Getting Started

Once you've installed and loaded Mica.jl, you're ready to begin modeling and detecting changepoints.

First Steps

Try a simple built-in model using the simulate_model function:

using Mica

model = exponential_ode_model()
sol = simulate_model(model)

This simulates a basic exponential decay model defined by an ordinary differential equation (ODE), using default parameters and initial conditions.

Core Workflow

A typical workflow in Mica.jl looks like this:

1. Define or select a model Use one of the built-in examples (exponential_ode_model, example_difference_model, etc.) or define your own model structure using the provided ODEModelSpec, DifferenceModelSpec, or RegressionModelSpec types.

2. Simulate the model Use simulate_model(model) to generate synthetic or fitted outputs.

3. Format data and objective Prepare your observed data in a matrix format and define a loss function to quantify the difference between model predictions and data.

4. Run changepoint detection Use detect_changepoints(...) to estimate both the changepoint locations and the model parameters in each segment. The algorithm applies evolutionary optimization to jointly minimize the simulation error and a penalty function.

Next Steps

To learn more:

• Explore Tutorials: Guided examples for ODEs and discrete systems
• Review Problem Types: Understand which models are supported
• Try Examples Realistic use cases and model setups
• See Reference: API documentation for all core types and functions