Abstract:
Unsupervised spectral feature selection (USFS) methods can produce interpretable and discriminative results by embedding a Laplacian regularizer in sparse feature selection to maintain training sample local similarity.
USFS methods use a general-graph or hyper-graph on the data to construct the Laplacian matrix. A hyper-graph can measure the relationship between at least two samples, while a general-graph can measure two samples.
The general-graph is a special case of the hyper-graph, which can capture more complex sample structure. Previous USFS methods separated feature selection from Laplacian matrix construction. Original data also have noise.
Each makes feature selection models unreliable. We propose a sparse feature selection method that dynamically constructs a hyper-graph-based Laplacian matrix. Our method outperformed state-of-the-art clustering and segmentation methods on real datasets.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
