Use Word for writing, and Excel for formulas and graphics. Conduct a time-series forecasting on a key variable that is deemed important for an organization using the three methodologies below:
– Decomposition,
– Holt-Winters’ exponential smoothing,
– Regression for time-series forecasting.
Data may be directly collected from one or more organizations of interest (such as freight forwarders and shipping companies) or from data websites.
But in any case, make sure that your data exhibits both trend and seasonality. Apply each model on the data you have collected with parameters set to appropriate values (use optimization if necessary). Then forecast m periods ahead, where m is the seasonality period in your data. Compare the accuracy of three methods using the appropriate statistics, and report on the pros and cons of each method.
