quant: Quantitative tests of pypomp

These quantitative tests, or simply quant tests, are designed to assess the accuracy and performance of pypomp for problems existing on a scale too large to be run on a laptop within the unit tests in pypomp:pypomp/tests.

The quant tests also provide additional examples of pypomp, focused on technical issues that extend the simpler examples in pypomp:tutorials.

🚀 Quick Links

💻 Core Repository — Source code and issue tracker for the pypomp package.
📖 Read the Documentation — Core library API reference and user guide.
🎓 Tutorials — Examples and tutorials for learning pypomp.

Quantitative Tests Index

Below is a list of quantitative test reports available in this repository:

1. SPX (S&P 500) Model

SPX Report (tests/spx): Compares parameter estimation traces and log-likelihood estimates on the SPX index dataset using pypomp (CPU/GPU) and R’s pomp.

2. Dhaka Cholera Model

Dhaka Report (tests/dacca): Analyzes the performance, runtime, and parameter convergence of IF2 versus IFAD for the Dhaka cholera model. Also checks that the particle filter yields the same distribution of log likelihoods in both pypomp and pomp.

3. Random Number Generators

Random Number Generators Benchmark & Comparison (tests/samplers): Benchmarks the execution speed and validates the statistical accuracy of pypomp’s fast approximate inverse CDF samplers (fast_poisson, fast_binomial, fast_gamma, fast_nbinomial) against jax.random and scipy.stats.

4. Measles Model

Log-Likelihood and Parameter Comparison: Pypomp vs R (tests/measles/R_comparison): Compares distributions of log-likelihood and parameter estimates obtained via pypomp versus R’s pomp.
Runtime Comparison: Pypomp vs R (tests/measles/R_comparison/speed_comparison): Benchmarks runtime of IF2 on a discrete measles model using pypomp versus pomp. This demonstrates the utility of the fast samplers in pypomp.random.

5. Panel Measles Model

Panel Measles (only unit-specific parameters) (tests/panel_measles/performance): Benchmarks panel parameter estimation via MPIF on a panel measles dataset using only unit-specific parameters.
Panel Measles (mixed parameters) (tests/panel_measles/performance_mixed): Benchmarks panel parameter estimation via MPIF on a panel measles dataset using both unit-specific and shared parameters.

6. Differentiated Process Off-Parameter Filtering

DPOP London Measles Benchmark (tests/dpop/london_measles): Compares DPOP training against an IF2 baseline on the single-unit London measles model, including likelihood distribution, elapsed-time trace, and runtime summaries.

7. Bayesian Iterated Filtering

BIF versus PMCMC Benchmarks (tests/bif/sir_dacca_pmcmc): Compares BIF against PMCMC on a four-parameter SIR model and a four-parameter, 100-observation Dacca cholera benchmark, including runtime, posterior marginal, interval, and PMCMC diagnostic summaries.