Explore BrainMass

Explore BrainMass

    Forecasting and Seasonality

    This content was COPIED from BrainMass.com - View the original, and get the already-completed solution here!

    I need to forecast out 12 months using 3 different techniques, and then decide which one is best based on MAD (mean absolute deviation).

    Using some historical data for the past 40 periods (see attachment) I need to make a forecast for the next 12 periods (periods 41 thru 52) using 3 of the simple forecasting techniques we have covered so far in our 2nd year class. These include: 1)Simple Moving Average 2)Weighted Moving Average 3)Exponential Smoothing 4)Linear Regression 5)Seasonal Forecasting via Regression, if we determine it is subject to seasonality (i.e., linear trends with multiplicative seasonality or nonlinear trends with multiplicative seasonality, and linear trends with additive seasonality or nonlinear trends with additive seasonality)

    Using Excel or the Excel add-in Crystal Ball (we may NOT use Minitab, SAS, JMP, or any other packages), we need to investigate AT LEAST 3 different ways to forecast this data, and present the MAD (mean absolute deviation) and parameter values of each technique we tried. And then pick whatever is the "best forecast" based on the MAD...

    I do not know how to do this. I think seasonality may be one technique we need to look at, because the prof mentioned it would be one to look out for...but how do we know if demand is affected seasonally???

    © BrainMass Inc. brainmass.com June 3, 2020, 5:06 pm ad1c9bdddf


    Solution Preview

    Using Excel, this assignment should be relatively straight-forward. I would suggest trying each of the models you've learned (if for no other reason then to get practice with each). Start with the easiest, simple moving average. If you wanted a 3-period moving average, for example, you'd start forecasting at period 4: find the average of the actual demand from the previous 3 quarters. (Set up the equation in the square next to the "Demand" for pd 4, then copy & paste all the way down the line to 1 period after the data ends.) In the next column, find the error for each period. Put the absolute value of each error in the next column and use that to determine the MAD value. You could use any # of periods to do the forecasting that you want, but I suggest using whatever your prof used as an example in class.
    <br>Then use the equations for the weighted avg and exponential smoothing models. You would need to make an assumption about the ...