Databases Reference
In-Depth Information
Appendix A
VisMiner Reference by Task
Dataset Preparation
Handle missing values
VisMiner does not support mining of datasets with any missing values. They are
represented in the Control Center by a dataset icon overlaid with a red triangle.
Before proceeding with any data exploration or mining tasks all missing values
must be handled. To do this:
1. Right click on dataset containing missing values.
2. Select “Handle missing values”.
3. Specify handling option for all attributes with missing values.
Outlier detection and isolation
The search for outliers in a dataset takes time, but is facilitated by the data
exploration viewers of VisMiner. For a thorough search, complete all of the
following:
1. View Summary Statistics for dataset. (Right-click on dataset in Control
Center.)
a. Look at minimum and maximum values of each numeric attribute;
check that they fall within acceptable range. For example, an attribute
GradePointAverage should not have a minimum value less than zero or
a maximum value greater than four.
