.png){kind=icon}
Return
Principal component analysis based on graph embedding
.png){kind=icon}
PRE
AI
DOI：10.1007/s11042-022-13620-6.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Dimensionality reduction plays an important role in image recognition and data mining Traditional methods extract features from data itself and ignore the structure information of data even though it is crucial for effective representation. Considering graph embedding method can capture and model the complicated relationships among data, therefore, we consider to incorporate graph convolution learning into principal component analysis (GCPCA) to abstract more effective features in this paper. The key idea of the proposed model is embedding graph convolutional to realize linear representation by fusing the relationship of data points. Then PCA is operated on projected data to extract effective features. The model can be solved to obtain a globally optimal closed-form solution, which is convenient for implementation and practical application. Experiments on some publicly available datasets demonstrate that the proposed GCPCA model show the better performance than the existing classical algorithms in terms of classification accuracy.
Keywords:
Dimensionality reduction
Graph embedding
Principal component analysis
Feature extraction
Journal
IF:
3
Papers:
2.0W
Citations:
3.2W
Organization
View more organizations
Cited Papers
.png){kind=icon}
ShareSave
ShareSaveShareSaveShareSaveShareSaveShareSaveShareSave View more