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