Localized factor models for multi-context recommendation

Agarwal, Deepak; Chen, Bee-Chung; Long, Bo

doi:10.1145/2020408.2020504

Cited by 58 publications

(41 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Three main paradigms are proposed for this purpose. The first paradigm is "Regression Based Factor Models" (RBFM) and its extensions, proposed by Agarwal, Chen and colleagues [1,2,3,36], which have been successfully used in a variety of recommendation scenarios, such as social networks [33,34], professional networks and content recommendation. The basic idea behind RBFM is to replace zero-mean Gaussian distributions usually used in a simple LFM with regression-based means.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Co-factorization machines

Liu

Doumith

Davison

2013

Proceedings of the Sixth ACM International Conference on Web Search and Data Mining

138

View full text Add to dashboard Cite

Users of popular services like Twitter and Facebook are often simultaneously overwhelmed with the amount of information delivered via their social connections and miss out on much content that they might have liked to see, even though it was distributed outside of their social circle. Both issues serve as difficulties to the users and drawbacks to the services.Social media service providers can benefit from understanding user interests and how they interact with the service, potentially predicting their behaviors in the future. In this paper, we address the problem of simultaneously predicting user decisions and modeling users' interests in social media by analyzing rich information gathered from Twitter. The task differs from conventional recommender systems as the cold-start problem is ubiquitous, and rich features, including textual content, need to be considered. We build predictive models for user decisions in Twitter by proposing Co-Factorization Machines (CoFM), an extension of a state-of-the-art recommendation model, to handle multiple aspects of the dataset at the same time. Additionally, we discuss and compare rankingbased loss functions in the context of recommender systems, providing the first view of how they vary from each other and perform in real tasks. We explore an extensive set of features and conduct experiments on a real-world dataset, concluding that CoFM with ranking-based loss functions is superior to state-of-the-art methods and yields interpretable latent factors.

show abstract

Section: Related Workmentioning

confidence: 99%

“…An additional possibility is suggested in Agarwal and Chen [3] where a "global" representation is assumed. The "local" representation is drawn from the "global" representation by a multivariate normal distribution.…”

Section: Cofm Via Latent Space Regularizationmentioning

confidence: 99%

Co-factorization machines

Liu

Doumith

Davison

2013

Proceedings of the Sixth ACM International Conference on Web Search and Data Mining

138

View full text Add to dashboard Cite

show abstract

“…Suppose all the existing rating locations of the test set are denoted as Ω. Like many recent research papers [29,1], we also take the Root Mean Squared Error (RMSE) to measure the effectiveness of an algorithm on solving recommendation problems:…”

Section: Experiments On Recommendation Datamentioning

confidence: 99%

Accelerated singular value thresholding for matrix completion

Zhang

et al. 2012

Proceedings of the 18th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

Recovering a large matrix from a small subset of its entries is a challenging problem arising in many real world applications, such as recommender system and image inpainting. These problems can be formulated as a general matrix completion problem. The Singular Value Thresholding (SVT) algorithm is a simple and efficient first-order matrix completion method to recover the missing values when the original data matrix is of low rank. SVT has been applied successfully in many applications. However, SVT is computationally expensive when the size of the data matrix is large, which significantly limits its applicability. In this paper, we propose an Accelerated Singular Value Thresholding (ASVT) algorithm which improves the convergencewhere N is the number of iterations during optimization. Specifically, the dual problem of the nuclear norm minimization problem is derived and an adaptive line search scheme is introduced to solve this dual problem. Consequently, the optimal solution of the primary problem can be readily obtained from that of the dual problem. We have conducted a series of experiments on a synthetic dataset, a distance matrix dataset and a large movie rating dataset. The experimental results have demonstrated the efficiency and effectiveness of the proposed algorithm.

show abstract

“…The general idea is to perform learning on each heterogeneous feature space independently and then summarize the results via ensemble. Recently, [7] proposes a recommendation model (collaborative filtering) that can combine information from different contexts. It finds a latent factor that connects all data sources, and propagate information through the latent factor.…”

Section: Related Workmentioning

confidence: 99%

Learning from Heterogeneous Sources via Gradient Boosting Consensus

Shi

Paiement

Grangier

et al. 2012

Proceedings of the 2012 SIAM International Conference on Data Mining

View full text Add to dashboard Cite

Multiple data sources containing different types of features may be available for a given task. For instance, users' profiles can be used to build recommendation systems. In addition, a model can also use users' historical behaviors and social networks to infer users' interests on related products. We argue that it is desirable to collectively use any available multiple heterogeneous data sources in order to build effective learning models. We call this framework heterogeneous learning. In our proposed setting, data sources can include (i) nonoverlapping features, (ii) non-overlapping instances, and (iii) multiple networks (i.e. graphs) that connect instances. In this paper, we propose a general optimization framework for heterogeneous learning, and devise a corresponding learning model from gradient boosting. The idea is to minimize the empirical loss with two constraints: (1) There should be consensus among the predictions of overlapping instances (if any) from different data sources; (2) Connected instances in graph datasets may have similar predictions. The objective function is solved by stochastic gradient boosting trees. Furthermore, a weighting strategy is designed to emphasize informative data sources, and deemphasize the noisy ones. We formally prove that the proposed strategy leads to a tighter error bound. This approach consistently outperforms a standard concatenation of data sources on movie rating prediction, number recognition and terrorist attack detection tasks. We observe that the proposed model can improve out-of-sample error rate by as much as 80%.

show abstract

Localized factor models for multi-context recommendation

Cited by 58 publications

References 35 publications

Co-factorization machines

Co-factorization machines

Accelerated singular value thresholding for matrix completion

Learning from Heterogeneous Sources via Gradient Boosting Consensus

Contact Info

Product

Resources

About