Clustering ensemble has been progressively popular in the ongoing years by combining several base clustering methods into a most likely better and increasingly robust one. Nonetheless, fuzzy clustering dependability (durability) has been unnoticed within the majority of the proposed clustering ensemble approach. This makes them weak against low-quality fuzzy base clusters. In spite of a few endeavors made to the clustering methods, it appears that they consider each base-clustering separately without considering its local diversity. In this paper, to compensate for the mentioned weakness a new fuzzy clustering ensemble approach has been proposed using a weighting strategy at fuzzy cluster level. Indeed, each fuzzy cluster has a contribution weight computed based on its reliability (dependability/durability). After computing fuzzy cluster dependability (reliability/durability), dependability based fuzzy cluster-wise weighted matrix (DFCWWM) is computed. As a final point, the final clustering is obtained by applying the FCM traditional clustering algorithm over DFCWWM. The time complexity of the proposed approach is linear in terms of the number of data-points. The proposed approach has been assessed on 15 various standard datasets. The experimental evaluation has indicated that the proposed method has better performance than the state-of-the-art methods.

中文翻译：

---

考虑集群可靠性的模糊聚类集成

通过将几种基本聚类方法组合成一种最有可能更好且越来越健壮的方法，聚类集成在过去的几年中逐渐流行起来。尽管如此，在大多数建议的聚类集成方法中，模糊聚类的可靠性（持久性）都没有被注意到。这使得它们对低质量的模糊基础集群很弱。尽管对聚类方法做出了一些努力，但似乎他们单独考虑了每个基聚类，而不考虑其局部多样性。在本文中，为了弥补上述不足，提出了一种新的模糊聚类集成方法，该方法使用模糊聚类级别的加权策略。实际上，每个模糊集群都有一个基于其可靠性（可靠性/耐用性）计算的贡献权重。在计算了模糊聚类的可靠性（可靠性/耐久性）之后，计算了基于可靠性的模糊聚类加权矩阵（DFCWWM）。作为最后一点，最终的聚类是通过在 DFCWWM 上应用 FCM 传统聚类算法获得的。所提出方法的时间复杂度与数据点的数量呈线性关系。所提出的方法已在 15 个不同的标准数据集上进行了评估。实验评估表明，所提出的方法比最先进的方法具有更好的性能。所提出方法的时间复杂度与数据点的数量呈线性关系。所提出的方法已在 15 个不同的标准数据集上进行了评估。实验评估表明，所提出的方法比最先进的方法具有更好的性能。所提出方法的时间复杂度与数据点的数量呈线性关系。所提出的方法已在 15 个不同的标准数据集上进行了评估。实验评估表明，所提出的方法比最先进的方法具有更好的性能。