BUSA3015 Business Forecasting
Regression model is created by following those step
Step 1 (Dummy variables):
Dummy variable is only applied for seasonality data. From the graph above, the data shows that it has seasonality and trend characteristics. There fore, we need dummy variables. Since this is a quarter data (March, June, September, and December), we are going to need 3 dummy variables for March (D1), June (D2), and September (D4). Thereis no dummy variable for December, which means that December would be base dummyvariable for this model. After that, we label three columns is D1, D2, and D3 respectively.
Step 2 (Time variable):
label a column that is next to dummy variable for September column “time”, the data starts from March 2016 to December 2019. Therefore, the base time milestone is March 2016, which means time variable for March 2016 is 1. Then, time variable for June 2016 is 2, for September 2016 is 3, keep this going on until December 2019 (time variable is16).
Set suitable dummy variable for each date. For example, if date is March 2016, then dummy variable for March (D1) will equal to 1 while other dummy variables are 0. The same thing goes on until December (D1 = D2 = D3 = 0)
After that I use data analysis tool to create the model. Household final expenditure is Y input while time and dummy variables are X input.
The household appliances expenditure is quarterly data, there will be 3 dummy variables, D1: dummy variable for March, D2: dummy variables for June, D3: dummy variables for September. December isexcluded from dummy variables, which means December is base or reference . The o verall equation is: