Factorization meets the neighborhood: a multifaceted collaborative filtering model

By Yehuda Koren. 2008

4,350 citations recorded by the source corpus.

Subfields: Information Systems, Management Science and Operations Research, Marketing, 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 Management Science and Operations Research

In Marketing

In Information Retrieval & Search

Citations in the corpus

On Sampling Strategies for Neural Network-based Collaborative Filtering (2017, methodology): “In most of existing matrix factorization techniques [17, 18, 27], each user/item ID is associated with a latent vector u or v (i.”
Evaluating Prerequisite Qualities for Learning End-to-End Dialog Systems (2015, methodology): “However, it has been shown to be outperformed by other ﬂavors of matrix factorization, in particular by using a ranking loss rather than squared loss (Weston et al., 2013) which we will compare to (cf. sec 3.2), as well as improvements like SVD++ (Koren, 2008).”
Aligning AI With Shared Human Values (2020, background): “Recommender systems learn individual preferences about products (Koren, 2008).”
Recurrent Latent Variable Networks for Session-Based Recommendation (2017, background): “Consequently, under this seing, conventional algorithmic approaches towards RS are confronted with hard challenges that stem from the unavailability of rich user proles (data sparsity) [19].”
Improved Recurrent Neural Networks for Session-based Recommendations (2016, methodology): “Collaborative filtering approaches [15, 14, 21] can define useruser similarity based on their history of purchases, or they”
Gaussian Ranking by Matrix Factorization (2015, methodology): “This assumption is not only a useful working hypothesis in practice, as illustrated in Section 7, but it may also be motivated as follows: due to data sparsity, the vast majority of items belongs to the same (ie negative) class.”
RippleNet: Propagating User Preferences on the Knowledge Graph for Recommender Systems (2018, methodology): “Traditional CF-basedmethods and their variants [11, 31] learn latent representations of users and items, then predict unknown ratings by directly applying a specific function to their representations such as inner product.”
Explainable Recommendation: A Survey and New Perspectives (2018, methodology): “The most classical model-based approach to recommendation could be Latent Factor Models (LFM) based on Matrix Factorization (MF) techniques Koren et al. [2009], which attempts to learn latent factors to predict the missing ratings in a user-item rating matrix.”
How algorithmic confounding in recommendation systems increases homogeneity and decreases utility (2017, methodology): “In considering collaborative ltering alone, neighborhood-based methods can be merged with latent factor approaches [34].”
Recommendations as Treatments: Debiasing Learning and Evaluation (2016, methodology): “Except for the propensities Pu,i that act like weights for each loss term, the training objective is identical to the standard incomplete matrix factorization objective (Koren, 2008; Steck, 2010; Hu et al., 2008) with MSE (using Eq.”
BPR: Bayesian Personalized Ranking from Implicit Feedback (2009, methodology): “We have chosen the two diverse model classes of matrix factorization [5, 12] and learned k-nearest-neighbor [8].”
A Review of Relational Machine Learning for Knowledge Graphs (2015, neutral): “ARE and the models of [84] and [85] are similar in spirit to the model of [119], which augments SVD (i.”
Practical Privacy Preserving POI Recommendation (2020, background): “and user-POI actions are private, whereas POI descriptions are public to all the users. The model refers to the built recommendation model, e.g., the latent factors of Matrix Factorization (MF) model [30], that predicts users’ preferences on POIs. First, besides the public POI data, users’ private d”
Exploring temporal effects for location recommendation on location-based social networks (2013, methodology): “To solve largescale recommendation problems, matrix factorization is state-ofthe-art technology that has been proven to be successful in the Netflix Competition [11, 12], and is being used for item recommendation and trust prediction on product review sites like Epinions and Ciao for research purpos”
Fast ALS-based tensor factorization for context-aware recommendation from implicit feedback (2012, methodology): “We set the number of features to 20 and 40 commonly used in literature [1,7].”