abstract

• In this paper, a novel algorithm for approximate string matching (ASM) is proposed. The novelty resides in the fact that, unlike most other methods, the proposed algorithm is not based on the Hamming or Levenshtein distances, but instead computes a score for each symbol in the search text based on a consensus measure. Those symbols with sufficiently high scores will likely correspond to approximate instances of the pattern string. To demonstrate the usefulness of the proposed method, it has been applied to the detection of QRS complexes in electrocardiographic signals with competitive results when compared against the classic Pan-Tompkins (PT) algorithm. The proposed method outperformed PT in 72%25 of the test cases, with no extra computational cost. © 2016 World Scientific Publishing Company.

authors

• Alba Cadena Francisco Alfonso
• Arce Santana Edgar Román
• Méndez García Martin Oswaldo
• Rubio-Rincón M.

publication date

• 2016-01-01

published in

• International Journal of Modern Physics C  Journal

keywords

• Approximate string matching; consensus measure; pattern recognition; QRS complex

Digital Object Identifier (DOI)

• 10.1142/S0129183116500297

start page

end page

volume

• 27

issue

• 3