The Maasai team is committed to share as much as possible the methodologies its members propose as R or Python libraries, or SaaS platforms. A non-exhaustive list of the software production of the team is proposed below. Please visit the Maasai member web pages for complete lists.
SaaS platforms
- Linkage: a statistical AI algorithm to analyze communication networks. It allows you to cluster the nodes of networks with textual edges while identifying topics which are used in communications. You can analyze with Linkage networks such as email networks or co-authorship networks. Linkage allows you to upload your own network data or to make requests on scientific databases (Arxiv, Pubmed, HAL).
- Topix: the platform implements an innovative AI-based solution allowing to summarize massive and possibly extremely sparse data bases involving text, through co-clustering. Topix is a versatile technology that can be applied in a large variety of situations where large matrices of texts / comments / reviews are written by users on products or addressed to other individuals (bi-partite networks).
R packages
- OrdinalLBM: The oLBM algorithm allows to simultaneously cluster the rows and the columns of a data matrix where each entry of the matrix is an ordinal data. Available on the CRAN.
- FunLBM: The funLBM algorithm allows to simultaneously cluster the rows and the columns of a data matrix where each entry of the matrix is a function or a time series. Available on the CRAN.
- SpinyReg: this package implements a generative model for Bayesian variable selection in high-dimensional linear regression. Available on the CRAN.
- FisherEM: the package provides the FisherEM algorithm for the clustering of high-dimensional data. Available on the CRAN.
- HDclassif: the package implements the HDDC and HDDA algorithms designed respectively for the clustering and classification of high-dimensional data. Available on the CRAN.
Python
- Tabular LIME theory: this package contains the code for the theory behind Tabular LIME.
- HDMI for image denoising: this notebook implements the High-Dimensional Mixture Models (HDMI) algorithm for unsupervised single image denoising.
- MIWAE and not-MIWAE: Jupyter notebooks that illustrate how to deal with missing data using deep generative models.
- GemClus: this package implements different discriminative clustering models that can be trained with distance-generalised mutual information (GEMINI)
Julia
- PEN: deep learning for approximate likelihood-free Bayesian inference
