Viewing a Mining Model with the Microsoft Association Rules Viewer

The Microsoft Association Rules Viewer in Microsoft SQL Server 2005 Analysis Services (SSAS) displays mining models that are built with the Microsoft Association algorithm. The Microsoft Association algorithm is an association algorithm for use in creating data mining models that you can use for market basket analysis. For more information about this algorithm, see Microsoft Association Algorithm.

When you browse a mining model in Analysis Services, the model is displayed on the Mining Model Viewer tab of Data Mining Designer, using the appropriate viewer for the model.

Following are the primary reasons for using the Microsoft Association algorithm:

  • To find itemsets that describe items that are typically found together in a transaction.
  • To discover rules that predict the presence of other items in a transaction, based on existing items.

Accordingly, the Microsoft Association Rules Viewer is more textual than the other Microsoft algorithm viewers in Analysis Services.

Viewer Tabs

The Microsoft Association Rules Viewer includes the following tabs:

  • Itemsets
  • Rules
  • Dependency Net

Each tab contains the Show long name check box, which you can use to show or hide the table from which the itemset originates in the rule or itemset.

Itemsets

The Itemsets tab displays the list of itemsets that the model identified as frequently found together. The tab displays a grid with the following columns: Support, Size, and Itemset. For more information about support, see Microsoft Association Algorithm. The Size column displays the number of items in the itemset. The Itemset column displays the actual itemset that the model discovered. You can control the format of the itemset by using the Show list, which you can set to the following options:

  • Show attribute name and value
  • Show attribute value only
  • Show attribute name only

You can filter the number of itemsets that are displayed in the tab by using Minimum support and Minimum itemset size. You can limit the number of displayed itemsets even more by using Filter Itemset and entering an itemset characteristic that must exist. For example, if you type "Water Bottle = existing", you can limit the itemsets to only those that contain a water bottle. The Filter Itemset option also displays a list of the filters that you have used previously.

You can sort the rows in the grid by clicking a column heading.

Back to Top

Rules

The Rules tab displays the rules that the association algorithm discovered. The Rules tab includes a grid that contains the following columns: Probability, Importance, and Rule. The probability describes how likely the result of a rule is to occur. The importance is designed to measure the usefulness of a rule. Although the probability that a rule will occur may be high, the usefulness of the rule may in itself be unimportant. The importance column addresses this. For example, if every itemset contains a specific state of an attribute, a rule that predicts state is trivial, even though the probability is very high. The greater the importance, the more important the rule is.

You can use Minimum probability and Minimum importance to filter the rules, similar to the filtering you can do on the Itemsets tab. You can also use Filter Rule to filter a rule based on the states of the attributes that it contains.

You can sort the rows in the grid by clicking a column heading.

Back to Top

Dependency Net

The Dependency Net tab includes a dependency network viewer. Each node in the viewer represents an item, such as "state = WA". The arrow between nodes represents the association between items. The direction of the arrow dictates the association between the items according to the rules that the algorithm discovered. For example, if the viewer contains three items, A, B, and C, and C is predicted by A and B, if you select node C, two arrows point toward node C — A to C and B to C.

The slider at the left of the viewer acts as a filter that is tied to the probability of the rules. Lowering the slider shows only the strongest links.

Back to Top

See Also

Concepts

Microsoft Association Algorithm
Using the Data Mining Tools

Other Resources

Mining Model Viewer Tab How-to Topics
Mining Model Viewer Tab How-to Topics
Viewing a Data Mining Model

Help and Information

Getting SQL Server 2005 Assistance