TY - GEN
T1 - Vote Parallel SVM
T2 - 4th IEEE SmartWorld, 15th IEEE International Conference on Ubiquitous Intelligence and Computing, Advanced and Trusted Computing, Scalable Computing and Communications, Cloud and Big Data Computing, Internet of People and Smart City Innovations, SmartWorld/UIC/ATC/ScalCom/CBDCom/IoP/SCI 2018
AU - Song, Yan
AU - Jin, Qun
AU - Yan, Ke
AU - Lu, Huijuan
AU - Pan, Julong
N1 - Funding Information:
ACKNOWLEDGMENT This work is in part supported by National Natural Science Foundation of China (Nos. 61272315 and 61602431), and International Cooperation Project of Zhejiang Provincial Science and Technology Department (No. 2017C34003).
Publisher Copyright:
© 2018 IEEE.
PY - 2018/12/4
Y1 - 2018/12/4
N2 - Support Vector Machine (SVM) is a set of machine learning algorithms, which has been widely used in diverse domains. With the increasing size of datasets, the traditional SVM training algorithms for large-scale datasets become infeasible. Mathematical optimization and cascade parallelism are both popular strategies for accelerating SVM training. In these parallel methods, the use of appropriate parallel framework to reduce SVM training time has become a priority issue and a research focus. In this paper, we investigate and overview mathematical optimization algorithms and parallel technologies of SVM, and summarize parallel SVM solutions and application problems under different frameworks. We propose a Vote Parallel SVM to reduce the training time. Finally, we show experimental results comparing with baseline methods.
AB - Support Vector Machine (SVM) is a set of machine learning algorithms, which has been widely used in diverse domains. With the increasing size of datasets, the traditional SVM training algorithms for large-scale datasets become infeasible. Mathematical optimization and cascade parallelism are both popular strategies for accelerating SVM training. In these parallel methods, the use of appropriate parallel framework to reduce SVM training time has become a priority issue and a research focus. In this paper, we investigate and overview mathematical optimization algorithms and parallel technologies of SVM, and summarize parallel SVM solutions and application problems under different frameworks. We propose a Vote Parallel SVM to reduce the training time. Finally, we show experimental results comparing with baseline methods.
KW - Cascade SVM
KW - Parallel SVM
KW - Parallel prameworks
KW - Vote Parallel SVM
UR - http://www.scopus.com/inward/record.url?scp=85060315658&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85060315658&partnerID=8YFLogxK
U2 - 10.1109/SmartWorld.2018.00325
DO - 10.1109/SmartWorld.2018.00325
M3 - Conference contribution
AN - SCOPUS:85060315658
T3 - Proceedings - 2018 IEEE SmartWorld, Ubiquitous Intelligence and Computing, Advanced and Trusted Computing, Scalable Computing and Communications, Cloud and Big Data Computing, Internet of People and Smart City Innovations, SmartWorld/UIC/ATC/ScalCom/CBDCom/IoP/SCI 2018
SP - 1942
EP - 1947
BT - Proceedings - 2018 IEEE SmartWorld, Ubiquitous Intelligence and Computing, Advanced and Trusted Computing, Scalable Computing and Communications, Cloud and Big Data Computing, Internet of People and Smart City Innovations, SmartWorld/UIC/ATC/ScalCom/CBDCom/IoP/SCI 2018
A2 - Loulergue, Frederic
A2 - Wang, Guojun
A2 - Bhuiyan, Md Zakirul Alam
A2 - Ma, Xiaoxing
A2 - Li, Peng
A2 - Roveri, Manuel
A2 - Han, Qi
A2 - Chen, Lei
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 7 October 2018 through 11 October 2018
ER -