aggrmodel_cluster: Fit aggregated model with clusters
In gabrielfranco89/aggrmodel: A R package to fit aggregated data model

Description Usage Arguments Value Examples

Fit aggregated model with clusters

aggrmodel_cluster(
  formula = NULL,
  data,
  market,
  Y = NULL,
  timeVar,
  groupVar,
  repVar,
  n_basis,
  n_basis_cov = NULL,
  n_cluster = NULL,
  n_trials = 42,
  basisFunction = "B-Splines",
  cicleRep = FALSE,
  n_order = 4,
  covType = "Homog_Uniform",
  corType = "exponential",
  optMethod = "L-BFGS-B",
  returnFitted = FALSE,
  optLk = TRUE,
  corPar_init = NULL,
  diffTol = 1e-06,
  itMax = 100,
  verbose = FALSE
)

`formula`	building...
`data`	Dataset containing group, replicates (if any), time and aggregated signal
`market`	Market data frame. MUST be a 3 column dataframe with the following order: Group, Type and Number of subjects
`Y`	Dependent variable: aggregated signal
`timeVar`	Name of time variable
`groupVar`	Name of grouping variable
`repVar`	Name of replicates variable
`n_basis`	Number of basis functions for basis expansion
`n_basis_cov`	Number of basis functions for variance functional expansion
`n_cluster`	Number of grouping clusters
`n_trials`	Number of random grouping trials for cluster initial values (Default: 42 and don't forget your towel!)
`basisFunction`	Character indicating which basis: 'B-Splines' (default) or 'Fourier'
`cicleRep`	Indicator TRUE/FALSE if replicates are cyclical
`n_order`	Order of basis Splines (Default: 4)
`covType`	Covariance functional type. One of "Homog_Uniform" (default), "Homog" or "Heterog"
`corType`	Correlation structure type. One of "periodic" (default) or "exponential"
`corPar_init`
`diffTol`	Tolerance of model covergence (Default: 1e-06)
`itMax`	Number of maximum iterations of EM algorithm (Default: 100)
`verbose`	TRUE/FALSE indicating if steps of optimization should be printed as messages (Default: FALSE)

An aggrmodel_cluster object

set.seed(81453)
df <- createSimuData(nRep=5)
mkt <- attr(df,"market")

fit_cl = aggrmodel_cluster(data = df, market=mkt, Y = obs, timeVar =
time,groupVar = group, repVar = rep, n_basis = 9, n_basis_cov = NULL,
n_cluster = 2,n_trials = 500, n_order = 4, corType = 'exponential',
returnFitted = TRUE, verbose=TRUE)

plot(fitCluster, scales="free")