The actual data set is acquired from the ABS website (http:www.abs.gov.au). This data is a migration data set - Oversea Arrivals and Departures, Australia Section, Table 6: Short-Term Movement, Resident Departures-Intended Length Stay and Main Reason for Business. The data starts from January 1991 to June 2003. The data set is shown in the graph below.

The last way to test the forecasting model is visual inspection. This technique is to look at the graph and find minimum differences between the actual data and the forecasting data. This technique is seemed easy to conduct but sometimes it is hard to justify which model is really the best one. This is because some forecasting models have minimum differences which visualization cannot guarantee a good result. With all three ways to test the forecasting models, it expects to have a best result for finding the best fit model.

The recommendation for this report is that the 4 quarter moving average is the best model to forecast the data set. However, it needs to monitor the trend of data together with environmental factors. As it can be seen in the actual graph, the resident departures for business had increased after the irregular events. This shows that people have a confident for going out of the country for business. Therefore the situation would be back to normal as it was before. Thus, the 4 quarter moving average is the best model at the present time. However, it cannot trust this model in 100 % because this model is only expected what is going to be happen in the future. Therefore, monitoring the situation and trend would be good to response and apply other forecasting models which might suitable for each period of time.

Once the deseasonalising data has been found, it is able to develop some forecasting models in term to find the best fit model.

With the results of Seasonal Indices, it can be develop of the Deseasonalising data. The findings of deseasonalising data series will show in a smoother data compared to the actual data. The table of Deseasonalising the data set is in the Appendix C. After collecting all deseasonalising data, it can be plot in the graph for comparing the actual data and the deseasonalising data. The graph is shown as below.

According to the graph, it can be observed clearly that the deseasonalised data is much smoother than the original series. It shows that the trend of deseasonalised data appears to be stationary, which means it does not appear to be an increasing or decreasing trend evident.

Some topics in this essay:
Forecasting Models, CONCLUSION RECOMMENDATIONS, War SARS, Reason Business, Seasonal Indices, Indices January, SI Appendix, moving average, quarter moving, quarter moving average, , data set, February March, average model, moving average model, Plot Data, resident departures, 4 quarter moving, forecasting models, 4 quarter, actual data, fit model, deseasonalising data, resident departures business, 6 quarter moving,