Mind the gaps: weighting the unknown in large-scale one-class collaborative filtering

By Rong Pan, Martin Scholz. 2009

170 citations recorded by the source corpus.

Subfields: Artificial Intelligence, 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 Artificial Intelligence

1. Xiangnan He
2. Tat-seng Chua
3. Xiang Wang
4. Meng Wang
5. Fuli Feng
6. Jun Xu
7. Yongfeng Zhang
8. Liqiang Nie

In Information Systems

1. Xiangnan He
2. Tat-seng Chua
3. Xiang Wang
4. Julian McAuley
5. Meng Wang
6. Fuli Feng
7. Steffen Rendle
8. Liqiang Nie

Citations in the corpus

• Fused Matrix Factorization with Geographical and Social Influence in Location-Based Social Networks (2012, methodology): “This is a standard way to solve the one-class problem in CF (Pan et al. 2008; Pan and Scholz 2009).”
• TFMAP: optimizing MAP for top-n context-aware recommendation (2012, methodology): “Some extensions following this approach are introduced in [23] and [29].”
• Content-Aware Point of Interest Recommendation on Location-Based Social Networks (2015, background): “(1), following the standard strategy of solving one-class CF problems (Pan and Scholz 2009; Pan et al. 2008), we sample 10% of unobserved check-ins from the training matrix, deem them as the check-in frequency of 0 and set their corresponding Wi j to 1.”
• One-class collaborative filtering with random graphs (2013, background): “If a parametric definition of confidence weights is given, a low rank approximation of the weighting scheme can also be included in an objective function [18].”
• Factorizing personalized Markov chains for next-basket recommendation (2010, methodology): “Pan and Scholz [6] also use case weights but with several optimization criteria like hinge-loss and least-square.”
• Improving one-class collaborative filtering by incorporating rich user information (2010, methodology): “So far, research on OCCF has focused on how to best model the missing examples [16, 15, 9, 20].”
• Bias and Debias in Recommender System: A Survey and Future Directions (2020, background): “Training Heuristic [69], [70], [71], [72], [73], [74], [75], [76] Sampling [30], [43], [74], [77], [78], [79] Exposure-based model [7], [8], [30], [31], [80] Others [5], [6], [81], [82], [83], [84], [85] Position Bias Click models [86], [87], [88], [89], [90], [91], [92] Propensity Score [5], [93], ”
• Fast Matrix Factorization for Online Recommendation with Implicit Feedback (2016, background): “ALS can be seen as an instantiation of CD and has been widely used to solve the whole-based MF [12, 20, 21, 30]; however, its inefficiency is the main obstacle for practical use [23, 31].”
• Collaborative matrix factorization with multiple similarities for predicting drug-target interactions (2013, methodology): “WLRA with Tikhonov regularization over A and B is also equivalent to one formalization for recommender systems, called one-class collaborative filtering, where an alternating least squares algorithm was presented for estimating A and B [20, 19, 15].”
• DELF: A Dual-Embedding based Deep Latent Factor Model for Recommendation (2018, background): “Recent works [ Pan et al. , 2008; Pan and Scholz, 2009; He et al. , 2014; 2016; Volkovs and Yu, 2015] has proposed non-uniform weighting strategies and sampling schemes to carefully select negative feedback from unobserved interactions.”
• Learning Hierarchical Representation Model for NextBasket Recommendation (2015, methodology): “[19] also introduce the weights to user-item pairs, and optimize the factorization with both least-square and hinge-loss criteria.”
• Recommender Systems (2016, unknown): “In the unary case, substituting missing entries with 0s also leads to some bias [457, 467, 468], although it is often small because the default assumption in implicit feedback data, such as buying data, is that the user will not buy most of the items.”