CAJ | 학술논문

随着的计算能力的不断提高和计算机体系结构的可编程性，将向着多核，众核的异质形核的方向继续发展。针对这一问题，本文对KFCM算法的类别间可分性优化进行了分析，通过对于FCM算法的数据集C划分、FCM算法和HCM算法的理论知识，解释了KFCM算法，对于样本的特征进行优化，将高维特征空间内的数据映射到内核函数中，将样本的有益特征扩大，到达快而准的聚类效果。经过仿真测试显示，KFCM算法模型聚类效果可以准确区分二者。
수착적계산능력적불단제고화계산궤체계결구적가편정성，장향착다핵，음핵적이질형핵적방향계속발전。침대저일문제，본문대KFCM산법적유별간가분성우화진행료분석，통과대우FCM산법적수거집C화분、FCM산법화HCM산법적이론지식，해석료KFCM산법，대우양본적특정진행우화，장고유특정공간내적수거영사도내핵함수중，장양본적유익특정확대，도체쾌이준적취류효과。경과방진측시현시，KFCM산법모형취류효과가이준학구분이자。
Along with multiple cores, the core processor to become the mainstream of computing devices, pattern recognition, realization of parallel system in the future will need corresponding parallel algorithm research as its theoretical basis. In order to solve this problem, in this paper, the nuclear class separability between the fuzzy c-means algorithm optimization are analyzed, through the fuzzy c-means algorithm for data set C, C- average clustering algorithm and fuzzy C- average clustering algorithm in detail, which leads to the nuclear fuzzy C-average clustering algorithm, this method is increased by the optimization of sample characteristics, application samples of kernel function of the input space is mapped to high-dimensional feature space, and clustering in the feature space, the kernel function clustering method in performance than classical clustering algorithm has great improvement, it through nonlinear mapping can distinguish useful feature, extracted and amplified, achieve a more accurate and more rapid clustering. Through the simulation experiments show that the clustering effect KFCM algorithm model can distinguish between two types of completely.