mcp-package: Regression with Multiple Change Points

In mcp: Regression with Multiple Change Points

Regression with Multiple Change Points

Description

The mcp package provides an interface to fit regression models with multiple change points between generalized linear segments, optionally with per-segment variance and autocorrelation structures.

The main function of mcp is the mcp() function, which uses a formula syntax to specify a wide range of change point models. Based on the supplied data, formulas, and additional information, it writes JAGS code on the fly and use rstan to fit the model, optionally in parallel to speed up sampling. You will need to install JAGS for mcp() to work.

A large number of post-processing methods can be applied. These include

• Summarise fits using summary(), fixef(), and ranef().

• Visualize fits using plot() and individual parameters using plot_pars().

• Test hypotheses using hypothesis() and loo().

Extensive documentation with worked examples is available at the mcp website.

mcp documentation built on April 1, 2023, 12:03 a.m.