Approximate regular expression searching with arbitrary integer weights (2002) [4 citations — 2 self]
Abstract:
Abstract. We present a bit-parallel technique to search a text of length n for a regular expression of m symbols permitting k dierences in worst case time O(mn = log k s), where s is the amount of main memory that can be allocated. The algorithm permits arbitrary integer weights and matches the best previous complexities, but it is much simpler and faster in practice. In our way, we dene a new recurrence for approximate searching where the current values depend only on previous values. 1 Introduction and Related Work The need to search for regular expressions arises in many text-based applications, such as text retrieval, text editing and computational biology, to name a few. A

