FLLLame

FLLLame is a machine learning environment for supervised learning (N-class problems) with various capabilities:

• Support for multiple machine learning data formats (import and export)
• Multiple concurrent data sets
• Various data operations (data type changes, duplication and deletion on attribute and instance basis, class attribute remapping, etc.)
• Multithreaded operations on multiple classifiers (training, classification, validation)
• Ability to export a trained classifier for later use in different environment
• Grid-Creation of classifiers to easily find the best model parameters for the given training set
• Various key figures to ease model comparison / selection / evaluation:
• Confusion matrix (absolute values / percentages)
• Actual classification results with confidences
• Various statistics
• Possibility to export evaluation results as model documentation
• Customizable window layout

See a screenshot here.