DEVELOPMENT OF A CLASSIFICATION MODEL BASED ON AN ARTIFICIAL IMMUNITY SYSTEM ALGORITHM AND EVALUATION OF ITS EFFICIENCY

In this paper, the possibilities of using artificial immune system algorithms in classification tasks are studied and analyzed. During the analysis process, the fundamental principles of the biological immune system, namely clonal selection, mutation, memory cell formation, and self-adaptation mechanisms, were modeled in the form of an artificial model. The article also develops a classification model based on the algorithm of artificial immune systems, which analyzes antibody-antigen relationships through affinity function. During clonal selection, the most effective antibodies are preserved, and the class boundaries are formed more accurately using their mutated copies. A classification algorithm based on artificial immune systems was tested on a standard dataset. To confirm the reliability of the results, the developed algorithm was compared with the results obtained from other popular classical classification algorithms, namely KNN, Naive Bayes, and SVM, and its effectiveness was evaluated.

Hali tarjima qilinmagan