Fast multiple keyword searching

Abstract

A new multiple keyword searching algorithm is presented as a generalization of a fast substring matching algorithm based on an n-gram technique. The expected searching time complexity is shown to be O((N/m+ml) log lm) under reasonable assumptions about the keywords together with the assumption that the text is drawn from a stationary ergodic source, where N is the text size, l number of keywords and m the smallest keyword size.