QUESTION ONE

This question refers to the data set that you are using for your project. For your data set determine the best forecasting method (average, naïve, seasonal naïve, exponential smoothing, Holt-Winters’ method, ETS, ARIMA). ). Use 80% of the data for training and 20% of the data for testing. Using the best model, forecast your data set for 6 periods into the future.

Load Data and Library Packages
Define as time-series and name varaible
indigo_data = ts(indigo_data, start=2002, frequency=4)
data = indigo_data[,3]
Set training data, test data, out of sample
train <- window(data,start=c(2002, 1),end=c(2014, 2))
test <- window(data, start=c(2014,2),end=c(2017, 3))
both <- window(data,start=c(2002, 1))
h=length(test)
Forecast using Average, Naïve, and Seasonal Naïve Method
Indigofit1 <- meanf(train, h=h)
Indigofit2 <- naive(train, h=h)
Indigofit3 <- snaive(train, h=h)
Plot forecasts for Average, Naïve, and Seasonal Naïve Method
plot(Indigofit1, PI=FALSE,
     main="Forecasts for quarterly Indigo sales")
lines(Indigofit2$mean,col=2)
lines(Indigofit3$mean,col=3)
legend("topleft",lty=1,col=c(4,2,3),
       legend=c("Mean method","Naive method","Seasonal naive method"),bty="n")

Forecast using Simple Exonential moving averages
Indigofit4 <- ses(train, h = h)
plot(Indigofit4)