COBRAPerformance: 'COBRAPerformance' object and constructor

In csoneson/iCOBRA: Comparison and Visualization of Ranking and Assignment Methods

COBRAPerformance object and constructor

Description

The COBRAPerformance class holds various types of calculated performance measures. Objects from this class are typically generated from COBRAData objects by means of the function calculate_performance.

Usage

COBRAPerformance(
  fdrtpr = data.frame(),
  fdrtprcurve = data.frame(),
  fdrnbr = data.frame(),
  fdrnbrcurve = data.frame(),
  fsrnbr = data.frame(),
  fsrnbrcurve = data.frame(),
  tpr = data.frame(),
  fpr = data.frame(),
  splv = "",
  roc = data.frame(),
  fpc = data.frame(),
  deviation = data.frame(),
  onlyshared = NA,
  overlap = data.frame(),
  maxsplit = NA_integer_,
  corr = data.frame(),
  scatter = data.frame()
)

Arguments

fdrtpr	A data frame containing observed FDR and TPR values at various adjusted p-value thresholds.
fdrtprcurve	A data frame containing observed FDR and TPR values for a (potentially large) number of cutoffs applied to a 'score' (that can be p-value, adjusted p-value or a more general score).
fdrnbr	A data frame containing observed FDR and the number of features considered to be significant at various adjusted p-value thresholds.
fdrnbrcurve	A data frame containing observed FDR and number of features considered to be significant for a (potentially large) number of cutoffs applied to a 'score' (that can be p-value, adjusted p-value or a more general score).
fsrnbr	A data frame containing observed false sign rate (FSR) and the number of features considered to be significant at various s-value thresholds
fsrnbrcurve	A data frame containing observed false sign rate (FSR) and number of features considered to be significant for a (potentially large) number of cutoffs applied to the s-value
tpr	A data frame containing observed TPR values at various adjusted p-value thresholds.
fpr	A data frame containing observed FPR values at various adjusted p-value thresholds.
splv	A character string giving the name of the stratification factor, "none" if the results are not stratified.
roc	A data frame containing observed FPR and TPR values for a (potentially large) number of cutoffs applied to a 'score' (that can be p-value, adjusted p-value or a more general score), which can be used to generate a ROC curve.
fpc	A data frame containing observed numbers of false positive findings among the N top-ranked features (ranked by p-values, adjusted p-values or more general scores), for a (potentially large) number of Ns, which can be used to generate a false positive curve.
deviation	A data frame containing deviations between observed scores and true scores.
onlyshared	A logical value indicating whether only features shared between the results and the truth should be retained, or if all features present in the truth should be used.
overlap	A data frame or list of data frames with binary values indicating, for each of a number of methods and number of features, whether the method consider the feature 'positive' (significant, 1) or 'negative' (non-significant, 0). If it is a list of data frames, each list element corresponds to one level of a stratifying factor.
maxsplit	A numeric value indicating the largest number of levels to retain if the results have been stratified by an annotation.
corr	A data frame containing observed (Pearson and Spearman) correlation values between observed and true scores.
scatter	A data frame containing observed 'scores' (p-values, adjusted p-values or more general scores) and true scores, which can be used to generate scatter plots.

Value

A COBRAPerformance object.

Author(s)

Charlotte Soneson

Examples

## Empty COBRAPerformance object
COBRAPerformance()

csoneson/iCOBRA documentation built on May 3, 2024, 5:23 a.m.