Welcome to TFilterPy’s Documentation!
Where we say: Sawubona, Molo, Hallo, Dumela, Ndaa, Xewani, and a heartfelt Dinstang in pure Sipitori style 🇿🇦✨.
TFiltersPy is your go-to open-source Python library for applying state-of-the-art Bayesian filtering — built with the power of Dask for scale, the elegance of Kalman and Particle Filters for precision, and a sprinkle of local flavor to make it proudly African.
✨ This library is part of the Ubunye Artificial Intelligent Ecosystems (UIAE) — a collaborative initiative to build powerful, locally rooted, and globally relevant AI tools. Explore more projects at 👉 https://github.com/ubunye-ai-ecosystems
🚀 Whether you’re building AI for space tech, smart grids, autonomous cars, or township telemetry — if it’s noisy, dynamic, and uncertain, we’ve got you covered.
Features
Bayesian Filtering: Supports both linear (Kalman) and nonlinear/non-Gaussian (Particle) filtering for robust state estimation in dynamic systems.
Distributed Computation: Built on Dask, enabling parallel and out-of-core filtering for large-scale or streaming data.
Uncertainty Quantification: Includes tools to quantify estimation confidence through residual analysis, covariance estimation, and adaptive strategies.
Parameter Estimation: Advanced methods for estimating system parameters using Bayesian techniques, including maximum likelihood and cross-validation.
User-Friendly API: Clean, modular, and scikit-learn like API — with examples, documentation, and sensible defaults so you can go from idea 💡 to insight 📈 fast.