Used tslm function to fit a multiple regression model with time series data while checking residuals for auto-correlation, trend and seasonality
R libraries: fpp3, using regression as a modeling technique
- scatter plot
- The tslm function fits a regression model with time series data
- several relationships in the data
- can use tslm to fit a multiple regression model
- With time series regression it is always good to check the fit as a time series plot
- actual by predicted plot
- Check the residuals for auto-correlation
- trend and seasonality
- In tslm, trend and seasona are keywords that can be used to create a regression model
- When you have a good model, the "forecast" function can be applied to the fitted model
- Sometimes non-linear models will be more useful