b))


  if( min(e.sig2) <=0 | min(F.m) <=0 ){loglik=100000000}
  else{
    loglik <- sum(na.exclude((sqrt(F.m)*series^2) / e.sig2 + log(e.sig2) - 0.5*log(F.m) ))
    attr(loglik,"sigma")<-e.sig2
    loglik}
}