Abstract
We present an algorithm that matches queries to a set of advertising bid terms, including lexical variants. We use an enhanced digital tree as index that allows to match misspelled words, word segmentation and splitting, and some stems. The results, both in quality and performance, are competitive and improved revenue in several percentage points. To the best of our knowledge, this is the first pure lexical matching algorithm for sponsored search.
Cite
CITATION STYLE
Yates, R. B., & Wang, G. (2016). Lexical matching of queries and ads bid terms in sponsored search. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9954 LNCS, pp. 231–239). Springer Verlag. https://doi.org/10.1007/978-3-319-46049-9_22
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
