Latent Sparse Modeling of Longitudinal Multi-Dimensional Data

Authors: Ko-Shin Chen, Tingyang Xu, Jinbo Bi

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Computational results on synthetic datasets and realfile f MRI and EEG problems demonstrate the superior performance of the proposed approach over existing techniques.
Researcher Affiliation Collaboration 1Department of Computer Science and Engineering, University of Connecticut, Storrs, CT, USA ko-shin.chen@uconn.edu, jinbo.bi@uconn.edu 2 Tencent AI Lab, Shenzhen, China, tingyangxu@tencent.com
Pseudocode Yes Algorithm 1 Search for optimal ˆΦ
Open Source Code No The paper does not provide any links to open-source code or explicitly state that the code for their methodology is released.
Open Datasets Yes The f MRI data used in the experiment were collected by the Alzheimer s Disease Neuroimaging Initiative (ADNI)1. 1http://adni.loni.usc.edu/
Dataset Splits Yes We randomly select 80% of the subjects for training and the rest for testing...The λ1, λ2, and λ3 were tuned in a two-fold cross validation. In other words, the training records were further split into half: one used to build a model with a chosen parameter value from a range of 1 to 20 with a step size of 0.1; and the other used to test the resultant model.
Hardware Specification No The paper does not explicitly describe the hardware (e.g., specific CPU/GPU models, memory) used for running the experiments.
Software Dependencies No The paper does not provide specific version numbers for any software dependencies or libraries used in the experiments.
Experiment Setup Yes In our experiments, λ s are tuned as λ1 = λ2 = λ3 = 0.3 based on cross validation within training...The λ1, λ2, and λ3 were tuned in a two-fold cross validation. In other words, the training records were further split into half: one used to build a model with a chosen parameter value from a range of 1 to 20 with a step size of 0.1; and the other used to test the resultant model...The hyperparameters λ1, λ2, and λ3 in our approach and GEE/PGEE (one parameter) were tuned in a two-fold cross validation within the training data.