SARIMA(0,0,0)(4,0,4)[12]

set.seed(123)

n <- 120 + 60 # panjang data + burn-in data <- rnorm(n, mean = 0, sd = 1)

Seasonal MA coefficients (12, 24, 36, 48 lags)

theta <- c(0.5, -0.3, 0.4, -0.2) season <- 12 yt <- numeric(n)

for (t in (season4+1):n) { yt[t] <- data[t] + theta[1]data[t-season] + theta[2]data[t-2*season] + theta[3]data[t-3*season] + theta[4]*data[t-4*season] }

Hilangkan nilai awal

yt <- yt[-(1:(season*4))] yt_ts <- ts(yt, frequency = 12)

Plot Time Series

plot.ts(yt_ts, main=“Original Plot Time Series”, col=“black”)

#ACF & PACF Original Series par(mfrow=c(1,2)) acf(yt_ts, main=“ACF-No Differencing”) pacf(yt_ts, main=“PACF-No Differencingl”)

Seasonal Difference (lag = 12)

seasonal <- diff(yt_ts, lag=12)

ACF & PACF after Seasonal Differencing

par(mfrow=c(1,2)) acf(seasonal, main=“ACF - Seasonal Difference (lag=12)”) pacf(seasonal, main=“PACF - Seasonal Difference (lag=12)”)