Detect Modes in a Dataset — findModes • controlDim

Detects peaks (modes) in a numeric dataset x using kernel density estimation. For each mode, it determines the range along the X-axis where the density decreases toward the adjacent minima. The function can also plot the density and the detected modes.

Usage

findModes(x, bw = 2, Q1 = 0.95, plot = TRUE)

Arguments

x: Numeric vector. The data for which modes will be calculated.
bw: Numeric. Bandwidth for the density estimation. Default is bw = 2.
Q1: Numeric between 0 and 1. Percentile of the density used to filter the most relevant peaks. Default is Q1 = 0.95.
plot: Logical. If TRUE, generates a plot of the density with detected modes. Default is TRUE.

Value

A matrix with one row per detected mode and two columns (min and max) representing the X-axis range of each mode.

Examples

set.seed(123) # Seed
x <- c(
  rnorm(200, mean = 2, sd = 0.3),
  rnorm(5, mean = 3.5, sd = 0.1),
  rnorm(200, mean = 5, sd = 0.3)
)

# Compute and plot modes
modes <- findModes(x, bw = 0.2, Q1 = 0.5, plot = TRUE)


# X-axis ranges for each mode
modes
#>             min      max
#> mode1 0.7072493 3.156928
#> mode2 3.9328437 6.371437