Changes in version 0.3.5
  o lvnet is now no longer supported, with most functionality moved to the psychonetrics package. A message now warns users of this on loading the package.

Changes in Version 0.3.3
  o Added 'refitAll' argument to lvnetLasso to refit all models in the LASSO path before evaluating EBIC. This leads to similar functionality as qgraph::ggmModSelect(.., stepwise = FALSE), although the performance is slower.
  o The argument 'refit' has been renamed 'refitFinal'
  o Removed a warning on depracated dplyr functions
  o Fixed a bug in counting the number of parameters in RNM LASSO estimation

Changes in Version 0.3.3
  o 'ggmFit' is no longer exported by lvnet

Changes in Version 0.3.2
  o Summary output now contains parameter labels
  o Equality constraints now supported by supplying labels in model matrices

Changes in Version 0.3.1:
	o 'ggmFit' is now moved to qgraph and exported by lvnet
	    o Several updates to ggmFit in qgraph
	o Added 'lvnetRefit' to refit an lvnet model to new data

Changes in Version 0.3:
	o 'scale' now defaults to FALSE!
	  o Due to this change, some values such as the BIC will be different
	o Fixed a bug where higher order latent variables were not usuable
	o Added the 'mimic' argument. mimic = "lavaan" will mimic lavaan results (now default).
	  o lvnet should now return the same fit as lavaan when std.lv=TRUE is used in lavaan.
	o Added the function 'lav2lvnet' to easily generate lvnet input from a lavaan model
	o Added 'ggmFit' to fit a given partial correlation network (Gaussian graphical model; GGM)