Discuss MAD, and explain (in general) how it can be used to assist organizations in selecting among various forecasting models

Many organizations have trouble evaluating historical data, creating various models, and determining how to select the best decision model to use. In Week 4, please discuss MAD, and explain (in general) how it can be used to assist organizations in selecting among various forecasting models. Present an example to illustrate and explain how an organization would use MAD (specifically) to select between two forecasting approaches. (Please include an Excel attachment which shows the data you used and the derivation of the MAD values that you discuss within the forum. You must discuss key findings from your model in the Excel attachment in the main body of your discussion posting.)

In all discussion question responses ensure that you correctly reference sources you used in researching and analyzing your response. Please provide at least one, appropriate scholarly citation in each week of the course. ****

No examples or answers from the following website will be considered for grading, as it contains errors -http://wiki.answers.com/Q/How_do_you_calculate_mean_absolute_deviation NOTES ON THIS ASSIGNMENT

1) The DQ assignment is asking you to use two, different forecasting methods on the same set of data and compare using MAD. You cannot use MAD to compare forecasting approach if they are not based on the SAME actual data.

2)When comparing two types of forecasting approaches, the one with the smallest/lowest MAD is the best.

3) When using weighted methods, the higher weight is always applied to the most recent data. The software you use will have a note to tell you whether to put the weights in increasing or decreasing order.

4) When developing an exponential smoothing model, one has to have a ‘starter’ or ‘seed’ value to begin the process. Typically value is typically a value from prior years, OR it is a value that was predicted before the business began.

5)If you are asked to develop a trend line, then it means you are to develop a regression model using time as the x variable and you must state the trend line in Y = b0 +b1x1 format Statistics and Probability