Old MathJax webview

Use the time series data you have collected as part of your previous assignment to address the following questions. If the time series data you have utilized is not long enough for the purposes of this assignment, you may either make an attempt to extend it with new observations or decide to use a different data set. 1. Use Minitab or another statistical software of your choice to draw (a) a simple time plot and (b) a correlogram of your time series data. (c) Use these graphs to discuss the time series behavior of your data, specifically focusing on the existence/absence of a linearon-linear trend & seasonalon-seasonal cycles. 2. Fit a simple nave model, a nave model with a trend, and a seasonal nave model to your data and obtain one-period-ahead forecasts and forecast errors for each model. For each model, (a) plot a normal quantile- to-quantile plot of the forecast errors and (b) test the normality and independence of these error terms by applying the Jarque-Bera and Ljung-Box tests, respectively. (c) Briefly elaborate on your findings. 3. Fit the single and double moving average models and the single exponential smoothing model to your data and obtain one-period-ahead forecasts and forecast errors for each model. For each model, (a) plot a normal quantile-to-quantile plot of the forecast errors and (b) test the normality and independence of these error terms by applying the Jarque-Bera and Ljung-Box tests, respectively. (c) Briefly interpret your results. 4. For each of the six models above, (a) compute the mean error, mean absolute error, mean absolute percentage error, and root mean squared error and (b) prepare a table that compares all of the six models you have applied in terms of their statistical accuracy and adequacy. (c) Finalize your report by making a recommendation on what model is best in terms of its in-sample accuracy and adequacy. Use the time series data you have collected as part of your previous assignment to address the following questions. If the time series data you have utilized is not long enough for the purposes of this assignment, you may either make an attempt to extend it with new observations or decide to use a different data set. 1. Use Minitab or another statistical software of your choice to draw (a) a simple time plot and (b) a correlogram of your time series data. (c) Use these graphs to discuss the time series behavior of your data, specifically focusing on the existence/absence of a linearon-linear trend & seasonalon-seasonal cycles. 2. Fit a simple nave model, a nave model with a trend, and a seasonal nave model to your data and obtain one-period-ahead forecasts and forecast errors for each model. For each model, (a) plot a normal quantile- to-quantile plot of the forecast errors and (b) test the normality and independence of these error terms by applying the Jarque-Bera and Ljung-Box tests, respectively. (c) Briefly elaborate on your findings. 3. Fit the single and double moving average models and the single exponential smoothing model to your data and obtain one-period-ahead forecasts and forecast errors for each model. For each model, (a) plot a normal quantile-to-quantile plot of the forecast errors and (b) test the normality and independence of these error terms by applying the Jarque-Bera and Ljung-Box tests, respectively. (c) Briefly interpret your results. 4. For each of the six models above, (a) compute the mean error, mean absolute error, mean absolute percentage error, and root mean squared error and (b) prepare a table that compares all of the six models you have applied in terms of their statistical accuracy and adequacy. (c) Finalize your report by making a recommendation on what model is best in terms of its in-sample accuracy and adequacy