Basic Time Series Forecasting

Basic time series forecasting methods:

References

Hyndman, R. J., and Athanasopoulos, G. (2018) Exponential smoothing. Forecasting: principals and practice, 2nd Ed.
https://otexts.com/fpp2/expsmooth.html

Ravinder, H. V. (2016) Determining the optimal values of exponential smoothing constants – does Solver really work? American Journal of Business Education
https://files.eric.ed.gov/fulltext/EJ1054363.pdf

6 thoughts on “Basic Time Series Forecasting”

    • There are lots of ways of measuring trend and seasonality. E.g. referring to Holt-Winters (see the following webpage), I guess you can use beta as a measure of trend and gamma as a measure of seasonality. The slope of the line that best fits the data could be viewed as a measure of trend and the period could be viewed as a measure of seasonality (quarterly = 4, monthly = 12, etc.).
      https://real-statistics.com/time-series-analysis/basic-time-series-forecasting/holt-winters-method/
      Charles

      Reply
      • Thank you for your kind reply. In the above link, it seems that s represents seasonality. Is gamma also a measure of seasonality? By the way, s doesn’t measure how strong the seasonality is.

        Do we have to measure trend and seasonality depending on the model? Isn’t there a universal method which only uses the given time series?

        Reply
        • s represents the seasonality period, while gamma can be viewed as a measure of seasonality strength. There is also something called the seasonality index, which tells how far above or below the mean any season is. E.g. in a case where you have 4 seasons (quarters) with data (e.g. sales data) for each quarter, the seasonality index = the data value for that quarter divided by the average value for the year. This is shown in the table below. You can clearly see that Q4 sales are 1.5 times the yearly average, while Q3 sales are only 60% of the yearly average.

          Quarter Data Index
          1 23 0.766666667
          2 34 1.133333333
          3 18 0.6
          4 45 1.5
          mean 30

          Reply

Leave a Comment