For this assignment, our teacher gave us data for 144 periods and asks us...
"use the best forecasting technique for forecast 12 periods into the future. State all relevant assumptions, and briefly describe the technique(s) you used."
Now, I'm assuming this would be a seasonal forecast, but I'm not sure what technique to use.
We have only covered the following topics so far:
1) Time series analysis: SimpleMovingAverage, Weightedmovingaverage, Exponential Smoothing
3) Forecast errors: absolute deviation, mean absolute deviation (MAD), and tracking signal
4) Single linear regression
5) Seasonalized time series regression; and nonlinear seasonal forecasting
Not sure, but 5 seems likely...I do not know how to approach this....and what assumptions are there to be drawn...can you help?
With 144 periods of data, you have enough of a series to establish the behavior of the data. What you're most interested in is are there any patterns in the data. To be more specific, is an upward or downward trend apparent? Do the data show a cyclical or time series variation? Are the data points (when plotted) linear or curvilinear?
<br>The way an experienced statistician gets a first cut at these questions is to plot the data. Microsoft Excel can do that. What you want is called a scatter ...