The Application In The CRM Based On Immune Genetic Algorithm With Data Mining

In recent years, the rapid development of information industry, make people’s life rhythm faster and faster, we can get the information we need anytime, anywhere. In many areas, the data are growing rapidly in different forms, through a database search or query the resulting data has been unable to meet the results we need. How dig out the potential relationships based on the data in the database, and make critical decisions for management, and make the enterprises gain greater benefits. So it is very important to use the data mining technology in practical significance.Classification is one of the major tasks of data mining. There are many ways classification. Decision tree classification algorithm which is commonly used in data classification and prediction tools. C4.5decision tree is the most common method. The algorithm has low computational complexity and the high classification accuracy. However, The decision tree has been built by using the strategy that decreases its useless branches after it is built; so it is difficult to adjust the tree’s structure and content, and difficult to improve the performance of the decision tree. So it is easy to fall into the local optimal solution. Genetic algorithms are a random search algorithm which it simulate the laws of evolution and the evolution of the biosphere. It has strong global optimization capabilities to search for the optimal solution. Genetic algorithm fitness function to evaluate individual solutions, when an individual’s fitness value is large, the individual’s genes can spread rapidly in the population. leading to premature loss of population diversity solutions. so the algorithm premature convergence, it is often said that "premature" phenomenon. Immune genetic algorithm is introduced to the principles of biological immunology based on standard genetic algorithm, using the antibody concentration and affinity to select individuals, thereby increasing the diversity of individual choice, in order to overcome the problem of premature genetic algorithm to improve the global search capability.How to make use of the immune genetic algorithm optimization better C4.5classification algorithm, complementary advantages, improve the classification accuracy, is the purpose of this study. First of all, the use of building decision tree C4.5algorithm, using the fuzzy classification rules says each decision tree branches formed a complete set of classification rules. Again, using the immune genetic algorithm based on density and accessibility choose excellent individual output, on the basis of the formation of the new classification rules. Finally, when the population evolution algebra to specify termination of algebra, immune genetic algorithm to stop. Compared with the C4.5algorithm, the algorithm effectively improves the decision tree classification accuracy. Finally, the improved algorithm is applied to a bank customer relationship management, risk level assessment with the customers, gained good classification effect.