Multiple regression and Time Series

[sony]

Hi,

I'm in a college statistics course where I'm doing an assignement with Minitab.

I have a one month time series of electricity use (hour intervals).

I've attempted to remove the season effect from weekdays by index multiplication so all I'm left with (hopefully) is the effect from day/night, possible trend, og random variation.

By facing this with multiple regression I've created dummy variables for every hour in the day, and observation number as dependant variables.

So my question here is: In minitab you can choose to "FIT INTERCEPT" in regression options,

exactly what qualitative implications does this have for my model? IF i force it to fit intercept (introduce a CONSTANT to the regression equation) do i then "create" a deterministic trend??


Thank you!

 

[lippyka]

When you choose to "Fit Intercept" in regression options, this means that you are including a constant term in the regression equation. This can help account for any underlying trend that is present in your data and can also be used to adjust the intercept of the regression line so that it better fits the data. The constant term can also help to reduce the error in your prediction by accounting for any residual variance that is not explained by the other terms included in the model. By using the constant term, you are essentially creating a deterministic trend in your model.