Mining frequent web access patterns with partial enumeration
Peiyi Tang, Markus P. Turkia · 2007
In this paper, we extend the pattern-growth web access pattern mining algorithms [1, 2, 3] with partial enumeration. The extended algorithm can grow the frequent patterns with more than one symbol at a time and unifies the pattern-growth and apriori algorithms [4]. The experimental results show that for the databases of long sequences, the best performance is neither given by the pattern-growth algorithms nor by the full apriori enumeration algorithms, but rather by the mining with partial enumeration in the middle.