How to improve prediction using 18-point moving average?

[Mirisage]

Hi Forum,

Could anybody help me on this?

I have a monthly data series for the past 35 months (pl see attached book1.xls).

Based on this, my boss asked me to predict the expenditure for the next 25 months.

I have used a simple 18-point moving average method to do these predictions.

Question.
As is seen in the chart (pl see attached book1.xls), there is a big jump in the initial month of predicted data, otherwise the predicted series looks not too bad.

How could I get rid of this jump (or is there a better alternative method to do the prediction with precision?)

Appreciate your help
Mirisage

[Jakobshavn]

I would create a straight line linear fit to the data. As time progresses, replace each estimated value with the real data and re-do the fit.. See:

http://spreadsheetpage.com/index.php...line_formulas/

[Mirisage]

Hi Jakoshavan,

Thank you very much for this help.

I hope to ask for more help if I find difficulties in applying these equations.

Thanks again.

Mirisage