Privacy cloud calibration fuzzy learning for MEB
Wenjun Hu, Shitong Wang · Kongzhi yu juece · 2012
Under given conditions,Gaussian kernel density estimate with minimum integrated square error(ISE) criterion can be equivalent to the minimum enclosing ball(MEB).Based on this conclusion,a learning method of MEB with privacy cloud data is proposed,called privacy cloud calibration MEB(PCC-MEB).Meanwhile,PCC-MEB is extended to fuzzy privacy cloud calibration MEB(FPCC-MEB) by introducing a fuzzy membership function,which can resolve unclassifiable zones among classes.Experimental results on the artificial and real-word data sets show the effectiveness of presented method.