An Efficient Mapping for Score of String Matching
Tetsuya Nakatoh, Kensuke Baba, Daisuke Ikeda, Yasuhiro Yamada, Sachio Hirokawa · QIR (Kyushu University Institutional Repository) (Kyushu University) · 2003
This paper proposes an efficient algorithm to solve the problem of string matching with mismatches. For a text of length n and a pattern of length m over an alphabet Σ, the problem is known to be solved in O(|Σnlogm|)time by computing a score by the fast Fourier transformation (FFT). Atallah et al. introduced a randomized algorithm in which the time complexity can be decreased by the trade-off with the accuracy of the estimates for the score. The algorithm in the present paper yields an estimate with smaller variance compared to that the algorithm by Atallah et al., moreover, and computes the exact score in O(|Σ|nlogm)time. The present paper also gives two methods to improve the algorithm and an exact estimation of the variance of the estimates for the score. Keywords: string matching with mismatches, FFT, convolution, deterministic algorithm, randomized algorithm.