The columns of the RESULTS_EUROVIS2015.csv data describe the following information:

* fileName: filename of the dataset

* indexFile: unique identifier of the dataset

* indexDR: Dimension reception technique that was originally used to project this data (1:PCA, 2:RobPCA)

* classNum: shows which class was used as target class (target class is encoded as 1, all others as 0 in the data)

* scoreA: 5-point score of the first human coder: 5 is best separability, 1 worst separability
* scoreM: 5-point score of the second human coder: 5 is best separability, 1 worst separability

* scoreDes: aggregated score: (scoreA:4,scoreM:4),(4,5),(5,4),(5,5) -> 1 (separable) / (1,1),(1,2),(2,1),(2,2) -> 0 (non-separable)

* columns 8-42:  show the values of 15 current state-of-the-art measures and their parameterizations. 

All measures are linearly scaled, so that "100" is supposed to be best class separability and "0" worst separability.

The last row (829) shows the AUCBA scores for these measures, as a summary statistics for each measure.