Time series

Hi,

I have attached a graph representing data collected on a quarterly basis from 1995-2008. The x-axis represents the time - quarterly data.

Would an additive or multiplicative time series model be appropriate?

Or is it possible for me to use a simple smoothing exponential or double smoothing exponential to do forecasting?

Any suggestion would be appreciated.

Thank you very much

Perhaps a Markov switching model or a threshold autoregressive model?

Hi, I was wondering if an additive model without seasonal variation would be appropriate?

Thanks :)