Searching for a modified pattern in a changing text

Amihood Amir, Eitan Kondratovsky

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

Much attention has been devoted recently to the dynamic model of pattern matching. In this model the input is updated or changed locally. One is interested in obtaining the appropriate search result in time that is shorter than the time necessary to search without any previous knowledge. In particular, searching for a pattern P in an indexed text is done in optimal O(|P|) time. There has been work done in searching for a fixed pattern in a dynamic text, as well as finding all maximum common substrings of two strings in a dynamic setting. There are real-world applications where the text is unchanged and the pattern is slightly modified at every query. However, in the current state-of-the-art, a new search is required if the pattern is modified. In this paper we present an algorithm that reduces this search time to be sublinear, for a variety of types of pattern modification - in addition to the insertion, deletion, and replacement of symbols, we allow copy-paste and delete substring operations. We also make a step toward a fully dynamic pattern matching model by also supporting text changes, albeit much more modest than the pattern changes. We support dynamic pattern matching where symbols may also be either added or deleted to either the beginning or the end of the text. We show that we can support such a model in time O(log n) for every pattern modification or text change. We can then report all occ occurrences of P in the text in O(occ) time.

Original languageEnglish
Title of host publicationString Processing and Information Retrieval - 25th International Symposium, SPIRE 2018, Proceedings
EditorsTravis Gagie, Alistair Moffat, Gonzalo Navarro, Ernesto Cuadros-Vargas
PublisherSpringer Verlag
Pages241-253
Number of pages13
ISBN (Print)9783030004781
DOIs
StatePublished - 2018
Event25th International Symposium on String Processing and Information Retrieval, SPIRE 2018 - Lima, Peru
Duration: 9 Oct 201811 Oct 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11147 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference25th International Symposium on String Processing and Information Retrieval, SPIRE 2018
Country/TerritoryPeru
CityLima
Period9/10/1811/10/18

Bibliographical note

Publisher Copyright:
© Springer Nature Switzerland AG 2018.

Fingerprint

Dive into the research topics of 'Searching for a modified pattern in a changing text'. Together they form a unique fingerprint.

Cite this