AR(1)
A persistent autoregressive process with nine lagged candidate predictors, useful for forecast and variable-selection experiments.
synthesis · Statistical data generator
Build known systems. Test honestly.
synthesis generates controlled stochastic, nonlinear and chaotic datasets for testing prediction, variable-selection, classification and clustering methods.
Included online
Stochastic and chaotic benchmarks.
Logistic map
A compact nonlinear recurrence whose control parameter moves the system from stable behaviour into deterministic chaos.
Lorenz system
A three-state dynamical system integrated with fourth-order Runge–Kutta to reveal its characteristic attractor.
Online workflow
Configure → Generate → Inspect → Export.
Reproducible controls
Choose 100–5000 observations, set an integer seed and noise level, then adjust the parameters exposed for the selected model.
Immediate outputs
Inspect the time series, model-specific state-space plot and summary statistics, then download the complete generated dataset.
Software and research
Continue with the complete R package.
synthesis for R
The package contains a broader collection of linear, nonlinear, dynamic, classification and water-quality generators.
GitHub ↗ · CRAN ↗Advances in Water Resources · 2019
Assessing the sensitivity of hydro-climatological change detection methods to model uncertainty and bias.
Read the application paper ↗Generate a benchmark dataset.
Adjust a model, inspect its behaviour and download a reproducible CSV—all in the browser.