Partiview & Ndaona: A good way to study Machine Learning

Partiview is a piece of software for interactive 4D-viewer dataset. It is also useful for 3D visualization of dataset in Machine Learning algorithm, i.e. show manifolds or classification results.

Ndaona is a piece of Matlab functions for Partiview-compatible data conversion.

Downloads

• PartiView can be reached here (binaries + source codes)
• Ndaona is here

Screenshots

I borrow some screenshots from Ndaona’s website to illustrate here:

How well a linear svm (in LIBSVM) classifies Mandarin tones

How well does Principal Components Analysis separate faces

Resources on Information Visualization

• Maps of Science
• A Blog on Information Visualization
• 20 Useful Visualization Libraries
• GapMinder(Google)
• 35 Great Visualization Examples
• Speaking of Graphics
• History of Data Visualization
• The Information Visualization Community Platform
• The XMDVTool
• Multidimensional Visualization Methods http://pisuerga.inf.ubu.es/cgosorio/Visualization/?Multidimensional_visualization_methods
• An Investigation of Methods for Visualizing Highly Multivariate Datasets http://www.agocg.ac.uk/reports/visual/casestud/brunsdon/contents.htm
• Interactive Information Visualization of a Million Items http://www.cs.umd.edu/hcil/VisuMillion/
• Interactive Multimedia Technology http://interactivemultimediatechnology.blogspot.com/
• HCI Lab, Univ. Maryland http://www.cs.umd.edu/hcil/
• JavaScript InfoVis Toolkit http://thejit.org/demos/
• Ontology Online http://ontologyonline.org/
• Xebece Library http://xebece.sourceforge.net/index
• Visual Complexity http://www.visualcomplexity.com/vc/