What are time series forecasting methods?

Answer 1

Time series forecasting methods traditionally look only at the variable to be forecast, using no other explanatory variables. What you looking for is patterns in the data that can be extrapolated into the future in order to forecast future value. You are looking for such things as trends in the data (upward or downward ) as well as seasonal tends ( quarterly, monthly, weekly etc. )

The definitive treatment of time series is a book by Box and Jenkins.. The authors utilize a systematic approach to forming models by examining the various auto-correlations of the variable lagged by various time periods. After you fit a model you examine the error terms and look for any remaining patterns and continue with new models until the residual cease to have any pattern.

Feel free to contact me for more information.