Github address:

Research papers:

Spectrum: Fast density-aware spectral clustering for single and multi-view data

This spectral clustering algorithm can self-tune to the data, it adapts to local density and scale. It can integrate heterogenous data sources (multi-view) and cluster them, or cluster a single-view. It reduces noise through a tensor product graph diffusion technique. Spectrum can automatically detect Gaussian and non-Gaussian structures. Available from CRAN. Preprint. Published in Bioinformatics.


Featured in RStudio’s top 40 new CRAN packages of February 2019 link

M3C: Monte Carlo reference-based consensus clustering

M3C is a consensus clustering algorithm based on the Monti algorithm, which uses a Monte Carlo simulation to generate null reference datasets to eliminate overfitting and test the null hypothesis K=1. M3C uses an entropy objective function to measure consensus matrix uncertainty when deciding K, with the aim to minimise uncertainty in the system to find K. Available from the Bioconductor. Preprint. Published in Scientific Reports.


MLeval: Machine learning model evaluation

Evaluates machine learning models to make ROC curves, PR curves, PRG curves, and calibration curves. Also includes a number of key performance metrics with confidence intervals. MLeval can detect the caret output and automatically evaluate its results to speed up subsequent analyses. Available from CRAN.


Clusterlab: Flexible Gaussian cluster simulator

A flexible Gaussian cluster simulator for R that generates circles of Gaussian clusters and other shapes through vector rotations and transformations. Available from CRAN. Preprint. Published in Scientific Reports.