Partenaires



UNIVERSITE

PARIS DESCARTES

MAP5

Pages personnelles

Softwares Charles Bouveyron > Softwares

R packages:

- SpinyReg: this package implements a generative model for Bayesian variable selection in high-dimensional linear regression. It uses a spike-and-slab like prior distribution obtained by multiplying a deterministic binary vector with with a random Gaussian parameter vector. Such a model allows the use of an EM algorithm, optimizing a type-II log-likelihood, for inference. Available on the CRAN: [web].

- ProbFDA: this package proposes the probabilistic Fisher discriminant analysis technique for dimensionality reduction and classification. The pFDA method works at least as well as the traditional FDA method in standard situations and it clearly improves the modeling and the prediction when the dataset is subject to label noise and/or sparse labels. Available on the CRAN: [web].

- RobustDA: the package implements the robust mixture discriminant analysis (RMDA) which allows to build a robust supervised classifier from learning data with label noise. Available on the CRAN: [web].

- AdaptDA: this package provides the adaptive mixture discriminant analysis (AMDA) which allows to adapt a model-based classifier to the situation where a class represented in the test set may have not been encountered earlier in the learning phase. Available on the CRAN: [web].

- FunFEM: the package implements the funFEM algorithm for the clustering of functional data within a disciminative functional subspace. Available on the CRAN: [web].

- Rambo: the package proposes the VB-EM algorithm for the random subgraph model (RSM) which allows to cluster the vertices of a directed network with typed edges into clusters describing the connection patterns of subgraphs given as inputs. Available on the CRAN: [web].

- FisherEM: the package provides the FisherEM algorithm for the clustering of high-dimensional data. Available on the CRAN: [web].

- HDclassif: the package implements the HDDC and HDDA algorithms designed respectively for the clustering and classification of high-dimensional data. Available on the CRAN: [web].

- FunHDDC: the package implements the funHDDC algorithm which allows the clustering of functional data within group-specific functional subspaces. Available on the CRAN: [web].

- AdaptReg: the package provides tools for transfering a regression model on a reference population to a new population with only few observations. Download link: [tgz].

- LLN: the package provides tools for learning supervised classifiers on networks and classifying new arriving nodes in the network. Download link: [web].

Matlab toolboxes:

- PgpDA: the toolbox implements a classification method based on a family of parsimonious Gaussian process models. This allows in particular to use non-linear mapping functions which project the observations into infinite dimensional spaces. Download link: [tgz]

- HDDA/HDDC: the toolbox implements the HDDC and HDDA algorithms designed respectively for the clustering and classification of high-dimensional data. Dowload links: [web] and [web].


Photo