seqest (version 1.0.1)

seq_GEE_model: The The sequential method for generalized estimating equations case.

Description

seq_GEE_model estimates the the effective variables and chooses the subjects sequentially by the generalized estimating equations with adaptive shrinkage estimate method.

Usage

seq_GEE_model(formula, data = list(), clusterID, data_pool = list(),
  clusterID_pool, strategy, d = 0.4, family = stats::gaussian(link =
  "identity"), corstr = "independence", contrasts = NULL, ...)

Arguments

formula

An object of class "formula" (or one that can be coerced to that class): a symbolic description of the model to be fitted.

data

A data frame containing the initial random samples to obtain the initial estimate of the coefficient. Note that the first column of the data frame is the response variable, and the rest is the explanatory variables.

clusterID

The id for each subject in the initial samples. Note that the subjects in the same cluster will have identical id.

data_pool

A data frame containing all the random samples which we will choose subject from. The first column of the data frame is the response variable, and the rest is the explanatory variables.

clusterID_pool

The id for each subject in the data_pool. Note that the subjects in the same cluster will have identical id.

strategy

A character string that determines the sample selection criterion to be used, matching one of 'random' or 'D_optimal. The default value is 'D_optimal'.

d

A numeric number specifying the length of the fixed size confidence set for our model. The default value is 0.4.

family

A description of the error distribution and link function to be used in the model. See family for details of family functions. Matching one of 'gaussian' or 'binomial'

corstr

A character string specifying the correlation structure. The following are permitted: "independence", "exchangeable" and "ar1".

contrasts

An optional list. See the contrasts.arg of model.matrix.default.

...

Further arguments passed to or from other methods.

Value

a list containing the following components

d

the length of the fixed size confidence set that we specify

n

the current sample size when the stopping criterion is satisfied

is_stopped

the label of sequential iterations stop or not. When the value of is_stopped is TRUE, it means the iteration stops

beta_est

the parameters that we estimate when the the iteration is finished

cov

the covariance matrix between the estimated parameters

rho

estimate of correlation coefficient

nonZeroIdx

the index of the non zero coefficient

corstr

the correlation structure. The following are permitted: "independence", "exchangeable" and "ar1"

family

a description of the error distribution and link function to be used in the model

Details

seq_GEE_model fits the clustered data sequentially by generalized estimating equations with adaptive shrinkage estimate. It can detect the effective variables which have the impact on the response and choose the most representative sample point at the same time. Specifically, we fit a initial sample data and determine if the stop condition is reached. If not, we will select the most informative subjects by some criterion. Iteration stops once it meets our requirements.

References

Chen, Z., Wang, Z., & Chang, Y. I. (2019). Sequential adaptive variables and subject selection for GEE methods. Biometrics. doi:10.1111/biom.13160

See Also

seq_cat_model for categorical case

seq_bin_model for binary classification case

seq_ord_model for ordinal case.

Examples

# NOT RUN {
# generate the toy example
data <- gen_GEE_data(numClusters = 75, clusterSize = 5,
                    clusterCorstr = 'ar1', clusterRho = 0.3,
                    beta = c(1, -1.1, 1.5, -2, rep(0, 50)), family = gaussian(),
                    intercept = TRUE, xCorstr = 'ar1',
                    xCorRho = 0.5, xVariance = 0.2)
df <- data.frame(y = data$y, data$x)
clusterID <- data$clusterID
pool <- gen_GEE_data(numClusters = 8000, clusterSize = 5,
                     clusterCorstr = 'ar1', clusterRho = 0.3,
                     beta = c(1, -1.1, 1.5, -2, rep(0, 50)), family = gaussian(),
                     intercept = TRUE, xCorstr = 'ar1',
                     xCorRho = 0.5, xVariance = 0.2)
df_pool <- data.frame(y = pool$y, pool$x)
clusterID_pool <- pool$clusterID
d<- 0.25

# use seq_GEE_model to generalized estimating equations case.
# You can remove #' to run the command.
# seqRes.ASED <- seq_GEE_model(y ~ .-1, data = df, clusterID = clusterID,
#                             data_pool = df_pool,  clusterID_pool = clusterID_pool,
#                             strategy = "D-optimal",  d = d, family = gaussian(), corstr = 'ar1')
# }