【摘要】 由聚类目标函数导出求解最佳模糊聚类中心的迭代模型,提出判定多维样本聚类归并情况的贴近度指标;为了进一步消除冗余“潮汐力”等不利因素给归并效果带来的偏差,建立了冗余聚类判定及剔除法则·实例表明,该方法可任意控制归并程度,能节约大量钢材,比传统归并方法更有效·更多还原

【Abstract】 Iterative model for finding the optimal fuzzy clustering centers by objective function is presented, and the criterion of closeness for evaluating the result of fuzzy merging from multidimensional samples is provided. In order to delete the deviation that the redundant tide strength brings to the merging result, the principle of recognizing and rejecting redundant clusters is proposed. Practical examples show this method is more effective than the traditional way in controlling the granularity of merged classes and saving a lot of steel reinforcement.更多还原