The Empirica Signal application offers two options for grouping values, such as a list of drugs or a list of events, together for use in a data mining run without affecting source data. These features are:
· Custom Terms—see Define custom terms.
· Mapped Text Values—see Map text values.
While these options are similar, each has different advantages and limitations. The differences between these features are as follows:
Options |
Custom Terms |
Mapped Text Values |
MGPS runs can include: |
X |
X |
Logistic regression runs can include: |
X |
|
Set up using query logic: |
X |
|
Set up once, then reusable: (Manage Configurations permission) |
|
X |
Set up for each run: (Create Data Mining Run permission) |
X |
|
Can be set up for variables configured to use a hierarchy table: |
|
X |
Can be set up for variables configured to use a hierarchy account: |
X |
X |
When a custom term is added to a data mining run, that custom term cannot include a variable with mapped text values. Oracle recommends that when mapping text values, instead of applying this data transformation directly to a configuration variable, set up a new configuration variable based on the same table and column specifically to incorporate the mapping. When two different variables are offered in this way, the values of the original variable remain available for inclusion in a custom term. See About data transformations.