Overview of fingerprinting methods for local text reuse detection

Abstract

We overview several local text reuse detection methods based on fingerprinting techniques. We first define the context of local text reuse and situate it within the general spectrum of information retrieval in order to pinpoint its particular applicability and challenges. After a brief description of the major text reuse detection approaches, we introduce the general principles of fingerprinting algorithms from an information retrieval perspective. Three classes of fingerprinting methods (overlap, non-overlap, and randomized) are surveyed. Specific algorithms, such as k-gram, winnowing, hailstorm, DCT and hash-breaking, are described. The performance and characteristics of these algorithms are summarized based on data from the literature.