[Wu, Xiang-hua] Hunan City Univ, Network Informat Ctr, Yiyang, Peoples R China.
[Wu, Xiang-hua] Cent S Univ, Coll Informat Sci & Engn, Changsha, Hunan, Peoples R China.
[Cao, Li-jun] Hunan City Univ, Dept Comp Sci, Yiyang, Peoples R China.
通讯机构:
[Wu, Xiang-hua] H
Hunan City Univ, Network Informat Ctr, Yiyang, Peoples R China.
语种:
中文
关键词:
learning manifold;out of sample;dynamic neighborhood;LLE
期刊:
2011 INTERNATIONAL CONFERENCE ON FUTURE COMPUTER SCIENCE AND APPLICATION (FCSA 2011), VOL 3
年:
2011
页码:
500-503
会议名称:
International Conference on Future Computer Science and Application (FCSA 2011)
会议时间:
JUL 16-17, 2011
会议地点:
Kota Kinabalu, MALAYSIA
会议主办单位:
[Wu, Xiang-hua] Hunan City Univ, Network Informat Ctr, Yiyang, Peoples R China.^[Wu, Xiang-hua] Cent S Univ, Coll Informat Sci & Engn, Changsha, Hunan, Peoples R China.^[Cao, Li-jun] Hunan City Univ, Dept Comp Sci, Yiyang, Peoples R China.
会议赞助商:
Univ Malaya, Wuhan Univ, China Univ Geosciences, Huazhong Normal Univ, Int Ind Elect Ctr Hong Kong
主编:
Esa, R
出版地:
FLAT 4A YUE FAT BUILDING, 87-91 TAI PO ROAD, SHAM SHUI PO, KOWLOON 00000, PEOPLES R CHINA
出版者:
INT INDUSTRIAL ELECTRONIC CENTER
ISBN:
978-988-19116-9-8
机构署名:
本校为第一且通讯机构
院系归属:
信息与电子工程学院
摘要:
Manifold learning was designed to identify the meaningful low-dimensional structure hidden in the high-dimensional data, most of the existing algorithms are non-incremental. In this paper an incremental algorithm based on locally reconstruction is proposed, the algorithm encountered lead to the discussion on the impact of dynamic neighborhood choose, and ultimately the incremental algorithm based on locally reconstruction of dynamic neighborhood is designed. The algorithm takes the satisfying experime...