A New Class of Time Dependent Latent Factor Models with Applications
Authors: Sinead A. Williamson, Michael Minyi Zhang, Paul Damien
JMLR 2020 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Illustrations using synthetic and real data are provided. Keywords: Bayesian Nonparametrics, Latent Factor Models, Time Dependence... Finally, applications of time-dependent np LFMs are shown via simulated and real data analysis. In Section 5, using real data, we show the benefit of incorporating this dynamic, temporal feature persistence and contrast it to a static IBP, DIBP, and DDIBP. Section 5: Experimental Evaluation |
| Researcher Affiliation | Academia | Departments of Statistics and Data Science/Information, Risk and Operations Management University of Texas at Austin Austin, TX 78757, USA; Department of Computer Science Princeton University Princeton, NJ 08544, USA; Department of Information, Risk and Operations Management University of Texas at Austin Austin, TX 78757, USA |
| Pseudocode | No | Section 4 details the inference methods used to implement the models in Section 3, followed by synthetic and real data illustrations in Section 5. ... We obtain inferences for the IBP-distributed matrix Z and the lifetimes ℓnk using a Metropolis-Hastings algorithm described below. (The inference methods are described in prose, not structured pseudocode blocks.) |
| Open Source Code | No | We choose to compare with the IFDM over the related IFUHMM since it offers a more efficient inference algorithm, and because code was made available by the authors. (This refers to the IFDM authors, not the authors of this paper. The paper does not explicitly state that the code for their proposed method is open-source or provide a link.) |
| Open Datasets | Yes | Individual household electric power consumption data set1 available from the UCI Machine Learning Repository2. 2. http://archive.ics.uci.edu/ml ... The audio source that we will look at for this problem is a two minute long recording of various bird calls in Kerala, India3. 3. https://freesound.org/s/27334/ |
| Dataset Splits | Yes | We designated 10% of the observations as our test set. For each test set observation, we held out 30 of the 36 pixels... For validation, 10% of the data were set aside, with a randomly selected six out of seven dimensions being held out... a hold-out sample of 10% of the data, evenly spaced throughout the piece, was set aside. All but eight randomly selected dimensions were held out... a regularly spaced subsample of 2,000 observations, of which we held out 10% of the data randomly as a test set. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU models, CPU types, memory amounts) used for running its experiments. It only mentions 'computationally complex inference protocols' in general terms. |
| Software Dependencies | No | The paper describes the use of Markov chain Monte Carlo (MCMC) approaches, including Gibbs samplers and Metropolis-Hastings algorithms, but does not specify any software libraries, frameworks, or their version numbers used for implementation. |
| Experiment Setup | Yes | For the dynamic np LFM and the static IBP, we used our fully nonparametric sampler. For the DIBP, DDIBP and the IFDM we used code available from the authors. The DIBP and DDIBP codes use a weak limit sampler; we fixed K = 20 for the DDIBP and for the DIBP; the lower value for the DIBP is needed due to the much slower computational protocol for this method. All values are the final MSE averaged over 5 trials from the appropriate posterior distributions following convergence of the MCMC chain. |