Association Rules Startup Panel - Advanced Tab

Select the Advanced tab of the Association Rules Startup Panel to access options to determine various parameters that will guide the a priori algorithm for finding association rules, and determine the minimum confidence and maximum complexity of the rules that will be extracted; refer also to Computational Procedures and Terminology for details regarding these parameters and how they affect the computations.

Minimum support

Enter the minimum support for any association rule to be included in the final results. Given the general form of an association rule as if Body then Head (e.g., If (Car=Porsche and Age<20) then (Risk=High and Insurance=High)), the Support value is computed as the joint probability (relative frequency of co-occurrence) of the Body and Head of each association rule. The search algorithm will terminate when no additional rules can be extracted that meet the Minimum support condition (or when any of the other stopping criteria is satisfied). For additional details, see also Initial Pass Through the Data: The Support Value in Computational Procedures and Terminology.

Minimum confidence

Enter the minimum confidence (conditional probability) for any association rule to be included in the final results. Given the general form of an association rule as if Body then Head (e.g., If (Car=Porsche and Age<20) then (Risk=High and Insurance=High)), the confidence value denotes the conditional probability of the Head of the association rule, given the Body of the association rule. The search algorithm will terminate when no additional rules can be extracted that meet the Minimum confidence condition (or when any of the other stopping criteria is satisfied). For additional details, see also Second Pass Through the Data: The Confidence Value; Correlation Value in Computational Procedures and Terminology.

Minimum correlation

Enter the minimum correlation value for any association rule to be included in the final results. Given the general form of an association rule as if Body then Head (e.g., If (Car=Porsche and Age<20) then (Risk=High and Insurance=High)), the correlation value for an association rule is computed as the support value for the rule, divided by the square root of the product of the support values for the Body and Head computed separately. The search algorithm will terminate when no additional rules can be extracted that meet the Minimum correlation condition (or when any of the other stopping criteria is satisfied). For additional details, see also Second Pass Through the Data: The Confidence Value; Correlation Value in Computational Procedures and Terminology.

Maximum itemset size in body

Enter the maximum number of items (codes, text values) for the Body portion of an association rule. Given the general form of an association rule as if Body then Head (e.g., If (Car=Porsche and Age<20) then (Risk=High and Insurance=High)), the search algorithm will terminate when no additional rules can be extracted with fewer items in the Body of the rule that satisfy the Maximum itemset size in body condition (or when any of the other stopping criteria is satisfied). For additional details, see also Subsequent Passes Through The Data: Maximum Item Size in Body, Head in Computational Procedures and Terminology.

Maximum itemset size in head

Enter the maximum number of items (codes, text values) for the Head portion of an association rule. Given the general form of an association rule as if Body then Head (e.g., If (Car=Porsche and Age<20) then (Risk=High and Insurance=High)), the search algorithm will terminate when no additional rules can be extracted with fewer items in the Head of the rule that satisfy the Maximum itemset size in head condition (or when any of the other stopping criteria is satisfied). For additional details, see also Subsequent Passes Through The Data: Maximum Item Size in Body, Head in Computational Procedures and Terminology.