Survey of Clustering Algorithms

Abstract

Abstract-Data analysis plays an indispensable role for understanding various phenomena. Cluster analysis, primitive exploration with little or no prior knowledge, consists of research developed across a wide variety of communities. The diversity, on one hand, equips us with many tools. On the other hand, the profusion of options causes confusion. We survey clustering algorithms for data sets appearing in statistics, computer science, and machine learning, and illustrate their applications in some benchmark data sets, the traveling salesman problem, and bioinformatics, a new field attracting intensive efforts. Several tightly related topics, proximity measure, and cluster validation, are also discussed.Index Terms-Adaptive resonance theory (ART), clustering, clustering algorithm, cluster validation, neural networks, proximity, self-organizing feature map (SOFM).

Traffic accident segmentation by means of latent class clustering

“…to identify clusters which can be transferred to other data sets. In such research, it is important to validate the cluster results to provide a degree of confidence (Xu and Wunsch, 2005). One could even try to find a generic list of traffic accident types which can be used to reduce heterogeneity in general, although it is not certain such list exists.…”

“…Finite mixture models have been implemented in different software packages, such as MCLUST, GMDD, AutoClass, Multimix, EMMX, SNOB (Xu and Wunsch, 2005) and Latent Gold. All these software packages use the finite mixture model expressed by equation 1, but differ in regard to the implemented algorithm and probability distributions for ) , (…”

“…The remainder of this section will focus on the model-based clustering algorithm as implemented by this particular piece of software. For a more complete discussion and review of existing clustering techniques, the interested reader is referred to some classical texts as Gordon (1999) or Kaufman and Rousseeuw (2005) and the elaborated survey paper by Xu and Wunsch (2005). It should be noted that the goal of this paper is to investigate whether clustering techniques can be used to reduce traffic accident heterogeneity.…”

A More Relaxed Model for Graph-Based Data Clustering: s-Plex Editing

“…There are numerous approaches to clustering and "there is no clustering algorithm that can be universally used to solve all problems" [16]. To solve data clustering, one prominent line of attack is to use graph theory based methods [14].…”

A methodology for dealing with spatial big data

“…In particular spatial clustering algorithms, which group similar spatial objects into classes, can be used for knowledge discovery in spatial databases, i.e., databases managing 2D or 3D points, polygons, etc., or points in some d-dimensional feature space (see, e.g., Fayyad et al, 1996;Han et al, 2001;Koperski et al, 1998;Steinbach et al, 2003;Xu and Wunsch, 2005). …”