2. Preliminary conceptsLet D be a dataset with d attributes f g A1; A2; ... ; Ad and n denote records (objects), where each record has an object identifier(OID). Let C ¼ f g c1; c2; ... ; ck be a list of class labels (k is the number of classes). Let Constraint_Class be a subset of Ccontaining particular class labels considered by end users. A specific value of an attribute Ai and the mth record is denotedby ami (m 2 ½ 1; n ; i 2 ½ Þ 1; d and a specific value of class C is denoted by cx (x 2 ½ 1; k ).Definition 1. An item is described as an attribute and a specific value for that attribute, denoted by Ai f g ð Þ ; ami(m 2 ½ 1; n ; i 2 ½ 1; d ).Definition 2. An itemset is a set of items, denoted by Ai ð Þ ; ami ; ... ; Aj; amj (m 2 ½ 1; n ; i; j 2 ½ 1; d , and i – j).Definition 3. CAR R has the form Ai ð Þ ; ami ; ... ; Aj; amj ! cx, where Ai ð Þ ; ami ; ... ; Aj; amj is an itemset and cx 2 C is a classlabel.Definition 4. The actual occurrence ActOcc Rð Þ of rule R in D is the number of records of D that match R’s antecedent.Definition 5. The support of rule R, denoted by SupðRÞ, is the number of records of D that match R’s antecedent and R’sconsequent.
đang được dịch, vui lòng đợi..
