This paper discovers the major shortcomings of the Levenshtein Distance method, the longest common subsequence (LCS) method, and other general approaches to finding common parts, including the unjustified fragmentation of selected parts, the lack of sensitivity to transposition of large blocks, and no mechanisms to prevent accidental matches. The belief function theory leads to a flexible framework for similarity evaluation.The framework is aimed on new similarity models which are free of described shortcomings and can be effectively calculated. A sketch of better sequence alignment algorithm illustrates the framework’s utility.
CITATION STYLE
Znamenskij, S. (2015). A belief framework for similarity evaluation of textual or structured data. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9371, pp. 138–149). Springer Verlag. https://doi.org/10.1007/978-3-319-25087-8_13
Mendeley helps you to discover research relevant for your work.