In the previous chapter, we provide one concise description of some of the representative methods for clustering moderate-to-high-dimensional data, and summarize the analysis of the literature in Table 3.1. It allows us to identify two main desirable properties that are still missing from the existing techniques: (1) Linear or quasi-linear complexity and; (2) Terabyte-scale data analysis. Here we describe one work that focuses on tackling the former problem. Specifically, this chapter presents the new method Halite for correlation clustering [4, 5]. Halite is a novel correlation clustering method for multi-dimensional data, whose main strengths are that it is fast and scalable with regard to increasing numbers of objects and axes, besides increasing dimensionalities of the clusters. The following sections describe the new method in detail.
CITATION STYLE
Cordeiro, R. L. F., Faloutsos, C., & Traina Júnior, C. (2013). Halite. In SpringerBriefs in Computer Science (Vol. 0, pp. 33–67). Springer. https://doi.org/10.1007/978-1-4471-4890-6_4
Mendeley helps you to discover research relevant for your work.