Skip to contents

dormancy

dormancy is a novel R package for detecting and analyzing dormant patterns in multivariate data. Unlike traditional pattern detection methods that focus on currently active relationships, dormancy identifies statistical patterns that exist but remain inactive until specific trigger conditions emerge.

What Makes This Package Unique?

This is the first statistical package dedicated to dormant pattern detection. The concept is inspired by:

  • Biological dormancy: Seeds remaining dormant until conditions are right
  • Geological phenomena: Dormant faults that can trigger earthquakes
  • Epidemiology: Latent infections that activate under stress

In data analysis, dormant patterns are relationships that:

  • Are strong in specific data regions but weak overall
  • Only emerge when certain thresholds are crossed
  • Are masked by confounding variables
  • Could trigger cascade effects when activated

Installation 

# Install from CRAN (when available)
install.packages("dormancy")

# Install development version
# devtools::install_github("danymukesha/dormancy")

Quick Example 

library(dormancy)

set.seed(42)
n <- 500

# Create data with a dormant pattern
x <- rnorm(n)
condition <- sample(c(0, 1), n, replace = TRUE)

# Relationship only exists when condition == 1
y <- ifelse(condition == 1, 
            0.8 * x + rnorm(n, 0, 0.3), 
            rnorm(n))

data <- data.frame(x = x, y = y, condition = factor(condition))

# Overall correlation is weak
cor(data$x, data$y)  # ~0.35

# Detect the dormant pattern
result <- dormancy_detect(data, method = "conditional")
print(result)
#> Dormant pattern detected: x ~ y
#> Dormancy score: 0.72
#> Trigger: condition == 1

Core Functions

Function Description
dormancy_detect() Detect dormant patterns using 4 methods
dormancy_trigger() Identify activation trigger conditions
dormancy_depth() Measure how deeply dormant a pattern is
dormancy_risk() Assess activation risk and potential impact
dormancy_scout() Map data space for potential dormant regions
awaken() Simulate what happens when patterns activate
hibernate() Find patterns that have become dormant over time

Detection Methods

1. Conditional Detection

Finds patterns that are conditionally suppressed - active only under specific conditions.

2. Threshold Detection

Identifies patterns that emerge when variables cross specific thresholds.

3. Phase Detection

Detects patterns that exist in specific phase regions of the data space.

4. Cascade Detection

Finds patterns that could trigger chain reactions through other variables.

Why Dormancy Matters

Traditional correlation analysis misses dormant patterns because:

  • Aggregate statistics mask conditional relationships
  • Weak overall correlations may hide strong local correlations
  • Threshold effects create piecewise relationships
  • Phase-dependent patterns vary across the data space

Use Cases

Financial Risk 

# Detect dormant correlations that could activate during market stress
result <- dormancy_detect(returns_data, method = "threshold")
risk <- dormancy_risk(result, time_horizon = 30)

Quality Control 

# Find patterns that only emerge under certain conditions
result <- dormancy_detect(process_data, method = "conditional")
triggers <- dormancy_trigger(result)

Environmental Monitoring 

# Identify dormant patterns signaling ecological shifts
scout <- dormancy_scout(sensor_data)
hib <- hibernate(time_series_data, time_var = "date")

Healthcare Analytics 

# Detect latent risk factors
result <- dormancy_detect(patient_data, method = "cascade")
awakening <- awaken(result, intensity = 1)

Key Concepts

Dormancy Score

Measures how “dormant” a pattern is (0 = active, 1 = fully dormant).

Trigger Conditions

The specific circumstances under which a dormant pattern would activate.

Depth of Dormancy

How much change is needed to awaken the pattern:

  • Shallow: Minor perturbation could activate
  • Deep: Significant systemic shift required

Cascade Potential

Risk that activating one pattern triggers others.

Citation

If you use dormancy in your research, please cite:

@Manual{dormancy,
  title = {dormancy: Detection and Analysis of Dormant Patterns in Data},
  author = {Dany Mukesha},
  year = {2026},
  note = {R package version 0.1.0},
  url = {https://github.com/danymukesha/dormancy}
}

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

License

MIT License. See LICENSE for details.

Acknowledgments

This package develops a novel statistical framework inspired by concepts from biology, geology, and epidemiology. The idea of dormant patterns in data analysis provides a new perspective on hidden relationships and latent risks.