NEWS.md

In Prof-ThiagoOliveira/lcc: Longitudinal Concordance Correlation

lcc version 1.0.4

Minor updates

• update to use inherits() function

lcc version 1.0.3

Bug fixes and minor improvements

• Added a counter for the bootstrap number (lccBootstrap line 90)

• The got (goodness of fitness) value was changed. Now we are using the CCC between fitted values from mixed effect model and original observations. This statistic makes more sense as the goodness of fit of the model have a high impact in the LCC estimates.

• Fixed issue in ciCompute function to get the correct confidence intervals for the LA statistics. Versions before 1.0.3 the CI for the LA is computed using the logit transformation and versions equal or above 1.0.3 we are using the arcsin(sqrt(x)) rather than logit function.

• Updated the summary.lcc output (cleaner and more informative)

New functions

• print.lcc: print information about the longitudinal concordance correlation.

• fitted.lcc: fitted values from object of class lcc returned by modeling functions.

• plot.lcc: diagnostic plots for conditional error and random effects from the linear mixed-effects fit are obtained.

• print.summary.lcc: information summarizing the fitted longitudinal concordance correlation is printed.

• residuals.lcc: extract the residuals from the model used to estimate the longitudinal concordance correlation function.

• vcov.lcc: returns the variance-covariance matrix of a fitted lcc model object.

• getVarCov.lcc: returns the variance-covariance matrix of a fitted lcc model object.

• AIC.lcc and BIC.lcc: calculate the Akaike's 'An Information Criterion' or the BIC or SBC (Schwarz's Bayesian criterion) for an object of class lcc.

• ranef.lcc: the estimated random effects

• coef.lcc: the fixed effects estimated and corresponding random effects estimates are obtained at subject levels.

• logLik: extract fitted log-likelihood

• anova: Compare likelihoods of fitted models from an lcc object

lcc version 1.0.2

Bug fixes and minor improvements

• Fixed issue when the "all.plots" argument is declared FALSE. Now the list of plot can be split into multiple pages and save them as pdf using ggsave function. The marrangeGrob function of package gridExtra was used to solve this problem.

• Fixed issue in lccWrapper function get the correct rho:

  • previous code: if(length(rho)==1){ return(rho[[1]]) }else(if(is.na(rho[[2]])){ return(rho[[1]]) }else(return(rho[[n.delta]])))
  • current code: if(length(rho)==1){ return(rho[[1]]) }else(if(sum(is.na(rho[[2]]))!=0){ return(rho[[1]]) }else(return(rho[[n.delta]])))

• We include the parameter 'type' in the lccPlot() function. Now the user can choice among lcc, lpc, or la as plot output.

• Fixed issue when y-axis labels is changed.

• A new parameter called interaction was included in the lcc() function. This parameter allows to estimate or not the interaction effect among predictors variables in the fixed part of the linear predictor.

• no longer needs to import package gdata

lcc version 1.0.1

New argument

• We included an argument called "interaction " in the lcc function to estimate the interaction effect between time and method. If the interaction is declared FALSE, the lcc function considers only the main effects of time and method in the model.

Bug fixes and minor improvements

• Fixed issue when the wrong name of variable is declared.

• Fixed issue when three or more levels of methods is considered in the dataset.

• Fixed issue in lccWrapper function get the correct rho:

  • previous code: if(is.na(rho[[2]])=TRUE)
  • current code: if(length(rho)==1){ return(rho[[1]]) }else(if(is.na(rho[[2]])){ return(rho[[1]]) }else(return(rho[[n.delta]])))

• lccPlot - now all plots belong to the class ggplot.

• no longer needs to import packages grid and gridExtra.

lcc version 1.0.0

• New argument numCore can be used to establish more cores to work in parallel when performing bootstrap.

Prof-ThiagoOliveira/lcc documentation built on Dec. 9, 2023, 12:10 a.m.