Graph Clustering Methods Research Articles

A novel dynamic minimum spanning tree based clustering method for image mining

Image mining is more than just an extension of data mining to image domain. Image mining is a technique commonly used to extract knowledge directly from image. Image segmentation is the first step in image mining. Image segmentation is very difficult but important problem in computer vision. We present a novel Minimum Spanning Tree based approach to image segmentation which can be applied to either greyscale or color image. The assumption is the nearby pixels with similar color or greyscale intensities may belong to the same region or segment of the image. A graph representation for an image is derived from the similarity between the pixels, and then partitioned by computationally efficient graph clustering method. We treat image segmentation as graph partitioning problem. The minimum spanning tree based clustering algorithm is capable of detecting clusters with irregular boundaries. In this paper we propose minimum spanning tree based clustering algorithm using weighted Euclidean distance for edges, which is key element in building the graph from image. The algorithm produces best number of clusters (regions) with segments. An important characteristic of the method is its ability to preserve detail in low-variability image regions while preserving detail in high-variability regions.

Extracting Between-Pathway Models from E-MAP Interactions Using Expected Graph Compression

Abstract Genetic interactions (such as synthetic lethal interactions) have become quantifiable on a large-scale using the epistatic miniarray profile (E-MAP) method. An E-MAP allows the construction of a large, weighted network of both aggravating and alleviating genetic interactions between genes. By clustering genes into modules and establishing relationships between those modules, we can discover compensatory pathways. We introduce a general framework for applying greedy clustering heuristics to probabilistic graphs. We use this framework to apply a graph clustering method called graph summarization to an E-MAP that targets yeast chromosome biology. This results in a new method for clustering E-MAP data that we call Expected Graph Compression (EGC). We validate modules and compensatory pathways using enriched Gene Ontology annotations and a novel method based on correlated gene expression. EGC finds a number of modules that are not found by any previous methods to cluster E-MAP data. EGC also uncovers core submodules contained within several previously found modules, suggesting that EGC can reveal the finer structure of E-MAP networks.

Clustering Large Attributed Graphs

Social networks, sensor networks, biological networks, and many other information networks can be modeled as a large graph. Graph vertices represent entities, and graph edges represent their relationships or interactions. In many large graphs, there is usually one or more attributes associated with every graph vertex to describe its properties. In many application domains, graph clustering techniques are very useful for detecting densely connected groups in a large graph as well as for understanding and visualizing a large graph. The goal of graph clustering is to partition vertices in a large graph into different clusters based on various criteria such as vertex connectivity or neighborhood similarity. Many existing graph clustering methods mainly focus on the topological structure for clustering, but largely ignore the vertex properties, which are often heterogenous. In this article, we propose a novel graph clustering algorithm, SA-Cluster , which achieves a good balance between structural and attribute similarities through a unified distance measure. Our method partitions a large graph associated with attributes into k clusters so that each cluster contains a densely connected subgraph with homogeneous attribute values. An effective method is proposed to automatically learn the degree of contributions of structural similarity and attribute similarity. Theoretical analysis is provided to show that SA-Cluster is converging quickly through iterative cluster refinement. Some optimization techniques on matrix computation are proposed to further improve the efficiency of SA-Cluster on large graphs. Extensive experimental results demonstrate the effectiveness of SA-Cluster through comparisons with the state-of-the-art graph clustering and summarization methods.

Spectral methods for graph clustering – A survey

Graph clustering is an area in cluster analysis that looks for groups of related vertices in a graph. Due to its large applicability, several graph clustering algorithms have been proposed in the last years. A particular class of graph clustering algorithms is known as spectral clustering algorithms. These algorithms are mostly based on the eigen-decomposition of Laplacian matrices of either weighted or unweighted graphs. This survey presents different graph clustering formulations, most of which based on graph cut and partitioning problems, and describes the main spectral clustering algorithms found in literature that solve these problems.

Recognition of crowd behavior from mobile sensors with pattern analysis and graph clustering methods

Mobile on-body sensing has distinct advantages for the analysis and understanding of crowd dynamics: sensing is not geographically restricted to a specific instrumented area, mobile phones offer on-body sensing and they are already deployed on a large scale, and the rich sets of sensors they contain allows one to characterize the behavior of users through pattern recognition techniques. In this paper we present a methodological framework for the machine recognition of crowd behavior from on-body sensors, such as those in mobile phones. The recognition of crowd behaviors opens the way to the acquisition of large-scale datasets for the analysis and understanding of crowd dynamics. It has also practical safety applications by providing improved crowd situational awareness in cases of emergency. The framework comprises: behavioral recognition with the user's mobile device, pairwise analyses of the activity relatedness of two users, and graph clustering in order to uncover globally, which users participate in a given crowd behavior. We illustrate this framework for the identification of groups of persons walking, using empirically collected data. We discuss the challenges and research avenues for theoretical and applied mathematics arising from the mobile sensing of crowd behaviors.

Protein complex detection with semi-supervised learning in protein interaction networks

BackgroundProtein-protein interactions (PPIs) play fundamental roles in nearly all biological processes. The systematic analysis of PPI networks can enable a great understanding of cellular organization, processes and function. In this paper, we investigate the problem of protein complex detection from noisy protein interaction data, i.e., finding the subsets of proteins that are closely coupled via protein interactions. However, protein complexes are likely to overlap and the interaction data are very noisy. It is a great challenge to effectively analyze the massive data for biologically meaningful protein complex detection.ResultsMany people try to solve the problem by using the traditional unsupervised graph clustering methods. Here, we stand from a different point of view, redefining the properties and features for protein complexes and designing a “semi-supervised” method to analyze the problem. In this paper, we utilize the neural network with the “semi-supervised” mechanism to detect the protein complexes. By retraining the neural network model recursively, we could find the optimized parameters for the model, in such a way we can successfully detect the protein complexes. The comparison results show that our algorithm could identify protein complexes that are missed by other methods. We also have shown that our method achieve better precision and recall rates for the identified protein complexes than other existing methods. In addition, the framework we proposed is easy to be extended in the future.ConclusionsUsing a weighted network to represent the protein interaction network is more appropriate than using a traditional unweighted network. In addition, integrating biological features and topological features to represent protein complexes is more meaningful than using dense subgraphs. Last, the “semi-supervised” learning model is a promising model to detect protein complexes with more biological and topological features available.

IDENTIFICATION OF FUNCTIONAL MODULES IN A PPI NETWORK BY BOUNDED DIAMETER CLUSTERING

Dense subgraphs of Protein-Protein Interaction (PPI) graphs are assumed to be potential functional modules and play an important role in inferring the functional behavior of proteins. Increasing amount of available PPI data implies a fast, accurate approach of biological complex identification. Therefore, there are different models and algorithms in identifying functional modules. This paper describes a new graph theoretic clustering algorithm that detects densely connected regions in a large PPI graph. The method is based on finding bounded diameter subgraphs around a seed node. The algorithm has the advantage of being very simple and efficient when compared with other graph clustering methods. This algorithm is tested on the yeast PPI graph and the results are compared with MCL, Core-Attachment, and MCODE algorithms.

A garter snake transcriptome: pyrosequencing, de novo assembly, and sex-specific differences

BackgroundThe reptiles, characterized by both diversity and unique evolutionary adaptations, provide a comprehensive system for comparative studies of metabolism, physiology, and development. However, molecular resources for ectothermic reptiles are severely limited, hampering our ability to study the genetic basis for many evolutionarily important traits such as metabolic plasticity, extreme longevity, limblessness, venom, and freeze tolerance. Here we use massively parallel sequencing (454 GS-FLX Titanium) to generate a transcriptome of the western terrestrial garter snake (Thamnophis elegans) with two goals in mind. First, we develop a molecular resource for an ectothermic reptile; and second, we use these sex-specific transcriptomes to identify differences in the presence of expressed transcripts and potential genes of evolutionary interest.ResultsUsing sex-specific pools of RNA (one pool for females, one pool for males) representing 7 tissue types and 35 diverse individuals, we produced 1.24 million sequence reads, which averaged 366 bp in length after cleaning. Assembly of the cleaned reads from both sexes with NEWBLER and MIRA resulted in 96,379 contigs containing 87% of the cleaned reads. Over 34% of these contigs and 13% of the singletons were annotated based on homology to previously identified proteins. From these homology assignments, additional clustering, and ORF predictions, we estimate that this transcriptome contains ~13,000 unique genes that were previously identified in other species and over 66,000 transcripts from unidentified protein-coding genes. Furthermore, we use a graph-clustering method to identify contigs linked by NEWBLER-split reads that represent divergent alleles, gene duplications, and alternatively spliced transcripts. Beyond gene identification, we identified 95,295 SNPs and 31,651 INDELs. From these sex-specific transcriptomes, we identified 190 genes that were only present in the mRNA sequenced from one of the sexes (84 female-specific, 106 male-specific), and many highly variable genes of evolutionary interest.ConclusionsThis is the first large-scale, multi-organ transcriptome for an ectothermic reptile. This resource provides the most comprehensive set of EST sequences available for an individual ectothermic reptile species, increasing the number of snake ESTs 50-fold. We have identified genes that appear to be under evolutionary selection and those that are sex-specific. This resource will assist studies on gene expression and comparative genomics, and will facilitate the study of evolutionarily important traits at the molecular level.

A Search Algorithm for Clusters in a Network or Graph

A novel breadth-first based structural clustering method for graphs is proposed. Clustering is an important task for analyzing complex networks such as biological networks, World Wide Web and social networks. The clusters are of various shapes such as cliques and stars, for example, in ProteinProtein Interactive(PPI) networks. Traditional algorithms may detect clique-shaped clusters, but they fail to identify star-shaped clusters that are common in scale free networks, including PPI networks. We propose a novel clustering algorithm to solve the problem. Experimental results demonstrate it outperforms other algorithms in one or several aspects: Detecting clusters of mixed shapes, including both cliques and stars; Faster. Its running time on a network with n nodes and m links is O(n), which is much faster than O(mdlogn) of the fastest modularity-based algorithm(where d is the depth of the dendrogram describing the hierarchical cluster structure); A non-parametric algorithm. It can accomplish all goals without requiring any input parameters.

Ports in multi-level maritime networks: evidence from the Atlantic (1996–2006)

While maritime transport ensures about 90% of world trade volumes, it has not yet attracted as much attention as other transport systems from a graph perspective. As a result, the relative situation and the evolution of seaports within maritime networks are not well understood. This paper wishes verifying to what extent the hub-and-spoke strategies of ports and ocean carriers have modified the structure of a maritime network, based on the Atlantic case. We apply graph measures and clustering methods on liner movements in 1996 and 2006. The methodology also underlines which ports are increasing their position by carriers’ circulation patterns on various scales. This research demonstrates that the polarization of the Atlantic network by few dominant ports occurs in parallel with the increased spatial integration of this area by shipping lines.

Optimizing an organized modularity measure for topographic graph clustering: A deterministic annealing approach

This paper proposes an organized generalization of Newman and Girvan's modularity measure for graph clustering. Optimized via a deterministic annealing scheme, this measure produces topologically ordered graph clusterings that lead to faithful and readable graph representations based on clustering induced graphs. Topographic graph clustering provides an alternative to more classical solutions in which a standard graph clustering method is applied to build a simpler graph that is then represented with a graph layout algorithm. A comparative study on four real world graphs ranging from 34 to 1133 vertices shows the interest of the proposed approach with respect to classical solutions and to self-organizing maps for graphs.

Graph clustering based on structural/attribute similarities

The goal of graph clustering is to partition vertices in a large graph into different clusters based on various criteria such as vertex connectivity or neighborhood similarity. Graph clustering techniques are very useful for detecting densely connected groups in a large graph. Many existing graph clustering methods mainly focus on the topological structure for clustering, but largely ignore the vertex properties which are often heterogenous. In this paper, we propose a novel graph clustering algorithm, SA-Cluster , based on both structural and attribute similarities through a unified distance measure. Our method partitions a large graph associated with attributes into k clusters so that each cluster contains a densely connected subgraph with homogeneous attribute values. An effective method is proposed to automatically learn the degree of contributions of structural similarity and attribute similarity. Theoretical analysis is provided to show that SA-Cluster is converging. Extensive experimental results demonstrate the effectiveness of SA-Cluster through comparison with the state-of-the-art graph clustering and summarization methods.

Graph clustering

In this survey we overview the definitions and methods for graph clustering, that is, finding sets of “related” vertices in graphs. We review the many definitions for what is a cluster in a graph and measures of cluster quality. Then we present global algorithms for producing a clustering for the entire vertex set of an input graph, after which we discuss the task of identifying a cluster for a specific seed vertex by local computation. Some ideas on the application areas of graph clustering algorithms are given. We also address the problematics of evaluating clusterings and benchmarking cluster algorithms.

2A1-J08 二部グラフ分割問題の順序最適化としての解法(進化・学習とロボティクス)

There're a various networks which can represent bipartite graph in the real world. Most of the unipartite graph clustering methods have been researched but few of the bipartite graph. The bipartite graph clustering methods have been implemented by the unipartite folding. But this approach results in lack of link information, and existing approaches have a large computation orders. In this paper, we propose the bipartite clustering by TSP solution. We propose three cost function definitions with vertex adjacency vectors. Then, we apply the solution to computer-generated graphs with Local Clustering Organization(LCO). As results of the experiment, we obtained clusters by cutting high cost threshold, and we found that the cost function based on cosine construct better clusters.

Graph Clustering and Minimum Cut Trees

In this paper, we introduce simple graph clustering methods based on minimum cuts within the graph. The clustering methods are general enough to apply to any kind of graph but are well suited for graphs where the link structure implies a notion of reference, similarity, or endorsement, such as web and citation graphs. We show that the quality of the produced clusters is bounded by strong minimum cut and expansion criteria. We also develop a framework for hierarchical clustering and present applications to real-world data. We conclude that the clustering algorithms satisfy strong theoretical criteria and perform well in practice.

Graph- and multigrafh-theoretic partitioning clustering algorithms for large data sets

This work aims at introducing two clustering algorithms conceived to make the application of particular graph and multigraph theoretic clustering methods possible for large data sets. Results obtained by applying the suggested algorithms to simulated data sets are presented.

Character analysis, phylogeny, and classification of the moss genus Pohlia

Forty-one Pohlia taxa coded for their expression of 36 qualitative morphological characters are analyzed using cladistic and phenetic methods. Phenetic approaches include similarity graph clustering and divisive methods utilizing information theoretic criteria. The Wagner Tree parsimony method and character compatibility analysis provide estimates of cladistic relationships. Four major lineages within the genus differ primarily in sporophyte characters, with specific differences based largely on gametophytic features. Peristome simplification may have occurred several times within Pohlia. An infrageneric classification is proposed with diagnostic descriptions and nomenclature provided.