Database Reference
In-Depth Information
a per city or state level can improve querying eciency substantially.
Also, such information can help decide partitioning strategies for dis-
tributed applications (e.g., in a map-reduce setting). Clearly, incorpo-
rating domain knowledge in string similarity searching is a non-trivial
task. Furthermore, a large number of applications depend heavily on
ecient solutions to the approximate string matching problem as a
primitive operation. Examples include entity resolution, record linkage,
data cleaning, and deduplication [12, 21, 28, 27] data integration [35],
text analytics and more. Every application has unique properties and
presents idiosyncrasies that necessitate the development of specialized
solutions. It is our belief that approximate string matching will remain
of fundamental importance as new application domains and problems
in text data management arise, but a solid foundation will guarantee
robust solutions in the future.