Recursive Mode Detection for Distributions of Ordinal Data

Provides the function remode() for recursive modality detection in ordinal data. 'remode' is an algorithm specifically designed to estimate the number and location of modes in ordinal data while being robust to large sample sizes.

ReMoDe

ReMoDe is an R package designed for the robust detection of modes in data distributions. It uses statistical tests, including Fisher's exact test and binomial tests, to determine if a given maximum in a data distribution is a true local maximum. It was specifically designed for mode detection in ordinal distributions but can also be applied to continuous data.

Features

• Mode Detection based on recursive statistical testing to identify local maxima in the distribution
• Data formatting: Converts raw data into count data for analysis. Input can thus be in counts or raw
• Stability Analysis: Includes functionality to assess the stability of the number and location of detected modes using jackknife resampling
• Visualization: Provides methods to plot the histogram of data along with identified modes, as well as outcomes of the stabilty analysis

Usage

# count data as input
xt <- c(8,20,5,2,5,2,30)

# apply ReMoDe algorithm 
results <- remode(xt)

# plot result
plot(results)

# Perform stability analysis 
stability_info <- remode_stability(results)

Citation

Please cite the following paper

To be added