Sequential Click Prediction for Sponsored Search with Recurrent Neural Networks

By Yuyu Zhang, Hanjun Dai, Chang Xu, Jun Feng, Jiang Bian, Tie-Yan Liu. 2014

355 citations recorded by the source corpus.

Click prediction is one of the fundamental problems in sponsored search. Most of existing studies took advantage of machine learning approaches to predict ad click for each event of ad view independently. However, as observed in the real-world sponsored search system, user's behaviors on ads yield high dependency on how the user behaved along with the past time, especially in terms of what queries she submitted, what ads she clicked or ignored, and how long she spent on the landing pages of clicked ads, etc. Inspired by these observations, we introduce a novel framework based on Recurrent Neural Networks (RNN). Compared to traditional methods, this framework directly models the dependency on user's sequential behaviors into the click prediction process through the recurrent structure in RNN. Large scale evaluations on the click-through logs from a commercial search engine demonstrate that our approach can significantly improve the click prediction accuracy, compared to sequence-independent approaches.

Subfields: Computer Vision and Pattern Recognition, Information Systems

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Information Systems

Citations in the corpus

Recurrent Latent Variable Networks for Session-Based Recommendation (2017, background): “To address these issues, [33] introduced a novel framework based”
Improved Recurrent Neural Networks for Session-based Recommendations (2016, methodology): “In [31], the authors also use RNNs for click sequence prediction; they consider historical user behaviours as well as hand engineered features for each user and item.”
DeepFM: A Factorization-Machine based Neural Network for CTR Prediction (2017, background): “Some ideas extend CNN and RNN for CTR predition [Liu et al. , 2015; Zhang et al. , 2014], but CNN-based models are biased to the interactions between neighboring features while RNN-based models are more suitable for click data with sequential dependency.”
Evaluation of session-based recommendation algorithms (2018, methodology): “lar because some techniques like the one discussed in [53] require the existence of knowledge about certain item attributes. The most recent works on sequence modeling are based on RNNs. Zhang et al. [63], for example, used them for the prediction of user clicks in an advertisement scenario. Hidasi ”
Predicting the Next Location: A Recurrent Model with Spatial and Temporal Contexts (2016, methodology): “• RNN (Zhang et al. 2014): This is a state-of-the-art method for temporal prediction, which has been successfully applied in word embedding and ad click prediction.”
DeepMove: Predicting Human Mobility with Attentional Recurrent Networks (2018, methodology): “Because of its powerful representation ability, RNN have been applied to many fields like click prediction [44], recommendation system [13] and mobility prediction [9, 15, 20, 23, 37, 38].”
Product-Based Neural Networks for User Response Prediction (2016, methodology): “Recurrent neural networks (RNN) was leveraged to model the user queries as a series of user context to predict the ad click behavior [17].”
ATRank: An Attention-Based User Behavior Modeling Framework for Recommendation (2017, methodology): “• Bi-LSTM: We implement a Bi-LSTM method to encode the user behavior history, whose difference with (Zhang et al. 2014) is that we use a bidirectional LSTM as the encoder because of better performance in our experiments.”
Click Models for Web Search (2015, neutral): “, [Zhang et al., 2014].”
Where to Go Next: A Spatio-Temporal Gated Network for Next POI Recommendation (2019, methodology): “• RNN (Zhang et al. 2014): This method leverages the temporal dependency in user’s behavior sequence through a standard recurrent structure.”
Session-based Recommendation with Graph Neural Networks (2018, methodology): “RNN also has been applied successfully in numerous applications, such as the sequential click prediction (Zhang et al. 2014), location prediction (Liu et al. 2016), and next basket recommendation (Yu et al. 2016).”
Embedding-based News Recommendation for Millions of Users (2017, methodology): “[21] proposed a framework based on an RNN for click prediction of sponsored search advertising.”
Sequence-Aware Recommender Systems (2018, methodology): “tern Mining Frequent itemsets [48,56,83,86,106,118] Frequent sequences [78,83,86,95,115,125] Sequence modeling Markov Models [32,42,47,81] Reinforcement Learning [84,97,104] Recurrent Neural Networks [29,45,46,77,98,100,102,107,116,121] Distributed item representations Latent Markov embeddings [19,2”
STAN: Spatio-Temporal Attention Network for Next Location Recommendation (2021, methodology): “Among them, models based on RNN [40] are representative and quickly develop as strong baselines.”
A Dynamic Recurrent Model for Next Basket Recommendation (2016, methodology): “Recently, RNN approaches to word embedding for sentence modeling [5], sequential click prediction [10] have achieved much success in respective ﬁelds.”
Neural Attentive Session-based Recommendation (2017, methodology): “[45] also use RNN for the click sequence prediction, they consider historical user behaviors as well as hand-crafted features for each user and item.”
When Recurrent Neural Networks meet the Neighborhood for Session-Based Recommendation (2017, methodology): “[38] for example successfully applied RNNs in a related setting of predicting sequential advertisement clicks, in which the partitioning into sessions is less important.”