Usage
coefplot(
data,
dep = NULL,
covs = NULL,
model_type = "linear",
time_var = NULL,
include_intercept = FALSE,
coef_selection = "all",
specific_coefs = "",
ci_level = 0.95,
inner_ci_level = 0.8,
sort_coefs = "natural",
decreasing_sort = TRUE,
horizontal_plot = TRUE,
point_size = 3,
line_thickness = 1,
standardize = FALSE,
robust_se = FALSE,
exp_transform = FALSE,
compare_models = FALSE,
model2_covs = NULL,
model3_covs = NULL,
model_names = "Model 1, Model 2, Model 3",
show_coefficient_plot = TRUE,
show_model_summary = TRUE,
show_coefficient_table = FALSE,
custom_title = "",
custom_x_label = ""
)Arguments
- data
The data as a data frame for regression analysis.
- dep
The outcome variable for regression analysis. Can be continuous (linear regression), binary (logistic regression), or time-to-event.
- covs
Independent variables (predictors) to include in the regression model. These will be displayed as coefficients in the plot.
- model_type
Type of regression model to fit: - Linear: For continuous outcomes - Logistic: For binary outcomes (odds ratios) - Cox: For survival analysis (hazard ratios) - Poisson: For count outcomes (rate ratios)
- time_var
Time-to-event variable for Cox regression. Only required when model type is set to Cox regression.
- include_intercept
Include the intercept term in the coefficient plot. Usually excluded as it's often not of primary interest.
- coef_selection
How to select which coefficients to display in the plot.
- specific_coefs
Comma-separated list of coefficient names to include or exclude (depending on selection method). Leave blank to use all coefficients.
- ci_level
Confidence level for coefficient confidence intervals (e.g., 0.95 for 95\
inner_ci_levelOptional inner confidence interval for enhanced visualization. Set to 0 to disable inner CI. Common values are 0.8 or 0.9.
sort_coefsHow to order coefficients in the plot. Magnitude sorting can help identify the most important predictors.
decreasing_sortWhen sorting by magnitude or alphabetically, use decreasing order.
horizontal_plotDisplay coefficients horizontally (default) or vertically. Horizontal layout is typically preferred for readability.
point_sizeSize of the coefficient points in the plot.
line_thicknessThickness of the confidence interval lines.
standardizeStandardize coefficients by scaling predictors to have mean 0 and SD 1. Useful for comparing effect sizes across variables with different scales.
robust_seUse robust (sandwich) standard errors for confidence intervals. Recommended when there are concerns about heteroscedasticity.
exp_transformExponentiate coefficients to show odds ratios (logistic), hazard ratios (Cox), or rate ratios (Poisson). Automatically enabled for logistic and Cox models.
compare_modelsCreate comparison plots for multiple model specifications. Useful for sensitivity analysis or model selection.
model2_covsCovariates for second model comparison. Only used when comparing models.
model3_covsCovariates for third model comparison. Only used when comparing models.
model_namesComma-separated names for models when comparing multiple models. Will be used in the legend.
show_coefficient_plotDisplay the main coefficient plot with confidence intervals.
show_model_summaryDisplay statistical summary of the fitted model(s) including R-squared, AIC, and other fit statistics.
show_coefficient_tableDisplay detailed table of coefficients, standard errors, and p-values.
custom_titleCustom title for the coefficient plot. Leave blank for automatic title.
custom_x_labelCustom label for x-axis. Leave blank for automatic label based on model type.
A results object containing:
results$instructions | a html | ||||
results$coefficient_plot | an image | ||||
results$model_summary | a html | ||||
results$coefficient_table | a html |