Hierarchy graphs

Hierarchy graphs highlight situations in which there are interaction effects (such as drug interactions or event syndromes), where the score for an interaction or syndrome is not predicted by the scores of its components.

The score for an interaction is INTSS, which is calculated as the EB05 score for the combined drugs or events divided by the highest EB95 score found for an individual component drug or event. The graph is built within the context provided by the drug or event specified on the Select Criteria page. A hierarchy graph is available for three- dimensional results of MGPS data mining runs.

The key below the graph shows the statistical ranges that determine the cell colors. The ranges are for EBGM and INTSS values. On the Choose Graph page, you can modify the graph key (that is, the cutpoints and colors).

If Notes is checked when you display the graph, a Notes section provides information about the selection criteria and display options used for the graph.

The following example is for a three-dimensional data mining run, with the drug Tilidine selected. The scores for the two-way interactions between that drug and a single event appear in the left column. The scores for three-way interactions, between the drug and two events, appear in the right column.