几个常用的聚类任务总结
is basically a collection of objects on the basis of similarity and dissimilarity between them. (For single-view data)
传统的聚类任务就是我们说的clustering,其主要思想是利用数据表示之间的相似性和差异性对没有标记的数据进行组别的划分。当未指明数据类型时,一般默认针对单视图数据任务。
is a machine learning paradigm to classify similar subjects into the same group and dissimilar subjects into different groups by combining the available multi-view feature information, and to search for consistent clusterings across different views.
多视图聚类任务是区别于传统聚类任务的,其特点是,一个样本拥有多种不同的表示方式,这多种表示方式被称为描述这一样本的多个视图。其目标是利用多视图之间的相似性和互补性,在不同视图间搜索一致的聚类结果。
aims at exploring alternative clusterings to organize the data into meaningful groups from different perspectives. (For single-view data)
我们平时关注的clustering,其目的是利用数据的特征信息或者对象信息得到一致的聚类结果。但是,用户很难确定给定的解决方案是否确实适合当前的问题,尤其是在数据集很大且很复杂的情况下。此外,可能存在许多聚类方案,并且每个聚类解决方案都可以对数据进行有意义的划分,例如,在生物信息学中,可以针对结构或功能对蛋白质进行聚类,并且可以根据基因在不同实验条件下的表达谱对蛋白质进行聚类。通俗点来说,在这些情况下,希望向用户呈现多个替代集群供选择,并且这些集群应该是高质量的并且彼此不同。而得到这些集群的方法和过程就是multiple clustering。在未说明的情况下,默认针对的是单视图数据。
assumes that the individuality and commonality of multi-view data can be leveraged to generate high-quality and diverse clusterings.
是multiple clustering的多视图版本,多了multiple的修饰,整个任务的目的不再是利用不同视图的信息搜索一致性的聚类结果,而是去寻找多种更具有可解释性的分组结果。
is a pair of maps from rows to row-clusters and from columns to column-clusters. Co-clustering (a.k.a. bi-clustering, projective clustering, or two way clustering) can simultaneously group objects and features to explore local (or block) patterns in the data matrix.
这里贴一张来自ML Wiki的解释:
co-clustering.png当然这种技术也可以应用在multiple clustering上。
Existing solutions to multiple clustering mainly focus on one way clustering, that is , they cluster either the samples or the features. However, for many practical tasks, it is meaningful and desirable to explore alternative two-way clusterings (or co-clusterings), which capture not only the sample cluster structure but also the feature cluster structure.
现有的多聚类解决方案主要集中在一种聚类方式上,即对样本或特性进行聚类。然而,对于许多实际任务来说,探索可选的双向聚类(或共聚类)是有意义和可取的,它不仅能捕获样本聚类结构,还能捕获特征聚类结构。
To tackle this interesting and unresolved task, is proposed, to generate multiple alternative co-clusterings at the same time. (For single-view data)
也许在后续的任务中随着构造数据矩阵的方法的突破,也可能出现multi-view multiple co-clustering。
网友评论