An Approximate String Matching Algorithm Based upon the Candidate Elimination Method

In this paper, we consider the approximate string matching problem. We give a method to eliminate candidate locations in text T as there can be no substring S starting from those locations such that the edit distance between S and pattern P is smaller than or equal to a specified error bound k . Our method is simple to implement. Experimental results show that our method is effective, especially in the case when we perform natural language searching. For instance, for a DNA type data, when the text is 10000 characters long, patterns are 40 characters long and 3 = k , in average, 73% of locations are eliminated by our method. As for a English text with 4500 characters, for English patterns with 20 characters, in average, 99.9% of locations are eliminated if 1 = k and 73.3% for 6 = k .