v1.5

Updates in this release

• Kernel related parts were isolated and moved to two new files: kernel.h and kernel.cpp
• One-vs-All SVM Taxonomy added
• MCSVM Taxonomy using predicted labels added
• MCSVM with Equal Length Taxonomy using maximal similarity score added

v1.4.1

Hotfixes in this release

• Fixed a segment fault in FreeProblem(), now this function checks before freeing memory.

v1.4

Updates in this release

• README file added, including detailed information on LibVM
• Data file example iris_scale and iris_scale_t added

v1.3

Updates in this release

• Full featured SVM-EL (SVM with equal length) taxonomy has been added. Using option: -t 1 to choose this taxonomy.
• Full featured SVM-ES (SVM with equal size) taxonomy has been added. Using option: -t 2 to choose this taxonomy.
• Full featured SVM-KM (SVM with k-means clustering) taxonomy has been added. Using option: -t 3 to choose this taxonomy.
• Cross validation support for all taxonomies has been added, see vm-cv for details.

v1.2

Updates in this release

• A simple taxonomy for Support Vector Machine has been added. Using option: -t 1 to choose this taxonomy.
• KNN library which is used to hold knn-related functions has been redesigned.
• Probabilistic prediction metrics: brier score and logarithm loss score have been added to the output.

v1.1

Updates in this release

• Time measurement added to both online and offline mode

v1.0

Features includes in this release

• Venn Machine with k-Nearest Neighbors as taxonomy
• Online and Offline mode for Venn Machine
• Multi-class problem supported
• Classification only