Efficient multiple-click models in web search

By Fan Guo, Chao Liu, Yi Min Wang. 2009

300 citations recorded by the source corpus.

Subfields: Information Systems, Information Retrieval & Search

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Systems

In Information Retrieval & Search

Citations in the corpus

Contextual Dueling Bandits (2015, methodology): “…setting: as in their setting, we consider a family of K-armed dueling bandit problems arising from the ranker evaluation problem in IR, where the comparisons are carried out using Probabilistic Interleave (Hofmann et al., 2011) and using click models (Guo et al., 2009a,b) to simulate user behavior.”
MergeRUCB: A Method for Large-Scale Online Ranker Evaluation (2015, methodology): “Queries are sampled randomly and clicks are generated probabilistically by conditioning on these assessments using a user model that resembles the behavior of an actual user [12].”
Bandit Algorithms (2020, methodology): “[2016] study the dependent click model introduced by Guo et al. [2009]. This differs from the models proposed in this chapter because the reward is not assumed to be the number of clicks and is actually unobserved. We leave the reader to explore this interesting model on their own. The adversarial v”
Reusing historical interaction data for faster online learning to rank for IR (2013, methodology): “To generate clicks, we employ a user model based on the Dependent Click Model (DCM) [8],(4) an extension of the Cascade Model [6] that has been shown to be effective in explaining users’ click behavior in web search.”
Equity of Attention: Amortizing Individual Fairness in Rankings (2018, result): “The existence of position bias in rankings of search results has been revealed by a number of eye-tracking and other empirical studies [10, 11, 19].”
Click chain model in web search (2009, methodology): “5% better LL for DCM compared with a previous implementation in [7] on the same data set (average LL = -1.”
Copeland Dueling Bandits (2015, background): “Queries are sampled randomly and clicks are generated probabilistically by conditioning on these assessments in a way that resembles the behavior of an actual user [36].”
Relative confidence sampling for efficient on-line ranker evaluation (2014, methodology): “Queries are sampled randomly and clicks are generated probabilistically by conditioning on these assessments using a user model that resembles the behavior of an actual user [11, 12].”
Cascading Bandits: Learning to Rank in the Cascade Model (2015, methodology): “Therefore, the model has been extended to more realistic settings (Chapelle & Zhang, 2009; Guo et al., 2009a;b), such as multiple clicks and the persistence of the user.”
A probabilistic method for inferring preferences from clicks (2011, methodology): “The click model simulates user interactions according to the Dependent Click Model (DCM) [8, 9], an extension of the cascade model [6].”
Tailoring click models to user goals (2009, background): “A common assumption in existing click models [6, 7] is that query sessions are independent of each other and share the same user behavior model, while in the real world users may have different browsing strategies for different query terms.”
Time-Aware Click Model (2016, methodology): “Subsequently, the click chain model (CCM) [Guo et al. 2009b] uses Bayesian inference to obtain the posterior distribution of the relevance.”