balancR is an R package which provides tools for
balancing imbalanced data in scaling analyses using bootstrapping, and
for comparing regression coefficients (slopes and intercepts) between
original and balanced samples.
library(balancR)Balance an imbalanced, log-normally distributed dataset, fit an SMA model to both imbalanced and balanced datasets, and compare slopes:
# Simulated imbalanced dataset
library(balancR)
set.seed(123)
df <- data.frame(x = rlnorm(1000, meanlog = 3, sdlog = 1), y = rlnorm(1000, meanlog = 2, sdlog = 1))
# Create balanced sample and fit with SMA model
bal <- balanced_scaling(data = df, var_x = x, var_y = y, min_per_bin = 10, n_boot = 1, seed = 1, model_type = "power" )$first_boot
# Compare slopes between balanced and imbalanced datasets
compare_coefficients(original = df, balanced = bal, var_x = x, var_y = y, which_coefficient = "slope", model_type = "power")