Automated abstraction of language concepts

In this paper, we describe the notion of an abstraction and conceptualization of information obtained by symbolization of some real world element that has meaning to humans. Such abstracted concept is able to recognize already observed information and also to recognize similar ones. Similarities identification is based on repetition in structure of information symbolized as regular language strings. Approaches of repetition identification, longest repeating non-overlapping subsequence and longest common subsequence are discussed and applied to show experimental results.