The Cardinal Electronics Company must project the sales of cellular phones for the next year (called year 2). The number of cell phones sold by Cardinal Electronics in each of the past 6 months (July of year 1 through December of year 1) is shown below. Month Number of Cellular Phones Sold July 183 August 215 September 294 October 333 November 410 December 438 a. Forecast the number of cellular phones demanded in January of year 2 using a rolling (i.e. updated at end of each month when new demand data becomes available) 4-month linear regression model. For example, forecasts generated at the end of October are made by fitting a linear regression to the demands in months July through October and forecasts generated at the end of December are made by fitting a linear regression to the demands in months September through December. Forecast the number of cellular phones demanded in July of year 2. Calculate the MAD, MSE, and MAPE for a rolling 4-month linear regression model. b. Using Double Exponential Smoothing (DES) with ? = 0.15 and ?= 0.10 (i.e. assume the data has a linear trend pattern), forecast the number of cellular phones demanded in January of year 2. Forecast the number of cellular phones demanded in July of year 2. Calculate the MAD, MSE, and MAPE for your double exponential smoothing model. c. Based on the measures of forecast accuracy you calculated for the part a) linear regression model and the part b) DES (0.15, 0.10) model, which model is a better fit for this data?