3D4U, Ltd., a small manufacturing firm specializing in 3D printers for individual customers, has been gathering quarterly data on orders. Their data gathering efforts started in the third quarter of 2015 (year 1). The table below represents their orders since then: Quarter Year 1 2 1 2 5415 3 2559 5106 2034 4875 3591 4 3104 3592 5130 4288 3384 3222 2478 5130 3918 3 2090 4 5161 2375 5 The officers of the company - Arnie, Burt, Candace, Denise, and Ernie - cannot agree on a method for forecasting future orders. Each of them prefers to use a different method. . . Arnie wants to use linear regression with the time period as the independent variable; Burt believes that a trend x seasonality forecasting method is needed since the data is quarterly;/ Candace feels that exponential smoothing will work just fine; Denise thinks that a time series of the form ao + ait + az t2 is the way to go; and Ernie says "let's keep it simple and just use a horizontal line." . However, they all agree that the firm's eventual forecasting method needs to focus on minimizing the mean squared error (MSE). Based on the available data (where t = 1 corresponds to Q3 of Year 1), your job is to a. determine the best version of each of these proposed methods, and b. rank the methods using the agreed upon performance metric of minimizing MSE. Each officer wants to understand the analysis that is performed, so it is very important that your analysis is explained well and documented thoroughly. In addition, you must discuss the appropriateness of each forecasting method. 3D4U, Ltd., a small manufacturing firm specializing in 3D printers for individual customers, has been gathering quarterly data on orders. Their data gathering efforts started in the third quarter of 2015 (year 1). The table below represents their orders since then: Quarter Year 1 2 1 2 5415 3 2559 5106 2034 4875 3591 4 3104 3592 5130 4288 3384 3222 2478 5130 3918 3 2090 4 5161 2375 5 The officers of the company - Arnie, Burt, Candace, Denise, and Ernie - cannot agree on a method for forecasting future orders. Each of them prefers to use a different method. . . Arnie wants to use linear regression with the time period as the independent variable; Burt believes that a trend x seasonality forecasting method is needed since the data is quarterly;/ Candace feels that exponential smoothing will work just fine; Denise thinks that a time series of the form ao + ait + az t2 is the way to go; and Ernie says "let's keep it simple and just use a horizontal line." . However, they all agree that the firm's eventual forecasting method needs to focus on minimizing the mean squared error (MSE). Based on the available data (where t = 1 corresponds to Q3 of Year 1), your job is to a. determine the best version of each of these proposed methods, and b. rank the methods using the agreed upon performance metric of minimizing MSE. Each officer wants to understand the analysis that is performed, so it is very important that your analysis is explained well and documented thoroughly. In addition, you must discuss the appropriateness of each forecasting method
