Fits or generalized linear models either a regression with Autoregressive moving-average (ARMA) errors for time series data. The package makes it easy to incorporate constraints into the model's coefficients. The model is specified by an objective function (Gaussian, Binomial or Poisson) or an ARMA order (p,q), a vector of bound constraints for the coefficients (i.e beta1 > 0) and the possibility to incorporate restrictions among coefficients (i.e beta1 > beta2). The references of this packages are the same as 'stats' package for glm() and arima() functions. See Brockwell, P. J. and Davis, R. A. (1996, ISBN-10: 9783319298528). For the different optimizers implemented, it is recommended to consult the documentation of the corresponding packages.
ConsReg is my first R package that I decide to give to the R community. It is a compilation of functions that I have used in my day-to-day developing models.
ConsReg allows the estimation of GLM models with restrictions to parameters: both of limits (upper and lower) and between parameters (e.g. coef1 > coef2). This package is very easy to use and is ideal for situations where the logic of the model is more important than the error itself.
ConsReg has two main functions:
The first one, *ConsReg, as I said is a wrapper of glm function, while ConsRegArima allows to estimate regression models with Arima errors.