Comprehensive non-parametric regression methods including kernel regression, local polynomial smoothing (LOESS), spline regression, and robust regression techniques designed for clinical research where parametric assumptions are violated.
Usage
nonparametricregression(
data,
outcome,
predictors,
grouping_variable = NULL,
regression_type = "kernel",
kernel_type = "gaussian",
bandwidth_method = "cross_validation",
manual_bandwidth = 0.1,
loess_span = 0.75,
loess_degree = "linear",
loess_iterations = 3,
spline_type = "smooth",
spline_df = 4,
spline_lambda = 0.1,
quantile_tau = 0.5,
quantile_method = "interior_point",
gam_smoother = "cubic_spline",
gam_basis_dim = 10,
cv_folds = 10,
cv_repeats = 1,
model_selection_criterion = "cv_error",
robust_regression = FALSE,
robust_method = "huber",
outlier_threshold = 3,
confidence_level = 0.95,
bootstrap_ci = TRUE,
bootstrap_samples = 1000,
prediction_intervals = TRUE,
residual_diagnostics = TRUE,
influence_diagnostics = FALSE,
goodness_of_fit = TRUE,
show_fitted_curves = TRUE,
show_confidence_bands = TRUE,
show_prediction_bands = FALSE,
show_residual_plots = TRUE,
show_qq_plots = TRUE,
show_partial_plots = FALSE,
show_model_summary = TRUE,
show_parameter_estimates = TRUE,
show_model_comparison = FALSE,
show_interpretation = TRUE,
multivariate_method = "additive",
missing_data_handling = "complete_cases",
clinical_context = "general",
set_seed = TRUE,
seed_value = 42,
parallel_processing = FALSE,
n_cores = 2
)Arguments
- data
The data as a data frame for non-parametric regression analysis.
- outcome
.
- predictors
.
- grouping_variable
.
- regression_type
.
- kernel_type
.
- bandwidth_method
.
- manual_bandwidth
.
- loess_span
.
- loess_degree
.
- loess_iterations
.
- spline_type
.
- spline_df
.
- spline_lambda
.
- quantile_tau
.
- quantile_method
.
- gam_smoother
.
- gam_basis_dim
.
- cv_folds
.
- cv_repeats
.
- model_selection_criterion
.
- robust_regression
.
- robust_method
.
- outlier_threshold
.
- confidence_level
.
- bootstrap_ci
.
- bootstrap_samples
.
- prediction_intervals
.
- residual_diagnostics
.
- influence_diagnostics
.
- goodness_of_fit
.
- show_fitted_curves
.
- show_confidence_bands
.
- show_prediction_bands
.
- show_residual_plots
.
- show_qq_plots
.
- show_partial_plots
.
- show_model_summary
.
- show_parameter_estimates
.
- show_model_comparison
.
- show_interpretation
.
- multivariate_method
.
- missing_data_handling
.
- clinical_context
.
- set_seed
.
- seed_value
.
- parallel_processing
.
- n_cores
.
Value
A results object containing:
results$modelSummary | a table | ||||
results$parameterEstimates | a table | ||||
results$bandwidthSelection | a table | ||||
results$modelFit | a table | ||||
results$residualDiagnostics | a table | ||||
results$influenceDiagnostics | a table | ||||
results$predictionTable | a table | ||||
results$goodnessOfFit | a table | ||||
results$crossValidation | a table | ||||
results$modelComparison | a table | ||||
results$clinicalInterpretation | a html | ||||
results$methodsExplanation | a html | ||||
results$fittedCurvePlot | an image | ||||
results$residualPlots | an image | ||||
results$qqPlots | an image | ||||
results$partialEffectPlots | an image | ||||
results$influencePlots | an image | ||||
results$bandwidthPlots | an image |
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$modelSummary$asDF
as.data.frame(results$modelSummary)
Examples
# Kernel regression with automatic bandwidth selection
nonparametricregression(
data = clinical_data,
outcome = "biomarker_response",
predictors = c("age", "weight", "baseline_score"),
regression_type = "kernel",
bandwidth_method = "cross_validation"
)