Artificial Immune Algorithm Research And Its Application In Data Mining

Artificial Immune is an increasingly important area in the field of computational technology. Biological Immune is highly complexity system with capability in learning , memory acquisition , pattern recognition and so on. Simulating biological immune system, the new computational techniques can solve not only the science but also the engineering problems. Based on the Immune principles , this paper at first analyse the immune algorithm by solving TSP problem which improves the ability of population and increase the holistic performance. Presenting several data mining algorithm Based on the theory, this paper mainly includes the following two aspects.Firstly, a new frequent itemsets discovering based immune algorithm is proposed. Tradition algorithm attempted to improve the mining efficiency reducing the number of database passes to control the I/O cost , which at first derives candidate itemsets from tuples in database, and count support of candidate itemsets to get frequent itemsets. Changing the former way of discovering frequent itemsets, the paper understand the definition of frequent itemsets from another standpoint. Making use of artificial immune strategy , a new algorithm is presented. In the paper, the algorithm is introduced in detail. The cost of new algorithm is related to the support and the total number of database. Finally, simulation shows the correctness and validity compared with other algorithm.In practice, using Immune Algorithm , the paper builds the model of data mining in finance 's feature information cards. Its key is to construct feature-relational degree function as the affinity between antibody and antigen. The stimulation of antibodies helps convergence of solution. The paper introduced in detail the designs . The result of experiment testify the correctness and validity of this new algorithm. Compared with the other algorithms.