| Bayesian Dynamic Linear Model for time-series analysis

/main page/OpenBDLM

What is it?

OpenBDLM is a Matlab open-source software developed to use Bayesian Dynamic Linear Models for time series analysis having time steps in the order of one hour or higher. OpenBDLM is capable to process simultaneously several time series data to interpret, monitor and predict their long-term behavior. The approach works by modeling the raw time series using a superposition of hidden state variables describing the baseline, external effects and residuals. OpenBDLM includes an anomaly detection tool which allows to detect abnormal behavior in a fully probabilistic framework. OpenBDLM handles time series with missing data and non-uniform timestep vector.

Go see our YouTube channel for videos and tutorials

How to cite?

OpenBDLM, an Open-Source Software for Structural Health Monitoring using Bayesian Dynamic Linear Models
Gaudot, I., Nguyen, L.H., Khazaeli S. and Goulet, J.-A.
In the proceedings from the 13th International Conference on Applications of Statistics and Probability in Civil Engineering (ICASP13), May 2019
[PDF] [EndNote] [BibTex]

Where to download it?