rts-smoother 0.1.0

High-performance RTS Kalman smoother for column-wise financial matrices (local level model)

RTS Kalman Smoother (C++ × Python)

The package provides a high-performance Rauch–Tung–Striebel Kalman smoother for the local level model. This model is a canonical choice for asset prices and similar data (random walk latent state).
It smooths each asset (column) independently with global process/observation variances (Q, R) shared across assets.

Features

• Handles NaN observations (prediction-only step when missing)
• Returns smoothed states as the same object type
• Diffuse initialization

Install (from PyPI)

pip install rts_smoother

Install (from source)

pip install -U pip build twine
pip install -e .

Usage

from rts_smoother import smooth

# Provide Q and R manually
df_smoothed = smooth(df, Q=1e-4, R=1e-3)

Model

• State: x_t = x_{t-1} + w_t, w_t ~ N(0, Q)
• Obs: y_t = x_t + v_t, v_t ~ N(0, R)

Notes

• Built with scikit-build-core and pybind11.
• The compiled extension lives at rts_smoother/_core.* inside the wheel.