CAJ | 학술논문

基于模糊c-均值算法的无监督最优模糊聚类算法集合了模糊c均值算法与无监督最优聚类算法的优点,它通过逐渐改变聚类数c,依据一些有效性衡量尺度,能无监督搜索出最优聚类数c.通过对距离测量尺度的改进,使聚类不受类形状的影响,以达到具备更高准确率的聚类效果.仿真实验结果表明,新算法不仅能准确找出聚类数,而且跟单纯的模糊c均值算法比,具有更好的聚类效果.
기우모호c-균치산법적무감독최우모호취류산법집합료모호c균치산법여무감독최우취류산법적우점,타통과축점개변취류수c,의거일사유효성형량척도,능무감독수색출최우취류수c.통과대거리측량척도적개진,사취류불수류형상적영향,이체도구비경고준학솔적취류효과.방진실험결과표명,신산법불부능준학조출취류수,이차근단순적모호c균치산법비,구유경호적취류효과.
The unsupervised optimal fuzzy clustering algorithm based on the fuzzy c-means algorithm gathered the advantages of the fuzzy c-means algorithm and the unsupervised optimal clustering algorithm.By changing the number of the clustering c gradually, basing on some measures of the effectiveness,the optimal clustering can be found out without supervising. By improving the measurement of the distance, the clustering can not be influenced by the shape of the class,in order to achieve a higher rate of correct clustering results. The simulation result showed that the new algorithm can not only find out the number of the clustering, but also has a better clustering effect compared to the fuzzy c-means algorithm.