Software and Datasets at IDEAL

Hierarchical support vector machine code developed by Yangchi Chen. Its a Matlab package that uses Max-cut class decomposition a fast SVM multiclass classifier.

A Matlab package that includes BHC and BPC algorithms. Originally by Alex Henneguelle, updated by Suju Rajan and Goo Jun.

Gene DIVER v1.02 (Gene Density Interactive Visual ExplorER) - developed by Gunjan Gupta

Datasets for the paper:
Arindam Banerjee, Inderjit S. Dhillon, Joydeep Ghosh, and Suvrit Sra, "Clustering on the Unit Hypersphere using von Mises-Fisher Distributions", JMLR 6(Sep):1345--1382, 2005

Matlab scripts for HMMs and Coupled HMMs for Multi-channel EEG classification

Model-based Document Clustering Toolkit in Matlab download (read README file in the package first, some document datasets are here (16MB), each dataset contains a sparse document-term matrix and a class label vector that can be read into Matlab directly using 'load datasetname')

Weighted Sequence Clustering Toolkit README, download

Cluster Analysis and Cluster Ensemble Software (ClusterPack Matlab Toolbox) .

Principal Surfaces software/data

You can also access a single hidden layer feed forward neural network code called bp.

Code Structurally Adaptive Mixtures of Experts Code (see papers co-authored with V. Ramamurti)

The hierearchical mixture of experts network
The structurally adapting on-line localized mixture of experts network
Structurally adapting batch-mode localized mixture of experts network for model order determination
Localized mixture of non-linear experts (MLP experts)


IDEAL Benchmarks Home