Partial Least Squares Cox regression for high-dimensional survival data. Combines PLS dimensionality reduction with Cox proportional hazards modeling for analysis of genomic, proteomic, and other high-dimensional datasets.
Usage
plscox(
time,
status,
predictors,
pls_components = 5,
cross_validation = "k10",
component_selection = "cv_loglik",
scaling_method = "standardize",
pls_algorithm = "nipals",
max_iterations = 100,
tolerance = 1e-06,
bootstrap_validation = FALSE,
n_bootstrap = 200,
permutation_test = FALSE,
n_permutations = 100,
plot_components = TRUE,
plot_loadings = TRUE,
plot_scores = TRUE,
plot_validation = TRUE,
plot_survival = TRUE,
risk_groups = 3,
confidence_intervals = TRUE,
feature_importance = TRUE,
prediction_accuracy = TRUE
)Arguments
- time
.
- status
.
- predictors
.
- pls_components
Number of PLS components to extract
- cross_validation
Cross-validation method for component selection
- component_selection
Method for selecting optimal number of components
- scaling_method
Method for scaling predictor variables
- pls_algorithm
Algorithm for PLS computation
- max_iterations
Maximum iterations for PLS algorithm convergence
- tolerance
Tolerance for algorithm convergence
- bootstrap_validation
Perform bootstrap validation of model performance
- n_bootstrap
Number of bootstrap replications
- permutation_test
Perform permutation test for variable importance
- n_permutations
Number of permutations for significance testing
- plot_components
Create PLS component visualization plots
- plot_loadings
Display variable loadings for PLS components
- plot_scores
Show component scores and survival relationships
- plot_validation
Display cross-validation curves for component selection
- plot_survival
Generate risk-stratified survival curves
- risk_groups
Number of risk groups for survival stratification
- confidence_intervals
Calculate confidence intervals for hazard ratios
- feature_importance
Calculate and display variable importance scores
- prediction_accuracy
Assess model prediction accuracy using C-index and other metrics
Value
A results object containing:
results$todo | a html | ||||
results$modelSummary | a html | ||||
results$componentSelection | a table | ||||
results$modelCoefficients | a table | ||||
results$variableLoadings | a table | ||||
results$modelPerformance | a table | ||||
results$riskStratification | a table | ||||
results$componentPlot | an image | ||||
results$loadingsPlot | an image | ||||
results$scoresPlot | an image | ||||
results$validationPlot | an image | ||||
results$survivalPlot | an image | ||||
results$bootstrapResults | a html | ||||
results$permutationResults | a html | ||||
results$clinicalGuidance | a html | ||||
results$technicalNotes | a html |
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$componentSelection$asDF
as.data.frame(results$componentSelection)