IHW (version 1.0.2)

ihwResult-class: An S4 class to represent the ihw output.

Description

An S4 class to represent the ihw output.

Usage

adj_pvalues(object)
"adj_pvalues"(object)
"weights"(object, levels_only = FALSE)
thresholds(object, ...)
"thresholds"(object, levels_only = FALSE)
pvalues(object)
"pvalues"(object)
weighted_pvalues(object)
"weighted_pvalues"(object)
covariates(object)
"covariates"(object)
covariate_type(object)
"covariate_type"(object)
groups_factor(object)
"groups_factor"(object)
nfolds(object)
"nfolds"(object)
nbins(object)
"nbins"(object)
alpha(object)
"alpha"(object)
rejections(object, ...)
"rejections"(object)
rejected_hypotheses(object, ...)
"rejected_hypotheses"(object)
regularization_term(object)
"regularization_term"(object)
m_groups(object)
"m_groups"(object)
as.data.frame_ihwResult(x, row.names = NULL, optional = FALSE, ...)
"as.data.frame"(x, row.names = NULL, optional = FALSE, ...)
"nrow"(x)
"show"(object)

Arguments

object, x

A ihwResult object as returned by a call to ihw(...)

levels_only

Logical, if FALSE, return a vector of weights (thresholds) with one weight (threshold) for each hypothesis, otherwise return a nfolds x nbins matrix of weights (thresholds)

...

Parameters passed in to individual methods

row.names, optional

See ?base::as.data.frame for a description of these arguments.

Value

The different methods applied to an ihwResult object can return the following:1) A vector of length equal to the number of hypotheses tested (e.g. the adjusted p-value or the weight of each hypothesis).2) A matrix of dimension equal to nfolds x nbins (e.g. the weight of each stratum, fold combination, set by specifying levels_only=TRUE).3) A vector of length 1 (usually a parameter of the ihwResult object such as nfolds or the total number of rejections).4) A data.frame (as.data.frame) or just console output (show) for the extended Base generics.See section below for the individual methods.

Methods (by generic)

• adj_pvalues: Extract adjusted pvalues
• weights: Extract weights
• thresholds: Calculate ihw thresholds
• pvalues: Extract pvalues
• weighted_pvalues: Extract weighted pvalues
• covariates: Extract covariates
• covariate_type: Extract type of covariate ("ordinal" or "nominal")
• groups_factor: Extract factor of stratification (grouping) variable
• nfolds: Extract number of folds
• nbins: Extract number of bins
• alpha: Extract nominal significance (alpha) level
• rejections: Total number of rejected hypotheses by ihw procedure
• rejected_hypotheses: Get a boolean vector of the rejected hypotheses
• regularization_term: Extract vector of regularization parameters used for each stratum
• m_groups: Extract total number of hypotheses within each stratum
• as.data.frame: Coerce ihwResult to data frame
• nrow: Return number of p-values
• show: Convenience method to show ihwResult object

Slots

df

A data.frame that collects the input data, including the vector of p values and the covariate, the group assignment, as well as outputs (weighted p-values, adjusted p-values)

weights

A (nbins X nfolds) matrix of the weight assigned to each stratum

alpha

Numeric, the nominal significance level at which the FDR is to be controlled

nbins

Integer, number of distinct levels into which the hypotheses were stratified

nfolds

Integer, number of folds for pre-validation procedure

regularization_term

Numeric vector, the final value of the regularization parameter within each fold

m_groups

Integer vector, number of hypotheses tested in each stratum

penalty

Character, "uniform deviation" or "total variation"

covariate_type

Character, "ordinal" or "nominal"

adjustment_type

Character, "BH" or "bonferroni"

reg_path_information

A data.frame, information about the whole regularization path. (Currently not used, thus empty)

solver_information

A list, solver specific output, e.g. were all subproblems solved to optimality? (Currently empty list)

See Also

ihw, plot,ihwResult-method

Examples

save.seed <- .Random.seed; set.seed(1)
X   <- runif(n = 20000, min = 0.5, max = 4.5)       # Covariate
H   <- rbinom(n = length(X), size = 1, prob = 0.1)  # Is the null hypothesis (mean=0) true or false ?
Z   <- rnorm(n = length(X), mean = H * X)           # Z-score
.Random.seed <- save.seed

pvalue <- 1 - pnorm(Z)                              # pvalue
ihw_res <- ihw(pvalue, covariates = X, alpha = 0.1)
rejections(ihw_res)
colnames(as.data.frame(ihw_res))