SLIM: Sparse Linear Methods for Top-N Recommender Systems

By Xia Ning, George Karypis. 2011

774 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Information Systems, Management Science and Operations Research, Recommender 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

In Management Science and Operations Research

In Recommender Systems

Citations in the corpus

Variational Autoencoders for Collaborative Filtering (2018, background): “elihoods have typically received less attention in the collaborative filtering literature, especially with latent-factor models. Most previous work builds on Gaussian likelihoods (square loss, Eq. 3) [19, 33, 51] or logistic likelihood (log loss, Eq. 4) [14, 51] instead. We argue in Section 2.1 that”
On the challenges of learning with inference networks on sparse, high-dimensional data (2017, methodology): “With them, the same generative model , outperforms CDAE and WMF on both datasets, and marginally outperforms SLIM on ML-20M while achieving nearly state-of-the-art re-sults on Netﬂix.”
On the Effectiveness of Linear Models for One-Class Collaborative Filtering (2016, methodology): “To see the connection, observe that we may rewrite SLIM as (see also (Ning and Karypis 2011, Equation 4)) where C is as in Equation 8, (cid:96) ( y, v ) = ( y − v ) 2 is the square loss, and the feature matrix X and label vector y ( i ) is LRec can thus be seen as a variant of SLIM where: (1) one…”
Sparse linear methods with side information for top-n recommendations (2012, methodology): “SSLIM is an extension to the Sparse LInear Method [4] (SLIM).”
A Comprehensive Survey of Neighborhood-based Recommendation Methods (2011, methodology): “A representative neighborhood-learning recommendation method is the SLIM algorithm, developed by Ning et al [50].”
Transparent, Scrutable and Explainable User Models for Personalized Recommendation (2019, methodology): “Therefore, users are presented with 20 movies, pooled from the following sources: two baseline meth-ods (ItemKNN and BPRSLIM), the initial set-based model (i.e., no scrutinization), and the set-based model with different user model variants, based on the anticipated changes.”
Translation-based Recommendation (2017, methodology): “For implicit feedback data (like purchases, clicks, and thumbs-up), point-wise (e.g. [ Hu et al. , 2008; Pan et al. , 2008; Ning and Karypis, 2011]) and pairwise methods (e.g. [Rendle et al. , 2009]) based on MF have been proposed.”
RecBole: Towards a Unified, Comprehensive and Efficient Framework for Recommendation Algorithms (2020, neutral): “SLIMElastic [44] ICDM 2011 ctlg, ccard, ecmrc,Book-Crossing, MoiveLens, Netflix, Yahoo Music”
Recommender Systems (2016, neutral): “Item-item regression models have also been studied in the context of Sparse Linear Models (SLIM) [455], where an elastic-net regularizer is used on the linear model without restricting the coefficients to the”
FISM: factored item similarity models for top-N recommender systems (2013, methodology): “Extensive experiments in [7] have shown that SLIM outperforms the rest of the state-of-the-art top-N recommendation methods.”
Embarrassingly Shallow Autoencoders for Sparse Data (2019, methodology): “It was introduced in [16].”
Fusing Similarity Models with Markov Chains for Sparse Sequential Recommendation (2016, methodology): “Without parameterizing each user explicitly, SLIM relaxes the low-rank assumption enforced on user representations and has achieved higher recommendation accuracy (see [15] for details).”
A Generic Coordinate Descent Framework for Learning from Implicit Feedback (2016, background): “Both approaches are considered state-of-the-art and while CD outperforms BPR on certain datasets [8, 25, 15, 22, 26], BPR has been shown to work better on others [4, 17, 16, 8].”
Auto-Encoding User Ratings via Knowledge Graphs in Recommendation Scenarios (2017, methodology): “Another interesting result here is that increasing the number of ratings we need a higher number of neighbors in K ( u ) to compute w ( f ′ n , u ) in Equation (2) to reach results comparable with BPRSLIM.”
Collaborative Denoising Auto-Encoders for Top-N Recommender Systems (2016, background): “Instead, it learns a similarity matrix from data [12].”
Transfer Meets Hybrid: A Synthetic Approach for Cross-Domain Collaborative Filtering with Text (2019, background): “Since the relationships between items are shown to be important in improving recommendation performance [26, 36, 37, 41] for single domain, we”