TY - GEN
T1 - User-aware advertisability
AU - Yu, Hai Tao
AU - Sakai, Tetsuya
N1 - Copyright:
Copyright 2014 Elsevier B.V., All rights reserved.
PY - 2013
Y1 - 2013
N2 - In sponsored search, many studies focus on finding the most relevant advertisements (ads) and their optimal ranking for a submitted query. Determining whether it is suitable to show ads has received less attention. In this paper, we introduce the concept of user-aware advertisability, which refers to the probability of ad-click on sponsored ads when a specific user submits a query. When computing the advertisability for a given query-user pair, we first classify the clicked web pages based on a pre-defined category hierarchy and use the aggregated topical categories of clicked web pages to represent user preference. Taking user preference into account, we then compute the ad-click probability for this query-user pair. Compared with existing methods, the experimental results show that user preference is of great value for generating user-specific advertisability. In particular, our approach that computes advertisability per query-user pair outperforms the two state-of-the-art methods that compute advertisability per query in terms of a variant of the normalized Discounted Cumulative Gain metric.
AB - In sponsored search, many studies focus on finding the most relevant advertisements (ads) and their optimal ranking for a submitted query. Determining whether it is suitable to show ads has received less attention. In this paper, we introduce the concept of user-aware advertisability, which refers to the probability of ad-click on sponsored ads when a specific user submits a query. When computing the advertisability for a given query-user pair, we first classify the clicked web pages based on a pre-defined category hierarchy and use the aggregated topical categories of clicked web pages to represent user preference. Taking user preference into account, we then compute the ad-click probability for this query-user pair. Compared with existing methods, the experimental results show that user preference is of great value for generating user-specific advertisability. In particular, our approach that computes advertisability per query-user pair outperforms the two state-of-the-art methods that compute advertisability per query in terms of a variant of the normalized Discounted Cumulative Gain metric.
UR - http://www.scopus.com/inward/record.url?scp=84893319629&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84893319629&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-45068-6_39
DO - 10.1007/978-3-642-45068-6_39
M3 - Conference contribution
AN - SCOPUS:84893319629
SN - 9783642450679
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 452
EP - 463
BT - Information Retrieval Technology - 9th Asia Information Retrieval Societies Conference, AIRS 2013, Proceedings
T2 - 9th Asia Information Retrieval Societies Conference on Information Retrieval Technology, AIRS 2013
Y2 - 9 December 2013 through 11 December 2013
ER -