Here you can find specific information about current and previous stable ELEFANT releases.

Version 0.2.0, released 29th July 2008

• New implementation of fast kernel modules
• Re-implemented some of the existing algorithms for speed and efficiency
• New clustering algorithms: Diffusion maps and KMeans
• Some new features to framework
• Enhancements to user interface like docking windows and quick menu toolbar, configure kernel and loss objects as sub properties, data inspector module and performance monitoring
• New data filtering components slice data and Shift index to zero base

Version 0.1.0, released 30th Nov 2007

• Component for visualizing image data
• Data components now supports compressed data formats like Bz2 and zip
• Installation setup program for MAC, Windows and Linux platform
• Cover Tree for calculating the nearest neighbor
• String Kernels
• Support for Windows platform
• Unit tests
• Bug fixes

Version alpha-0.1.0, released 4th July 2007

• Light weight component based system design, plug and play kind of a architecture
• Component suite for basic as well as advanced machine learning algorithms
• Support for various data source formats
• Components for data visualizations
• Easy to use graphical user interface for visual programming and quick prototyping
• Intuitive application programming interface for advanced prototyping
• Python wrappers for high-performance parallel scientific packages like PETSc, TAO, and SLEPc
• Comprehensive system documentation

Following machine learning algorithms are implemented in the Elefant:

• Support Vector Machine (SVM) for classification, regression, quantile and novelty detection, online learning, Epsilon Insensitive and Laplacian support vector regression.
• Gaussian Process Regression, Heteroscedastic Gaussian Process regression, Multi-class transductive classification with Gaussian Process.
• Solvers for the quadratic programming problem
• BAHSIC feature selection
• Algorithms for fast computation and manipulation of kernel matrices. Linear, RBF and Dot Product Kernels
• Loopy Belief Propagation and Junction Tree algorithms.

Following features are not available in this release and will be available in release 1.0.0 very soon

• Cover Tree for calculating the nearest neighbor
• String Kernels
• Interface to external systems like UIMA using jpype
• Unit tests
• Installation program
• Reference user manual for python wrapper modules for TAO and SLEPc libraries.