Adaptive Hypergraph Learning for Unsupervised Feature Selection

Zhu, Xiaofeng; Zhu, Yu; Zhang, Shichao; Hu, Rongyao; He, Wei

doi:10.24963/ijcai.2017/501

Cited by 32 publications

(18 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The hyperedges generated from the original training data may result in an inaccurate hypergraph. To deal with this, we design to learn the hyperedges from low-dimensional training data, whose redundant and irrelevant label space information have been removed as much as possible [18]. Next we will describe in detail how to update L p :…”

Section: E Optimizationmentioning

confidence: 99%

Robust Cross-View Embedding With Discriminant Structure for Multi-Label Classification

Wang

2021

IEEE Access

View full text Add to dashboard Cite

Label embedding is an important family of multi-label classification algorithms which can jointly extract the information of all labels for better performance. However, few works have been done to develop the multi-label embedding methods that can effectively deal with the interference of noisy data during training process. The noise often makes the labels of a few samples incorrect (i.e., missing or mislabeled), which could lead to a poor learning performance. To address this issue, we propose a novel cross-view based model. It performs a robust and discriminant embedding, namely Robust Crossview Embedding with Discriminant Structure for Multi-label Classification (RCEDS). In RCEDS, a novel hypergraph fusion technique is designed to explore and utilize the complementary between the feature space and the label space to make the proposed RCEDS robust. Meanwhile, we use double-side metric learning to mine the consistency between the feature space and the label space to effectively improve the discriminative ability of our proposed RCEDS. Furthermore, we conduct a deep extension of RCEDS and effectively apply it to image annotation. Extensive experimental results on data sets with many labels demonstrate that our proposed approach can attain better classification performance than the existing label embedding algorithms.INDEX TERMS multi-label classification, label embedding, cross view, hypergraph fusion, double-side metric learning

show abstract

Section: E Optimizationmentioning

confidence: 99%

Robust Cross-View Embedding With Discriminant Structure for Multi-Label Classification

Wang

2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…In the literature, many classifiers are used for gesture recognition, e.g., Dynamic Time Warping (DTW) [43], [44], [45], [24], linear SVMs [41], neuro-fuzzy inference system networks [46], hyper rectangular composite NNs [47], 3D Hopfield NN [48], sparse coding [49], [50], [51], [52], [53], [54]. Due to the ability of modeling temporal signals, Hidden Markov Model (HMM) is possibly the most well known classifier for gesture recognition.…”

Section: Related Workmentioning

confidence: 99%

BoMW: Bag of Manifold Words for One-Shot Learning Gesture Recognition From Kinect

Zhang

Jiang

et al. 2018

IEEE Trans. Circuits Syst. Video Technol.

View full text Add to dashboard Cite

“…This way, local structure is also considered. Other unsupervised embedded algorithms include NDFS (Li et al 2012), RUFS (Qian and Zhai 2013), FSASL (Du and Shen 2015), SOGFS (Nie, Zhu, and Li 2016), AHLFS (Zhu et al 2017), UPFS (Li et al 2018b), DGUFS (Guo and Zhu 2018), etc. Typically, existing similarity preserving feature selection schemes face two major challenges: • High dimensionality.…”

Section: Introductionmentioning

confidence: 99%

IVFS: Simple and Efficient Feature Selection for High Dimensional Topology Preservation

2020

AAAI

View full text Add to dashboard Cite

Feature selection is an important tool to deal with high dimensional data. In unsupervised case, many popular algorithms aim at maintaining the structure of the original data. In this paper, we propose a simple and effective feature selection algorithm to enhance sample similarity preservation through a new perspective, topology preservation, which is represented by persistent diagrams from the context of computational topology. This method is designed upon a unified feature selection framework called IVFS, which is inspired by random subset method. The scheme is flexible and can handle cases where the problem is analytically intractable. The proposed algorithm is able to well preserve the pairwise distances, as well as topological patterns, of the full data. We demonstrate that our algorithm can provide satisfactory performance under a sharp sub-sampling rate, which supports efficient implementation of our proposed method to large scale datasets. Extensive experiments validate the effectiveness of the proposed feature selection scheme.

show abstract

Adaptive Hypergraph Learning for Unsupervised Feature Selection

Cited by 32 publications

References 9 publications

Robust Cross-View Embedding With Discriminant Structure for Multi-Label Classification

Robust Cross-View Embedding With Discriminant Structure for Multi-Label Classification

BoMW: Bag of Manifold Words for One-Shot Learning Gesture Recognition From Kinect

IVFS: Simple and Efficient Feature Selection for High Dimensional Topology Preservation

Contact Info

Product

Resources

About