Selecting data mining results criteria

To view data mining results, you specify criteria such as the drugs and events to include in the table or graph of results on the Select Criteria page.

1.         Click the Data Mining Results tab.

The Select Criteria page appears.

2.         From the Run name drop-down list, select a data mining run, or click Browse to view all runs and select a run.

3.         To view details about the selected run, click View Run Details. Click Close to close the detail dialog box for the run.

4.         Specify one or more drugs and one or more events. Only exact matches, including the case (upper, lower, or mixed), are found in the run results.

5.         To specify item variable values, use one or more of the following options:

• If a variable has an associated hierarchy of terms and your user preference enables use of the hierarchy, you can click Select <hierarchy-name> Terms and select terms from the hierarchy. Custom terms used in the run are not in the hierarchy.

If the run option to store the primary paths of terms in run results was used, you can search for terms at various levels of the hierarchy. Click the radio button for the level at which you want to search. If you select terms at a higher level than the level of the term used in the run, you see results only for terms that have that higher level in their primary path.

• To select values from a list, click Select Available Values. The available values include custom terms used in the run.

• To use a saved list of terms, click Select Saved List. A saved list can include custom terms.

• To type the terms, review the rules in Typing values in text boxes.

Note: To prevent spelling and capitalization errors, Oracle recommends that you select rather than type values.

Note: Drugs or events that have a missing (null) value in the source data have the value of the user preference, Replace missing value with, that was in effect for the user who created the run. (Your own user preference does not affect run results generated by another user.)

• To look up a drug name, click Trade/Generic Lookup. You can find the generic name for a drug (if you only know its trade name) or vice versa. (This link is available only if the mapping of trade/generic names has been defined as part of preparing source data.)

6.         To include only scores that exceed a minimum threshold value, use the Limit to field to select a statistic and supply that minimum number. For a description of the statistics that you can use to limit results, see Data mining results for MGPS runs or Data mining results for logistic regression runs.

For example, to review only MGPS run results with an EB05 score greater than 2, select EB05 and supply 2. The application uses the setting as the default (for results that you have not yet viewed) until you change it.

Notes:

If you provide a limit for an MGPS run and then view a graph, the application applies the limit to the highest dimension in the graph only. For a sector map graph, the limit is ignored.

When you view one-dimensional run results for an MGPS run,the application ignores this limit unless you specify a limit for the N statistic.

This limit does not apply to sector map graphs or to the tables of covariate and interaction results for a logistic regression run.

To supply defaults for this statistic and value the first time you look at results for a run, you can set the user preference, Display Only Results with <score-type> Scores Over <number>. If you select a statistic other than N, a default of LROR > 0 is used for logistic regression runs.

7.         To select additional criteria, click Show Advanced. This link is available if additional fields can be used to limit the results to display. (If you do not generally use these fields you can click Hide Advanced.)

1. For an MGPS run In the Dimension field, select a dimension to limit the results that appear. You can select 1, up to the highest dimension included in the run, or All to display results for all of the dimensions included in the run. By default, the dimension is the highest dimension that was selected when the data mining run was created. You can change the dimension to less than the default. The number of Pattern fields that appear depends on the number in the Dimension field.
2. The Subset field appears if the run used subsetting and at least two subset categories were specified for the run. Select the subset for which you want to view results, or select All to display results for all subsets.
3. In the Pattern fields, specify a pattern such as Drug+Event. See Dimensions and patterns.
4. If your user preference, Include SQL WHERE Clause for Advanced Results Selection, is checked, you can specify a SQL WHERE clause to restrict which run results you view. You can reference any of the columns that can be displayed in the results table, regardless of whether they are currently included in the display. To view the complete list of columns, click the Columns link above the results table. The SQL WHERE clause entered here is ignored if you view a table of covariates or interactions for a logistic regression run, or when you view a sector map graph.

8.         If you want to remove your selections from the fields on this page, click Clear All.

9.         Click View Results Table to view a results table or click Choose Graph to choose a graph type to display. To view a graph, you must specify at least one drug or one event. For MGPS runs, you can view results graphically for two-, or three- dimensional drug-event combinations.

Note: When you view results for a logistic regression run, all criteria defined on this page are used to select the combinations that appear in the results table. However, only the drug and event values specified, if any, are used to define the results that appear in the Covariates Table and Interactions Table.