Fantail MLKit

Fantail ML Kit

Fantail is a collection of machine learning algorithms for ranking prediction, multi-target regression, label ranking and metalearning related data mining tasks. The algorithms can be called from your own Java code. It is also well-suited for developing new algorithms. Fantail is a multi-target learning extension to WEKA, and is at the early development stage. New algorithms and tools will be added to the library gradually, please send me an Email if you would like to join the team :-)

A key difference between Fantail and another popular preference learning package WEKA-LR is: Fantail uses the rank vector format (similar to the multi-target regression setting) rather than the order/explicit preference vector format. So in Fantail, label ranking is treated as a special case of the multi-target regression problem. The advantage of the Fantail approach is that both multi-target and label ranking algorithms can be used and tested under a unified framework.

Potential applications:
Survey/rank data modelling and analysis, metalearning, sports analytics, recommender systems, learning to rank



See for an example

Benchmark Datasets

iris_.arff (an example dataset showing the data format used by Fantail)
A collection of 26 label ranking datasets can be downloaded

Algorithms (8)


Evaluation Metrics


Citing Fantail

If you want to refer to Fantail in a publication, please cite the following paper:
Quan Sun and Bernhard Pfahringer. Pairwise Meta-Rules for Better Meta-Learning-Based Algorithm Ranking. Machine Learning, 93(1):141-161, Springer US, 2013, DOI: 10.1007/s10994-013-5387-y


Many TODOs, so please send me an Email if you would like to join the team :-)
Quan Sun