Title :
Group-Air grouping algorithm based on support vector clustering
Author :
Qi Linghui ; Zhang An ; Bi Wenhao
Author_Institution :
Sch. of Electron. & Inf., Northwestern Polytech. Univ., Xi´an, China
Abstract :
Aiming at the clustering problem of no noise, support vector machine training algorithms in support vector clustering (SVC) have been optimized, and the improved SVC algorithm was applied in the study of Group-Air grouping. This paper introduced the maximum entropy principle to solve the Lagrange multipliers, as the result, the Support vectors (SVs) are effectively reduced, and the performance of the support vector clustering process is improved. Group-Air grouping model is described, and the attributes set of target point during clustering is set up. Experiment verified the improved M-SVC (maximum entropy-SVC) could accomplish Group-Air grouping using clustering Battlefield situation information. Experimental results show the feasibility and effectiveness of the improved M-SVC algorithm.
Keywords :
command and control systems; maximum entropy methods; optimisation; pattern clustering; support vector machines; Lagrange multipliers; M-SVC algorithm; battlefield situation information clustering; group-air grouping algorithm; maximum entropy principle; maximum entropy-SVC algorithm; support vector clustering process performance improvement; support vector machine training algorithm optimisation; support vector reduction; target point attribute set; Clustering algorithms; Entropy; Kernel; Labeling; Static VAr compensators; Support vector machines; Training; Group Air grouping; Support Vector Clustering (SVC); Support Vector Machine Training; maximum entropy;
Conference_Titel :
Control Conference (CCC), 2013 32nd Chinese
Conference_Location :
Xi´an
