Question

Differentiate between an additive model and a multiplicative model when accounting for seasonal data in a casual forecasting model.

Answer 1

Additive model and multiplicative model are different in the way the data is calculated using the components.

In additive models the assumption is that the different components affect the time series additively.

So data in additive model will be something like this:

Data=Seasonal effect + Trend + Cyclical + Residual.

It is used where change is measured in absolute quantity.

In multiplicative model the assumption is the components act as the multiplication result. So the data will be like:

Data= Seasonal effect* Trend *Cyclical*Residual.

Used in cases where the change is a percent change depending on the components.

Usually for fitting multiplicative models logarithm is applied on the data to change multiplication to addition.

log(Data)=log(Seasonal effect)+log(Trend)+log(Cyclical)+log(Residual)

Happy learning.