Predicting the cellular localization sites of protein patterns with k-NN classification algorithm based artificial immune system

With the development of technology, for solving the problems we face in the real life, the concern for the works being done on biologic systems is increasing year by year. Artificial neural networks, evolutionary calculation and artificial immune system are some of these approaches. Artificial immune system is put forward for better understanding the interactions in the immune system, forming a model of immune system and for reckoning the events occurring in the system. In this thesis, a system aimed at classifying the medical data by using artificial immune system and k-NN classification algorithm, is proposed. In the proposed classification structure, with the artificial immune system, the features which characterize the data are selected, and with k-NN, these features which are reduced from data, are classified. For this purpose, experiments are done about the data of protein settlement place of E.coli bacterium which are taken from UCI datum base. Keywords: Artificial immune system, k-NN Algorithm, E.coli, Protein, Pattern Recognition, Classification.