changepoint.forecast: changepoint.forecast: An R package aimed at monitoring...

In grundy95/changepoint.forecast: Sequential Changepoint Detection of Forecast Errors

Description

The changepoint.forecast package monitors provided forecast errors for changepoints in first and second order changes. These include the common mean and variance changes. The main function is cptForecast which looks for changepoints in a set of forecast errors and then the function updateForecast can be used to analyse new forecast errors from the same forecasting model without repeating the initial analysis. This allows for the online implementation of the method as new forecast errors become available.

Main functions

cptForecast: Perform sequential changepoint analysis on forecast errors.

updateForecast: Update sequential changepoint analysis on new forecast errors.

cptSeqCUSUM: Perform generic CUSUM based sequential changepoint analysis on data.

cptFor S4 class

The main functions return a cptFor object by default. This S4 class object can then be passed to updateForecast along with new forecast errors to update the analysis. The methods show, summary and plot can be used on the cptFor object to see the main results of the analysis. Furthermore the slots of cptFor can be accessed using the @ (this works similarily to the $ for S3 objects). Moreover, a set of retriever functions are available for example to access the errors slots of an object ans, you would call errors(ans).

References

\insertRef

Grundy2021changepoint.forecast

\insertRef

Fremdt2014changepoint.forecast

grundy95/changepoint.forecast documentation built on Dec. 20, 2021, 1:45 p.m.