dynaris is a JAX-powered Python library for Dynamic Linear Models -- from composable DLM components to Kalman filtering, smoothing, forecasting, and parameter estimation, all with automatic differentiation.
pip install dynaris
# or
uv add dynarisFull documentation is available at dynaris.readthedocs.io.
from dynaris import LocalLevel, DLM
from dynaris.datasets import load_nile
# Load data
y = load_nile()
# Build a local-level model and fit
dlm = DLM(LocalLevel(sigma_level=38.0, sigma_obs=123.0))
dlm.fit(y).smooth()
# Forecast and plot
fc = dlm.forecast(steps=10)
print(dlm.summary())
dlm.plot(kind="panel")Build models by combining components with +:
from dynaris import LocalLinearTrend, Seasonal, Cycle
model = (
LocalLinearTrend(sigma_level=1.0, sigma_slope=0.1)
+ Seasonal(period=12, sigma_seasonal=0.5)
+ Cycle(period=40, damping=0.95)
)| Component | State dim | Description |
|---|---|---|
LocalLevel |
1 | Random walk + noise |
LocalLinearTrend |
2 | Level + slope |
Seasonal |
period - 1 | Dummy or Fourier form |
Regression |
n_regressors | Dynamic/static coefficients |
Autoregressive |
order | AR(p) in companion form |
Cycle |
2 | Damped stochastic sinusoid |
import jax.numpy as jnp
from dynaris import LocalLevel
from dynaris.estimation import fit_mle
def model_fn(params):
return LocalLevel(
sigma_level=jnp.exp(params[0]),
sigma_obs=jnp.exp(params[1]),
)
result = fit_mle(model_fn, y, init_params=jnp.zeros(2))
print(f"Log-likelihood: {result.log_likelihood:.2f}")| Dataset | Loader | N | Frequency | Domain |
|---|---|---|---|---|
| Nile river flow | load_nile() |
100 | Annual | Hydrology |
| Airline passengers | load_airline() |
144 | Monthly | Transportation |
| Lynx population | load_lynx() |
114 | Annual | Ecology |
| Sunspot numbers | load_sunspots() |
288 | Annual | Astronomy |
| Global temperature | load_temperature() |
144 | Annual | Climate |
| US GDP growth | load_gdp() |
319 | Quarterly | Economics |
MIT License. See LICENSE for details.