Group Sparse Additive Machine
Authors: Hong Chen, Xiaoqian Wang, Cheng Deng, Heng Huang
NeurIPS 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results on synthetic data and seven benchmark datasets consistently show the effectiveness of our new approach. In Section 4, experimental results on both simulated examples and real data are presented and discussed. |
| Researcher Affiliation | Academia | Hong Chen1, Xiaoqian Wang1, Cheng Deng2, Heng Huang1 1 Department of Electrical and Computer Engineering, University of Pittsburgh, USA 2 School of Electronic Engineering, Xidian University, China |
| Pseudocode | No | For given {τj}, the optimization problem of Group SAM can be computed efficiently via an accelerated proximal gradient descent algorithm developed in [30]. Due to space limitation, we don’t recall the optimization algorithm here again. |
| Open Source Code | No | The paper does not provide an explicit statement about releasing the source code for Group SAM or a link to a code repository. |
| Open Datasets | Yes | In this subsection, we use 7 benchmark data from UCI repository [12] to compare the classification performance of different methods. The 7 benchmark data includes: Ecoli, Indians Diabetes, Breast Cancer, Stock, Balance Scale, Contraceptive Method Choice (CMC) and Fertility. |
| Dataset Splits | Yes | In comparison, we adopt 2-fold cross validation and report the average performance of each method. We tune the hyper-parameters via 2-fold cross validation on the training data and report the best parameter w.r.t. classification accuracy of each method. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., CPU, GPU models, or memory) used for running the experiments. |
| Software Dependencies | No | The paper mentions using 'LIBSVM toolbox [2]' but does not specify its version number or any other software dependencies with versions. |
| Experiment Setup | Yes | We determine the hyper-parameter of all models, i.e., parameter C of SVM, L1SVM and Gaussian SVM, parameter λ of SAM, parameter λ of Group Sp AM, parameter λ in Eq. (6) of Group SAM, in the range of {10-3, 10-2, . . . , 103}. We tune the hyper-parameters via 2-fold cross validation on the training data and report the best parameter w.r.t. classification accuracy of each method. In the accelerated proximal gradient descent algorithm for both SAM and Group SAM, we set µ = 0.5, and the number of maximum iterations as 2000. |