Imbalanced dataset for benchmarking

Imbalanced dataset for benchmarking
=======================

The different algorithms of the `imbalanced-learn` toolbox are evaluated on a set of common dataset, which are more or less balanced. These benchmark have been proposed in [1]. The following section presents the main characteristics of this benchmark.

Characteristics
-------------------

|ID   |Name                   |Repository & Target                |Ratio |# samples| # features |
|:---:|:----------------------:|--------------------------------------|:------:|:-------------:|:--------------:|
|1    |Ecoli                      |UCI, target: imU                     |8.6:1 |336             |7                 |
|2    |Optical Digits        |UCI, target: 8                         |9.1:1 |5,620           |64               |
|3    |SatImage              |UCI, target: 4                          |9.3:1 |6,435           |36              |
|4    |Pen Digits             |UCI, target: 5                          |9.4:1 |10,992         |16              |
|5    |Abalone                |UCI, target: 7                          |9.7:1 |4,177           |8                |
|6    |Sick Euthyroid      |UCI, target: sick euthyroid     |9.8:1 |3,163           |25              |
|7    |Spectrometer       |UCI, target: >=44                   |11:1  |531              |93              |
|8    |Car_Eval_34         |UCI, target: good, v good        |12:1  |1,728           |6                |
|9    |ISOLET                  |UCI, target: A, B                      |12:1  |7,797           |617            |
|10   |US Crime             |UCI, target: >0.65                   |12:1  |1,994          |122             |
|11   |Yeast_ML8            |LIBSVM, target: 8                    |13:1  |2,417          |103             |
|12   |Scene                   |LIBSVM, target: >one label    |13:1  |2,407          |294             |
|13   |Libras Move         |UCI, target: 1                          |14:1  |360             |90               |
|14   |Thyroid Sick        |UCI, target: sick                      |15:1  |3,772          |28               |
|15   |Coil_2000            |KDD, CoIL, target: minority    |16:1  |9,822          |85               |
|16   |Arrhythmia          |UCI, target: 06                        |17:1  |452             |279             |
|17   |Solar Flare M0     |UCI, target: M->0                    |19:1  |1,389         |10                |
|18   |OIL                       |UCI, target: minority               |22:1  |937            |49                |
|19   |Car_Eval_4          |UCI, target: vgood                   |26:1  |1,728         |6                  |
|20   |Wine Quality       |UCI, wine, target: <=4            |26:1  |4,898         |11                |
|21   |Letter Img           |UCI, target: Z                           |26:1  |20,000       |16                |
|22   |Yeast _ME2          |UCI, target: ME2                      |28:1  |1,484         |8                  |
|23   |Webpage             |LIBSVM, w7a, target: minority|33:1  |49,749       |300              |
|24   |Ozone Level        |UCI, ozone, data                      |34:1  |2,536         |72               |
|25   |Mammography   |UCI, target: minority                |42:1  |11,183       |6                 |
|26   |Protein homo.     |KDD CUP 2004, minority         |111:1|145,751     |74               |
|27   |Abalone_19         |UCI, target: 19                         |130:1|4,177         |8                 |

References
----------
[1] Ding, Zejin, "Diversified Ensemble Classifiers for H
ighly Imbalanced Data Learning and their Application in Bioinformatics." Dissertation, Georgia State University, (2011).

[2] Blake, Catherine, and Christopher J. Merz. "UCI Repository of machine learning databases." (1998).

[3] Chang, Chih-Chung, and Chih-Jen Lin. "LIBSVM: a library for support vector machines." ACM Transactions on Intelligent Systems and Technology (TIST) 2.3 (2011): 27.

[4] Caruana, Rich, Thorsten Joachims, and Lars Backstrom. "KDD-Cup 2004: results and analysis." ACM SIGKDD Explorations Newsletter 6.2 (2004): 95-108.