XML data mining, Part 2: Mining XML association rules

Discover association rules from static and dynamic XML documents

Return to article

Text version of calculation of new support and confidence for an effective association rule (Figure 10)
Let X <- walkman and Y <- mobile phone
A <- 5     'transactions containing X and Y before changes
B <- 15    'transactions containing X before changes
A' <- AB' <- B
From the set of changes C1:
        Y deleted => A' = A' - 1 = 4 and B' <- B'
From the set of changes C2:
        No change applicable to rule Re
From the set of changes C3:
        Y deleted => A' = A' - 1 = 3 and B' <- B'

New_sup(Re) =  A' / D' = 3 / 24 = 12.5%
New_conf(Re) =  A' / B' = 3 / 15 = 20%

Return to article