What are custom terms?

A custom term is a pseudo-value specified in a data mining run for inclusion in cases that meet specified criteria. When you perform data mining, Oracle Empirica Signal acts as if the custom term is present for those cases and generates statistics for the custom term. Custom terms are defined and used within the context of data mining runs and do not affect the source data.

You can define multiple custom terms in a data mining run. Oracle Empirica Signal adds the suffix, (Custom Term), to the term that you specify.

A typical use of custom terms is to group together similar terms for products or events as a way to prevent the signal dilution that may occur when similar, but not the same, terms are used in reporting. For example, you can create the custom term, Hepatic Terms, to group the following PTs: Coma hepatic, Hepatic encephalopathy, Hepatic failure, Hepatorenal failure, and Hepatorenal syndrome. The data mining run then computes statistics for Hepatic Terms as if it is a PT in the source data, although the custom term is not added to the source data. The data mining run also computes statistics for the constituent PTs.

When estimating the shrinkage parameters for EBGM scores, MGPS does not consider custom terms. The raw RR scores for combinations involving custom terms are shrunk by the Bayesian formula, but they do not participate in the determination of the formula itself. For more information, see MGPS computations.

Note:

For another way of mapping values to a term for use in a data mining run, see Map text values. For information on the differences between these features, see Custom terms and mapped text values.

Custom terms are defined using queries. Reports that match the query condition are considered to have the pseudo value. When defining a new query, you have the option to save it to the Query Library. You can then use it to define custom terms for other data mining runs.

When you re-run a run, the same custom terms are used in the run. However, you cannot switch run types. Thus, if you plan to use a custom term for different run types, make sure that you save the query you define for the custom term for the first run you set up; then you can select that query as the custom term for the other run type(s).