The Dynamic Harmonic Regression model in R requires the input of parameters K, the length of which depends on the number of seasonality in the forecast data.

According to https://otexts.com/fpp2/dhr.html K is used to control the smoothness of the forecast model. The higher the magnitude of the parameters, the rougher the forecast graph will be.

In Prof. Hyndman's lecture on Datacamp, https://campus.datacamp.com/courses/forecasting-using-r/advanced-methods?ex=4

m refers to the number of seasonal periods. How does m differ from K? What does exactly K depend on?
I have two electricy consumption datasets spanning for 6 months, with minute by minute granularity, from two different households. Both the data sets have weekly and the daily seasonality, I get K as c(17,2) for one of the dataset and K as c(15,5) for the other datasets. How can I use these K values to infer about the nature of the data sets? Why is it different for both datasets? Do these parameters 'K' need to be tuned from time to time for the model or do they stay constant throughout the year? is there a rock solid definition for 'K' parameters used in DHR ? and what do they exactly depend on?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.