具有数据选择偏差的去相关聚类
Decorrelated Clustering with Data Selection Bias
摘要：
现有的聚类算法大多没有考虑数据的选择偏差。然而，在许多实际应用中，无法保证数据是无偏的。选择偏差可能会导致特征之间的非预期相关性，忽略这些非预期相关性会影响聚类算法的性能。因此，如何消除这些意想不到的相关性诱导的选择偏差是非常重要的，但在很大程度上尚未探索的聚类。本文提出了一种新的去相关正则化K-Means算法（DCKM），用于数据选择偏差的聚类。具体地说，去相关正则化器旨在学习能够平衡样本分布的全局样本权重，以便去除特征之间的意外相关性。同时，将学习的权重与k-means相结合，使得重新加权的k-means聚类在固有的数据分布上，而不会产生意外的相关性影响。此外，我们推导出更新规则，有效地推断DCKM中的参数。在真实世界数据集上的大量实验结果表明，我们的DCKM算法取得了显着的性能增益，表明在聚类时，消除由选择偏差引起的意外特征相关性的必要性。
英文摘要：
Most of existing clustering algorithms are proposed without considering the selection bias in data. In many real applications, however, one cannot guarantee the data is unbiased. Selection bias might bring the unexpected correlation between features and ignoring those unexpected correlations will hurt the performance of clustering algorithms. Therefore, how to remove those unexpected correlations induced by selection bias is extremely important yet largely unexplored for clustering. In this paper, we propose a novel Decorrelation regularized K-Means algorithm (DCKM) for clustering with data selection bias. Specifically, the decorrelation regularizer aims to learn the global sample weights which are capable of balancing the sample distribution, so as to remove unexpected correlations among features. Meanwhile, the learned weights are combined with k-means, which makes the reweighted k-means cluster on the inherent data distribution without unexpected correlation influence. Moreover, we derive the updating rules to effectively infer the parameters in DCKM. Extensive experiments results on real world datasets well demonstrate that our DCKM algorithm achieves significant performance gains, indicating the necessity of removing unexpected feature correlations induced by selection bias when clustering.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
