Description
## Exercise set.seed(54)
myts <- ts(c(rnorm(50, 34, 10),
rnorm(67, 7, 1),
runif(23, 3, 14)))
#5. Plot the data, explain the statistical characters of the data
#6. Use 80% of the data as the training set and the rest as testing set - This is to make sure the forecast models
#do not carry any information of the testing set (the rest 20% of the data) reserved for accuracy analysis.
#7.Set up three forecasting models using the training set.
#8.Get a plot with the three forecast models, add a legend. Which method looks more promising?
#9.Perform accuracy analysis to get the error measures and compare them; do the results match the
#visual impression (plot the residual)? if not, why?
#10. Check relevant statistical traits: Mean of zero; equal variance;
#standard distribution of the residual.
