Generalization of String Transformation Rules using Optimized Concept Lattice Construction Method

The sequences play an important role on many application areas. The strings as symbol sequences are investigated mainly in natural language processing and biological modeling. This paper investigates the transformation process of the strings and it proposes a method for induction of the string transformation rules. The presented method is based on the application of an FCA lattice for generalization of the transformation rules. The paper introduces an efficient construction algorithm for building the conceptualization lattice. Based on the performed comparison tests, the proposed method outperforms the main incremental methods and for some parameter ranges, its efficiency is comparable with the efficiency of the best batch methods.