Uniform interface for positive definite matrices of various structures. Positive definite matrices are widely used in machine learning and probabilistic modeling, especially in applications related to graph analysis and Gaussian models. It is not uncommon that positive definite matrices used in practice have special structures (e.g. diagonal), which can be exploited to accelerate computation. PDMats.jl supports efficient computation on positive definite matrices of various structures. In particular, it provides uniform interfaces to use positive definite matrices of various structures for writing generic algorithms, while ensuring that the most efficient implementation is used in actual computation.
Features
- PDMats.jl supports efficient computation on positive definite matrices of various structures
- All subtypes of AbstractPDMat share the same API
- It provides uniform interfaces to use positive definite matrices of various structures for writing generic algorithms
- Uniform interface for positive definite matrices of various structures
- Documentation available
- PRs to implement generic fallbacks are welcome
Categories
Data VisualizationLicense
MIT LicenseFollow PDMats.jl
You Might Also Like
Rate This Project
Login To Rate This Project
User Reviews
Be the first to post a review of PDMats.jl!