Dynamic Bayesian Probabilistic Matrix Factorization
Authors: Sotirios Chatzis
AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We present experimental results to demonstrate the superiority of our temporal model. |
| Researcher Affiliation | Academia | Sotirios P. Chatzis Department of Electrical Engineering, Computer Engineering, and Informatics Cyprus University of Technology Limassol 3603, Cyprus soteri0s@mac.com |
| Pseudocode | No | The paper describes its inference algorithm using mathematical equations but does not provide pseudocode or a clearly labeled algorithm block. |
| Open Source Code | No | The paper does not explicitly state that the authors are releasing the source code for their described methodology, nor does it provide a link to a code repository. |
| Open Datasets | Yes | To assess the efficacy of our approach, we experiment with the Netflix dataset1. This dataset comprises users ratings to various movies, given on a 5-star scale. 1http://archive.ics.uci.edu/ml/datasets/Netflix+Prize |
| Dataset Splits | Yes | Apart from the training set, the Netflix dataset also provides a probe set which comprises 1, 408, 395 uniformly selected users. From the resulting dataset, we randomly select 30% of the ratings, and no more than 10 ratings in any case, from each user as the test set, and use the rest for training. This procedure is similar to the way Netflix Prize created their test set. We perform 10-fold cross-validation to alleviate the effects of random selection of training and test samples on the obtained performance measurements. |
| Hardware Specification | Yes | We run our experiments as a single-threaded MATLAB process on a 2.53 GHz Intel Core 2 Duo CPU. |
| Software Dependencies | No | The paper mentions running experiments as a "single-threaded MATLAB process" but does not specify a version number for MATLAB or any other software dependencies. |
| Experiment Setup | Yes | BPMF is initialized with λU = λV = 1, m U = m V = 0, ηU = ηV = 30, SU = SV = I, similar to (Salakhutdinov and Mnih 2008). Similar hyperparameter values are selected for our model, to ensure fairness in our comparisons. |