Mining weighted association rules based on weighted Fp tree

摘要

The article presents a new algorithm for mining weighted frequent itemsets without generating candidate, which based on weighted Fp-tree and the weighted model proposed by Feng Tao. For solving the problem that the weighed support may be bigger than 1, the weight set of attributes was normalized. The new algorithm is testified to satisfy weighted downward closure property and an effectively mining pruning strategy base of weighed Fp-tree is structured. Thorough research and scientific analysis, the new algorithm solves the problem that the importance of association rule does not increase with the amount of attribute in practical application.