TY  - CONF
ID  - www20093
UR  - http://www2009.eprints.org/3/
A1  - Agarwal, Deepak
A1  - Chen, Bee-Chung
A1  - Elango, Pradheep
Y1  - 2009/04//
N2  - We propose novel spatio-temporal models to estimate clickthrough rates in the context of content recommendation. We track article CTR at a ?xed location over time through a dynamic Gamma-Poisson model and combine information from correlated locations through dynamic linear regressions, signi?cantly improving on per-location model. Our models adjust for user fatigue through an exponential tilt to the ?rstview CTR (probability of click on ?rst article exposure) that is based only on user-speci?c repeat-exposure features. We illustrate our approach on data obtained from a module (Today Module) published regularly on Yahoo! Front Page and demonstrate signi?cant improvement over commonly used baseline methods. Large scale simulation experiments to study the performance of our models under different scenarios provide encouraging results. Throughout, all modeling assumptions are validated via rigorous exploratory data analysis.
TI  - Spatio-Temporal Models for Estimating Click-through Rate
SP  - 21
M2  - Madrid, Spain
AV  - public
EP  - 21
T2  - 18th International World Wide Web Conference
ER  -