Hill, LaPan, Li and Haney (2007) develop models to predict which cells in a
high content screen were well segmented. The data consists of 119 imaging
measurements on 2019. The original analysis used 1009 for training and 1010
as a test set (see the column called Case).
data frame of cells
The outcome class is contained in a factor variable called Class with
levels "PS" for poorly segmented and "WS" for well segmented.
The raw data used in the paper can be found at the Biomedcentral website.
Versions of caret < 4.98 contained the original data. The version now
contained in segmentationData is modified. First, several discrete
versions of some of the predictors (with the suffix "Status") were removed.
Second, there are several skewed predictors with minimum values of zero
(that would benefit from some transformation, such as the log). A constant
value of 1 was added to these fields: AvgIntenCh2,
FiberAlign2Ch3, FiberAlign2Ch4, SpotFiberCountCh4 and
TotalIntenCh2.
A binary version of the original data is at http://topepo.github.io/caret/segmentationOriginal.RData.