Hi John, i hope you are well, I was watching f2 time series and regression analysis lectures, I had a doubt.
After finding Trend/centered avg. and then finding Seasonal variations in additive model. and then we find avg. seasonal variations.
now my doubt is, If i was asked to forecast future values with seasonal variations. should i use the avg. seasonal variations OR i make sure those additive model averages add up to ZERO and if they don't ,i use the difference to adjust in avg. seasonal variations and then use the adjusted figures in forecating.
i mean same thing will be applicable for multiplcative model except it should addup to 100*no. of seasons/ period
Ask the Tutor ACCA PM
Time series
4y ago#1With the additive model you should calculate the average seasonal variation for each season and then adjust them so that they add up to zero.
With the multiplicative model you should calculate the average seasonal % and then adjust so that they add up to 100 x number of seasons %.
Sign into reply to this topic.