Single Linkage Clustering Algorithm Research Articles

The Effect of Different Similarity Distance Measures in Detecting Outliers Using Single-Linkage Clustering Algorithm for Univariate Circular Biological Data

The procedure of outliers detection in univariate circular data can be developed using clustering algorithm. In clustering, it is necessary to calculate the similarity measure in order to cluster the observations into their own group. The similarity measure in circular data can be determined by calculating circular distance between each point of angular observation. In this paper, clustering-based procedure for outlier detection in univariate circular biological data with different similarity distance measures will be developed and the performance will be investigated. Three different circular similarity distance measures are used for the outliers detection procedure using single-linkage clustering algorithm. However, there are two similarity measures namely Satari distance and Di distance that are found to have similarity in formula for univariate circular data. The aim of this study is to develop and demonstrate the effectiveness of proposed clustering-based procedure with different similarity distance measure in detecting outliers. Therefore, in this study the circular similarity distance of SL-Satari/Di and another similarity measure namely SL-Chang will be compared at certain cutting rule. It is found that clustering-based procedure using single-linkage algorithm with different similarity distances are applicable and promising approach for outlier detection in univariate circular data, particularly for biological data. The result also found that at a certain condition of data, the SL-Satari/Di distance seems to overperform the performance of SL-Chang distance.

Scalable Event-Based Clustering of Social Media Via Record Linkage Techniques

We tackle the problem of grouping content available in social media applications such as Flickr, Youtube, Panoramino etc. into clusters of documents describing the same event. This task has been referred to as event identification before. We present a new formalization of the event identification task as a record linkage problem and show that this formulation leads to a principled and highly efficient solution to the problem. We present results on two datasets derived from Flickr — last.fm and upcoming — comparing the results in terms of Normalized Mutual Information and F-Measure with respect to several baselines, showing that a record linkage approach outperforms all baselines as well as a state-of-the-art system. We demonstrate that our approach can scale to large amounts of data, reducing the processing time considerably compared to a state-of-the-art approach. The scalability is achieved by applying an appropriate blocking strategy and relying on a Single Linkage clustering algorithm which avoids the exhaustive computation of pairwise similarities.

Natural scene text localization using edge color signature

Localizing text regions in images taken from natural scenes is one of the challenging problems dueto variations in font, size, color and orientation of text. In this paper, we introduce a new concept socalled Edge Color Signature for localizing text regions in an image. This method is able to localizeboth Farsi and English texts. In the proposed method rst a pyramid using different scales of theinput image is created. Then for each level of the pyramid an edge map is extracted. Afterward,several geometric features are employed to lter out the non-text edges from the extracted edges.At this stage we describe an edge using colors of its neighboring pixels. We use the mean-Shiftalgorithm to obtain the color modes surrounding each edge pixel. Subsequently, the connectededge pixels with similar color signatures are clustered using Single-Linkage clustering algorithm toconstruct meaningful groups. Finally, each of the clusters is labeled as text or non-text using anMLP based cascade classi er. The proposed method has been evaluated on well-known ICDAR 2013and our Farsi dataset, the result is very promising.

Transport anomalies emerging from strong correlation in ionic liquid electrolytes

Strong ionic interactions in concentrated ionic liquids is shown to result in significant correlations and deviations from ideal solution behavior. We use rigorous concentrated solution theory coupled with molecular dynamics simulations to compute and explain the unusual dependence of transport properties on cation concentration in the Na+-Pyr13+-FSI− ionic liquid electrolyte. When accounting for intra- and inter-species correlation, beyond the commonly used uncorrelated Nernst-Einstein equation, an anomalously low and even negative transference number emerges for xNaFSI lower than 0.2. With increasing concentration the transference number monotonically increases, approaching unity, while the total conductivity decreases as the system transitions to a state resembling a single-ion solid-state electrolyte. The degree of spatial ionic association is explored further by employing unsupervised single-linkage clustering algorithm. We emphasize that strong ion-ion coupling in the electrolyte significantly impacts the trade-off between key electrolyte transport properties, and consequently governs the power density of the battery.

An effective Sorensen‐single linkage clustering hybrid algorithm for cell formation problems in cellular manufacturing industry

SummaryThe paper presented here entails technique that is based upon Sorensen's similarity coefficient as well as the Sorensen‐ Single Linkage Clustering (SLC) algorithm, which essentially is a hybrid clustering approach that has been deployed here to show issue related to cell formation during the process of cellular engineering. In the following paper and study that has been presented here, the suggested hybrid algorithm comparison has been drawn versus the existing clustering algorithms and their performances have been compared specifically the MOD‐SLC modified single linkage or SLC also known as the single linkage clustering algorithm. The Sorensen‐SLC technique and the inferences drawn on the same show that it performs comparatively better when compared with the MOD‐SLC and SLC algorithms. Additionally, the computation that is necessary for this hybrid algorithm is absolutely a bare minimal and the computation processing is effective and easy.

Patterning of writing style evolution by means of dynamic similarity

This paper suggests a new methodology for patterning writing style evolution using dynamic similarity. We divide a text into sequential, disjoint portions (chunks) of the same size and exploit the Mean Dependence measure, aspiring to model the writing process via association between the current text chunk and its predecessors. To expose the evolution of a style, a new two-step clustering procedure is applied. In the first phase, a distance based on the Mean Dependence between each pair of chunks is evaluated. All document chunks in a pair are embedded in a high dimensional space using a Kuratowski-type embedding procedure and clustered by means of the introduced distance. In the next phase, the rows of the binary cluster classification documents matrix are clustered via the hierarchical single linkage clustering algorithm. By this way, a visualization of the inner stylistic structure of a texts’ collection, the resulting classification tree, is provided by the appropriate dendrogram. The approach applied to studying writing style evolution in the “Foundation Universe” by Isaac Asimov, the “Rama” series by Arthur C. Clarke, the “Forsyte Saga” of John Galsworthy, “The Lord of the Rings” by John Ronald Reuel Tolkien and a collection of books prescribed to Romain Gary demonstrates that the suggested methodology is capable of identifying style development over time. Additional numerical experiments with author determination and author verification tasks exhibit the high ability of the method to provide accurate solutions.

A novel specialized single-linkage clustering algorithm for taxonomically ordered data

Similarities among ortholog genes for a given set of species S can be expressed by alignment matrices, where each matrix cell results from aligning a gene transcript against the genome of a species within S. Gene clusters can be computed by using single-linkage clustering in time n × m, where n denotes the number of ortholog genes and m denotes the number of inspected assemblies. Our approach can break the O(n × m) complexity of single-linkage clustering by exploiting an order among species that results from an in-order traversal of a given phylogenetic tree. The order among species allows the reduction of the inspected scope of the matrix to taxonomically related combinations of assemblies and genes, thus lowering the computational efforts necessary for creating the alignment matrix without affecting cluster quality. We present two novel approaches for clustering. First, we introduce a hierarchical clustering with, omitting the initial sorting of |S| elements, amortized O(|S|) time behavior, where it holds |S|≤n+m. Then, we propose a consecutive clustering having a linear time complexity O(|S|). Both approaches compute identical clusters, whereas dendrograms can only be obtained from the hierarchical one. We prove that our approaches deliver higher cluster densities than single linkage clustering. Additionally, we show that we compute clusters of superior quality, which ensures that our approaches are generally less error prone.

A three-level optimization methodology for the partitioning of shale gas wellpad groups

Shale gas exploitation incurs considerable financial risk due to the high uncertainty of gas production; reducing the full life-cycle costs would play an important role in achieving expected economic benefits. In this paper, a three-level optimization methodology is proposed to optimally partition shale gas wellpad groups with the objective of minimizing the construction costs of the gas-gathering system. The first-level model is built to determine the annual optimal infield battery sites and the subordinated relationship among wellpads and infield batteries, aiming to minimize the construction costs of the gas-gathering system for each development year. The second-level model is proposed to determine the global optimal infield battery sites for the entire field life by using the hierarchical clustering method to divide all annual optimal infield battery sites into several limited clusters. The third-level model is established to determine a globally optimal subordinated relationship among wellpads and infield battery sites, with the objective of minimizing the total distances from the wellpads to the global optimal infield battery sites. The unconventional features, i.e., continuous tie-in, production decline and shut-down or abandonment of shale gas wellpads during field development, have been considered and integrated into the three-level methodology. The hybrid genetic algorithm and particle swarm optimization algorithm (HGAPSO) is used to solve the first- and third-level models, whereas the single-linkage clustering algorithm is adopted to solve the second-level model. Finally, a practical application of the methodology is performed to validate its effectiveness in partitioning wellpad groups in a real shale gas field in Sichuan Province, China. The achievements provide an effective method for partitioning shale gas wellpads in either the conceptual design or the following realistic development phases.

Genie: A new, fast, and outlier-resistant hierarchical clustering algorithm

The time needed to apply a hierarchical clustering algorithm is most often dominated by the number of computations of a pairwise dissimilarity measure. Such a constraint, for larger data sets, puts at a disadvantage the use of all the classical linkage criteria but the single linkage one. However, it is known that the single linkage clustering algorithm is very sensitive to outliers, produces highly skewed dendrograms, and therefore usually does not reflect the true underlying data structure – unless the clusters are well-separated. To overcome its limitations, we propose a new hierarchical clustering linkage criterion called Genie. Namely, our algorithm links two clusters in such a way that a chosen economic inequity measure (e.g., the Gini- or Bonferroni-index) of the cluster sizes does not increase drastically above a given threshold. The presented benchmarks indicate a high practical usefulness of the introduced method: it most often outperforms the Ward or average linkage in terms of the clustering quality while retaining the single linkage speed. The Genie algorithm is easily parallelizable and thus may be run on multiple threads to speed up its execution further on. Its memory overhead is small: there is no need to precompute the complete distance matrix to perform the computations in order to obtain a desired clustering. It can be applied on arbitrary spaces equipped with a dissimilarity measure, e.g., on real vectors, DNA or protein sequences, images, rankings, informetric data, etc. A reference implementation of the algorithm has been included in the open source genie package for R.

Auto Secured Text Monitor in Natural Scene Images

Monitoring the text in content based images is an important task. An accurate and robust method to monitor texts in natural scene images. In this paper fast and effective decision tree algorithm (MSERs) is used to extract a character participant by reducing regularized character dissimilarities. Character participant are grouped into text participant by using single-link clustering algorithm. Text participant with high non-text probabilities are removed and texts are establish with a text classifier. Valuable information may be broken in many content-based images, such as web images. We need to detect text before recognizing and retrieving.

RLT-S: A Web System for Record Linkage.

BackgroundRecord linkage integrates records across multiple related data sources identifying duplicates and accounting for possible errors. Real life applications require efficient algorithms to merge these voluminous data sources to find out all records belonging to same individuals. Our recently devised highly efficient record linkage algorithms provide best-known solutions to this challenging problem.MethodWe have developed RLT-S, a freely available web tool, which implements our single linkage clustering algorithm for record linkage. This tool requires input data sets and a small set of configuration settings about these files to work efficiently. RLT-S employs exact match clustering, blocking on a specified attribute and single linkage based hierarchical clustering among these blocks.ResultsRLT-S is an implementation package of our sequential record linkage algorithm. It outperforms previous best-known implementations by a large margin. The tool is at least two times faster for any dataset than the previous best-known tools.ConclusionsRLT-S tool implements our record linkage algorithm that outperforms previous best-known algorithms in this area. This website also contains necessary information such as instructions, submission history, feedback, publications and some other sections to facilitate the usage of the tool.AvailabilityRLT-S is integrated into http://www.rlatools.com, which is currently serving this tool only. The tool is freely available and can be used without login. All data files used in this paper have been stored in https://github.com/abdullah009/DataRLATools. For copies of the relevant programs please see https://github.com/abdullah009/RLATools.

Robust Text Detection in Natural Scene Images.

Text detection in natural scene images is an important prerequisite for many content-based image analysis tasks. In this paper, we propose an accurate and robust method for detecting texts in natural scene images. A fast and effective pruning algorithm is designed to extract Maximally Stable Extremal Regions (MSERs) as character candidates using the strategy of minimizing regularized variations. Character candidates are grouped into text candidates by the single-link clustering algorithm, where distance weights and clustering threshold are learned automatically by a novel self-training distance metric learning algorithm. The posterior probabilities of text candidates corresponding to non-text are estimated with a character classifier; text candidates with high non-text probabilities are eliminated and texts are identified with a text classifier. The proposed system is evaluated on the ICDAR 2011 Robust Reading Competition database; the f-measure is over 76%, much better than the state-of-the-art performance of 71%. Experiments on multilingual, street view, multi-orientation and even born-digital databases also demonstrate the effectiveness of the proposed method.

Constraint consensus concentration for identifying disjoint feasible regions in nonlinear programmes

It is usually not known in advance whether a nonlinear set of constraints has zero, one, or multiple feasible regions. Further, if one or more feasible regions exist, their locations are usually unknown. We propose a method for exploring the variable space quickly, using constraint consensus (CC) to identify promising areas that may contain a feasible region. Multiple CC solution points are clustered to identify regions of attraction. A new inter-point distance frequency distribution technique is used to determine the critical distance for the single-linkage clustering algorithm, which in turn determines the estimated number of disjoint feasible regions. The effectiveness of multistart global optimization is increased due to better exploration of the variable space, and efficiency is also increased because the expensive local solver is launched just once near each identified feasible region. The method is demonstrated on a variety of highly nonlinear models.

Pattern Recognition Algorithms for Cluster Identification Problem

There are large datasets in high dimensional database to solve the cluster identification problem and for identifying dense clusters in a noisy data. In this paper by using densely intra connected sub graphs our analysis identified the clusters. We have used pattern recognition algorithms for representing the graphs and solve the cluster identification problem using K-Means-Mode, single linkage clustering and K-Nearest Neighbor Algorithm. The computational analysis show that when running on 160 CPU’s, one of our algorithm can solve a cluster identification problem on a dataset with 1,000,000 data points almost 1000 times faster than on single CPU, indicating that the problem for handling large data is done in an efficient manner.

Interactive visualization of multi-resolution urban building models considering spatial cognition

Multi-resolution visualization of massive urban buildings is one of the most important components for a cyber city. Urban is a highly humanized system, and thus visualizing urban buildings needs to abide by people's habits of cognizing spatial relations between objects for their accurate and quick understanding of urban spatial information. This article proposes an approach to generalize and render urban building models in the context of Gestalt psychology and urban legibility. We introduce a new distance measurement method as the distance metric for the single-link clustering algorithm, which is used to group building footprints into clusters. Each cluster is merged based on the Delaunay triangulation and the polyline generalization algorithm. We then construct a hierarchical tree to store multi-resolution building models and implement interactive three-dimensional visualization of large-scale and high-density urban buildings. Experimental results indicate that the proposed methodology not only reduces the geometric complexity of urban models but also preserves urban legibility successfully and follow Gestalt principles.

Clustering of European countries based on ISO 9000 certification diffusion

PurposeOne of the stated goals behind the formation of the International Organization for Standardization was to develop standards to facilitate international trade in goods and services. This goal has caused different country‐group trends in the diffusion of ISO certification. The purpose of this paper is to create a taxonomy of ISO 9000 certification diffusion in Europe.Design/methodology/approachThe diffusion of ISO certificates in Europe was analyzed using the single linkage clustering algorithm. Then the features of each cluster were highlighted.FindingsEuropean nations were found to belong to three macro areas which differ for patterns of ISO 9000 certification diffusion: the constant growth area, the saturation area and the decline area.Research limitations/implicationsThe present study deals with European nations, focusing on ISO 9000 standards. Future research will be addressed toward the analysis of other standards in a wider geographical area in order to see if the results found for the European reality may be extended. In addition the outcomes of such analysis may be used to enhance the existing models for the diffusion of ISO 9000 certification.Practical implicationsThe analysis of the evolution of the certification market may arouse the interest of both companies and certification bodies that, from this study, may gain an insight into the future possible demand of certification. Furthermore, this study can be interesting from a legislature point of view, providing an answer to a more general question: “What is the general life cycle of standards or regulations?”Originality/valueThe novelty of the paper is the clustering approach that allows the grouping of different nations according to similar diffusion dynamics.

A note on single-linkage equivalence

We introduce the concept of single-linkage equivalence of edge-weighted graphs, we apply it to characterise maximal spanning trees and “ultra-similarities”, and we discuss how it relates to the popular single-linkage clustering algorithm.

Dynamics and Structure of the 30 Largest North American Companies

In this paper we describe a method to analyze the structure and dynamics of the 30 largest North American companies. The method combines the tools of symbolic time series analysis (Daw et al. in Rev Sci Instrum 74:916---930, 2003) with the nearest neighbor single linkage clustering algorithm (Mantegna and Stanley in An introduction to econophysics: Correlations and complexity in finance, Cambridge University Press, UK, 2000). Data symbolization allows to obtain a metric distance between two different time series that is used to construct a minimal spanning tree allowing to compute an ultrametric distance. From the analysis of time series data of companies included in Dow Jones Industrial Average, we derive a hierarchical organization of these companies. In particular, we detect different clusters of companies which correspond with their common production activities or their strong interrelationship. The obtained classification of companies can be used to study deep relationships among different branch of economic activities and to construct financial portfolios.

The international hotel industry in Spain: Its hierarchical structure

This paper empirically analyzes the indebtedness of the main companies in the Spanish hotel industry. In particular, it introduces a method to analyze the structures and dynamics of the largest companies in this sector. The method combines the Pearson correlation coefficient with the nearest neighbour single linkage clustering algorithm (Mantegna, R. N. (1999). Hierarchical structure in financial markets. The European Physical Journal B, 11, 193–197). The Pearson correlation coefficient determines a metric distance between two different multidimensional time series that is then used to construct a Minimal Spanning Tree. From this tree, an ultra metric distance that is used to derive the Hierarchical Tree can be computed. From the analysis of the time series data of companies included in the SABI, or the Iberian Balance Sheet Analysis System 2005, it is possible to derive a hierarchical organization of the Spanish hotel firms. In particular, different dynamic clusters of companies, which correspond to their common production and indebtedness strategies, can be detected.

PhyloPat: an updated version of the phylogenetic pattern database contains gene neighborhood

Phylogenetic patterns show the presence or absence of certain genes in a set of full genomes derived from different species. They can also be used to determine sets of genes that occur only in certain evolutionary branches. Previously, we presented a database named PhyloPat which allows the complete Ensembl gene database to be queried using phylogenetic patterns. Here, we describe an updated version of PhyloPat which can be queried by an improved web server. We used a single linkage clustering algorithm to create 241 697 phylogenetic lineages, using all the orthologies provided by Ensembl v49. PhyloPat offers the possibility of querying with binary phylogenetic patterns or regular expressions, or through a phylogenetic tree of the 39 included species. Users can also input a list of Ensembl, EMBL, EntrezGene or HGNC IDs to check which phylogenetic lineage any gene belongs to. A link to the FatiGO web interface has been incorporated in the HTML output. For each gene, the surrounding genes on the chromosome, color coded according to their phylogenetic lineage can be viewed, as well as FASTA files of the peptide sequences of each lineage. Furthermore, lists of omnipresent, polypresent, oligopresent and anticorrelating genes have been included. PhyloPat is freely available at http://www.cmbi.ru.nl/phylopat.