.png){kind=icon}
Return
Semi-supervised deep density clustering
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2023.110903.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Deep clustering generally obtains promising performance by learning deep feature representations. However, there are two limitations: specialIntscript end-to-end deep density clustering needs to be explored; specialIntscript prior information is ignored to guide the learning process. To overcome these limitations, we propose a novel semi-supervised deep density clustering (SDDC). Specifically, a convolutional autoencoder is applied to learn embedded features, and semi-supervised density peaks clustering is designed to identify stable cluster centers. Meanwhile, prior information is introduced to instruct the preferable clustering process. By integrating prior information, a joint clustering loss is directly built on embedded features to perform feature representation and cluster assignment simultaneously. Extensive experiments validate the power of SDDC for initializing and the effectiveness on clustering tasks.
Keywords:
Density-based clustering
Deep clustering
Deep density clustering
Semi-supervised deep clustering
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W
Organization
No organization information available
Cited Papers
.png){kind=icon}
ShareSaveShareSaveRNN-DBSCAN: A Density-Based Clustering Algorithm Using Reverse Nearest Neighbor Density Estimates
OAAI
OAAIShareSaveShareSaveShareSaveShareSaveShareSave
View more