Integrating Features and Similarities: Flexible Models for Heterogeneous Multiview Data
Authors: Wenzhao Lian, Piyush Rai, Esther Salazar, Lawrence Carin
AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate the effectiveness of our framework on several real-world and benchmarks datasets. In this section, we first apply our framework for analyzing a real-world dataset from cognitive neuroscience. We then present results on benchmark datasets for recommender system matrix completion and classification, respectively. |
| Researcher Affiliation | Academia | Wenzhao Lian, Piyush Rai, Esther Salazar, Lawrence Carin ECE Department, Duke University Durham, NC 27708 {wenzhao.lian,piyush.rai,esther.salazar,lcarin}@duke.edu |
| Pseudocode | No | The paper describes the inference algorithm with equations and step-by-step descriptions of updates, but it does not include a clearly labeled pseudocode or algorithm block. |
| Open Source Code | No | The paper does not provide any statement or link regarding the availability of open-source code for the described methodology. |
| Open Datasets | Yes | For this task, we consider two benchmark datasets1, Epinion and Ciao, both having two views: ordinal rating matrix (range 1-5) and similarity matrix. 1http://www.public.asu.edu/ jtang20/datasetcode/truststudy.htm/ For this experiment, we choose two benchmark datasets: UCI Handwritten Digits (Kumar, Rai, and Daum e III 2011) and Protein Fold Prediction (G onen 2012). |
| Dataset Splits | Yes | To do so, we first split the data at random into training (50%) and testing (50%) sets. We run 10 different splits with 50% of the observed ratings as training set and the remaining 50% ratings as test set. We split the data into 100 digits for training and 1900 digits for test. For Protein data, we split the data equally into training and test sets. |
| Hardware Specification | No | The paper does not provide any specific hardware details used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific software dependencies with version numbers used for the experiments. |
| Experiment Setup | Yes | We perform analysis considering Km = 20 (for similarity-based views), R = 30 latent factors, and prior hyperparameters aα = bα = 0.01. |