Recommender system model based on artificial immune system

Recommendation and prediction problems mostly rely on recognition and classification tasks. Artificial immune systems, based on natural immunological principles, are computational paradigm for solving such tasks. Additional context dependent response theories like Danger theory explain usage of signaling in recognition process. Recommender system model proposed in this paper addresses construction of a Web portal news article recommender based on artificial immune system combined with Danger theory. System knowledge represents learned user preferences using implicit tracking of user actions. System also adapts to evolution of user´s opinion and expresses results in personalized recommendation list format