BNSP: Bayesian Non- and Semi-Parametric Model Fitting Algorithms

Georgios Papageorgiou developed a collection of MCMC algorithms and processing functions for Bayesian non- and semi-parametric models. The software implements methods for single and multivariate response regression, joint mean-covariance modeling, and Dirichlet process mixtures, as documented in multiple peer-reviewed publications. The code is associated with research from Wuhan University and is available under an Open Access (gold) license.

Use Cases

• Fitting single-response multiple regression models based on the referenced Papageorgiou (2018) methodology.
• Performing multivariate regression with nonparametric models for means, variances, and correlations, including variable selection as described in Papageorgiou and Marshall (2020).
• Modeling joint mean-covariance structures for multivariate responses based on the Papageorgiou (2022) paper.
• Implementing Dirichlet process mixture models as outlined in Papageorgiou (2019).

Strengths

• Implements methods from four peer-reviewed publications, providing a documented theoretical foundation.
• Covers a range of advanced Bayesian modeling techniques, including multivariate and nonparametric approaches.
• Released under an Open Access (gold) license, facilitating use and modification.

Limitations

• Column-level documentation is absent; field semantics must be inferred from the academic papers or code.
• Row count is unknown, which may limit suitability assessment for specific applications.
• Last update date is unknown; freshness unverified.

Provenance

Source

Georgios Papageorgiou, Wuhan University

Collection Method

Software implementation of published statistical methodologies.