CAJ | 학술논문

相对于抑制式模糊C-均值聚类算法，半抑制式模糊C-均值聚类算法引入抑制门限，在保持收敛速度变化不大的前提下改善了聚类的效果。然而，该算法也存在参数选择较多的问题。针对该问题，我们引入面向隶属度修正的模糊聚类参数选择方法，提出了改进的半抑制式模糊C-均值聚类算法。实验表明，该算法具有较好的可操作性，在具有较快收敛速度的同时，也维持了较好的收敛效果。
상대우억제식모호C-균치취류산법，반억제식모호C-균치취류산법인입억제문한，재보지수렴속도변화불대적전제하개선료취류적효과。연이，해산법야존재삼수선택교다적문제。침대해문제，아문인입면향대속도수정적모호취류삼수선택방법，제출료개진적반억제식모호C-균치취류산법。실험표명，해산법구유교호적가조작성，재구유교쾌수렴속도적동시，야유지료교호적수렴효과。
Comparison of suppressed fuzzy C-means clustering algorithm, half suppressed fuzzy C-means clustering algorithm introduces suppression threshold parameter, which improves the clustering effect while keeping with little change in convergence speed. However, the algorithm has the problem of parameter selection. To solve the problem, an parameter selection method for mem-bership correction fuzzy clustering is introduced, and the improved half suppressed fuzzy C- means clustering algorithm is put for-ward. Experiments show that, the new algorithm has a better operability, faster convergence speed, and a good convergence effect.