Based On Artificial Immune Model Of Data Mining And Its Applications

Biology immune system is a highly evolved biochemistry system. It aims at telling the external baneful antigens apart the self organizations, which can clean up the antigens and stabilize the organism. From the angel of the computation, biology immune system is a highly synchronic, distributional, self-adaptive and self-organizational system. It has the strong ability of study, recognition, memorization and character extraction. People hope to be enlightened from the mechanism of the immune system. So they develop the computational models of the immune system facing the application——Artificial Immune System.The artificial immune system is an intelligent method for imitating the function of the natural immune system. It has the study ability enlightened from the biology immune system that studies the natural defense mechanism of the external materials. It offers the study mechanism including noise tolerance, self-teaching, self-organization and memory. Besides it takes the advantages of the classification machine, the neural network and the machine reasoning. So that it has the potential of solving some new problems.In the last few years, people have learned more and more immune mechanisms with the research of the immunology. In order to acquire the excellent system going with the same capabilities, they begin to implement the immune system in the engineering field for improving the old ways or developing the new ones. Nowadays, the immune system has been accepted by more and more research fields, such as intelligent control, computer safe and defense against viruses, abnormity and intrusion detection, knowledge discovery and optimization.The artificial immune system is still at it's beginning, although it has the strong potential. Compared with the genetic algorithms, the artificial immune system is hardly utilized in data mining because of the limit of its development. It is mostly used in the function optimization, engineering control and intrusion detection. On the other hand, the artificial immune system in data mining mainly adopts the basic immune algorithms, which ignore the unique characters of the other immune mechanisms such as the immune networks. These algorithms are mostly at the experimental phases. It's to say that they aim at proving the feasibility via some small databases but not the applications on marvel ones.Much attention has been paid to the relevant problems of the artificial immune system as well as its application in data mining in this paper. It brings the advantages of the artificial immune network models in data mining into play, due to its special immune mechanism. And then it is carried out in a marvel database. The main work is as follows:1. Discuss and study some artificial immune models Three representative ones are introduced: aiNet model, bone marrow model andresource limited artificial immune system. aiNet model is focused on.2. Propose two cluster algorithms based on the aiNet modelThe cluster technology based on the artificial immune network models are discussed and two new ones based on the aiNet model are put forward——the aiFCM algorithm and the aiNHA algorithm. The two algorithms are separately proved efficiently through the emulative experiments; therefore the utilization of the artificial immune network models on the knowledge discovery is realized.3. Carry out the cluster technology based on the artificial immune network models on the database of the hotel foreground managementThe database system facing the hotel foreground management is constructed with SQL Server 2000 and the cluster technology based on the artificial immune model is put into use. The design result is quite satisfying, which shows that the immune algorithms behaves well when facing the marvel databases because of its abilities of efficient search, study memory and searching the best systematically.