Recommending Groups to Users Using User-Group Engagement and Time-Dependent Matrix Factorization

Authors: Xin Wang, Roger Donaldson, Christopher Nell, Peter Gorniak, Martin Ester, Jiajun Bu

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments indicate that the time-varying implicit engagement-based model provides the best top-K group recommendations, illustrating the benefit of the added model complexity. Experiments Our experiments consider Deviant Art users and groups from 5 May 2011 to 31 August 2014.
Researcher Affiliation Collaboration Zhejiang Provincial Key Laboratory of Service Robot, College of Computer Science, Zhejiang University, China School of Computing Science, Simon Fraser University, Burnaby, B.C., Canada Department of Mathematics, The University of British Columbia, Vancouver, B.C., Canada Deviant Art, Inc. Vancouver, B.C., Canada
Pseudocode No The paper describes optimization procedures and mathematical formulations in text, but does not include a clearly labeled pseudocode or algorithm block/figure.
Open Source Code No The paper does not provide any concrete access information for open-source code related to the described methodology.
Open Datasets No The paper uses 'three real-world datasets from Deviant Art' and references 'Deviant Art. http://www.deviantart.com/', which is the platform. It does not provide concrete access information (like a specific link, DOI, or formal citation to a downloadable dataset) for the datasets used in the experiments, implying they are internal.
Dataset Splits Yes The first 9 intervals serve as training/validation data; we withhold the last 4 months for testing. The static model aggregates the first 9 intervals into a single training-validation set, on which a 5-fold cross validation is used to select the optimal parameters, while the temporal model performs the matrix factorization on each of the first 9 intervals separately, making recommendations on the 10th interval using our temporal scheme.
Hardware Specification No The paper does not provide specific hardware details (like exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., library names and their versions) needed to replicate the experiment.
Experiment Setup Yes Parameters γ, which scales the strength of user-item ratings, λ, which regularizes matrix factors, and k, the dimension of the latent space, are chosen by experiment. Choosing θ = 0.9 appears to be a good compromise between fast convergence and local minima avoidance in relaxing updates for all of Xu, Yg and w. Grid search is used to explore various values of p in [1, 8] and we present results with p = 3 which achieve the best performance.