Application of hybrid ant colony algorithm for mining maximum frequent item sets

In order to avoid the effect of initial pheromones for Ant Colony Algorithm, Ant Colony Algorithm and Genetic Algorithm are integrated to mine maximum frequent item sets in this paper. First of all, this paper introduces the ideas of Hybrid Ant Colony Algorithm and briefly explains how to process the data of database. In the second place, Ant Colony Algorithm and Genetic Algorithm are designed respectively, including the method of encoding, parameters setting, the choice of evaluation function and so on. Finally, the Hybrid Ant Colony Algorithm is compared with Max-Min Ant Colony Algorithm (MMAS) on the issue of mining maximum frequent item sets. The results show that the qualities of maximum frequent item sets mined by Hybrid Ant Colony Algorithm are better than MMAS.

[1]  Song Yu An Algorithm and Its Updating Algorithm Based on FP-Tree for Mining Maximum Frequent Itemsets , 2003 .

[2]  Chen Chen,et al.  MMFI: An Effective Algorithm for Mining Maximal Frequent Itemsets , 2008, 2008 International Symposiums on Information Processing.

[3]  Roberto J. Bayardo,et al.  Efficiently mining long patterns from databases , 1998, SIGMOD '98.

[4]  Haifeng Li,et al.  A GPU-based maximal frequent itemsets mining algorithm over stream , 2010, 2010 International Conference on Computer and Communication Technologies in Agriculture Engineering.