Single Linkage Clustering Research Articles

Sunspot nests as traced by a cluster analysis

The appearance of sunspot groups between August 1959 and December 1964 is investigated in search of sunspot nests. A sunspot nest is a relatively small space on the surface within which a succession of spot groups appears. A single-linkage clustering technique is used to trace clusters in the three dimensions longitude, latitude and time. The number of genuine sunspot nests is estimated and their properties are determined.

Clustering algorithms for the design of a cellular manufacturing system - an analysis for their performance

Abstract The heuristic clustering methods based on similarity coefficient are considered to be very efficient for providing modularity and flexibility in a cellular manufacturing systems (CMS's). Various algorithms have been implemented in these heuristic methods. However, these algorithms suffer from string effect which is also known as “chaining”. Sedveral studies have reported this problem, yet not much research has been conducted to investigate its impact on actual clustering process. This paper presents results from an analytical study performed to determine the severity of chaining problem and other characteristics associated with the clustering process of four selected algorithms. The four algorithms are Single linkage clustering (SLINK), Average linkage clustering (ALINK), Weighted average linkage clustering (WLINK), and Complete linkage clustering (CLINK). A sample of fifty problems with randomly generated data sets was used to determine feasible solutions consisting of machine cells and corresponding part families from each of the four algorithms. A quantitative measure is proposed for evaluating the performance of different algorithms. The study concludes that the chaining effect for CLINK, WLINK, ALINK and SLINK progresively worsens from CLINK to SLINK in the same order. The study also provides important guidelines to designers of a CMS in selecting the most efficient algorithm for a given problem data. Several important statistical results are also presented.

Stoffwechselphysiologische Charakterisierung der Boraginaceen unter besonderer Berücksichtigung ökologischer Aspekte

It is well established that morphological and anatomical plasticities are accompanied by physiological plasticity (E pstein & J efferies 1964). There have been several attempts to describe and classify plants by physiological data. One resulted in the "physiotype-concept" of Kinzel (1972), who described a "physiotype" as the pattern of specific physiological properties common to a certain taxonomic unit. The Austrian species of the Boraginaceae have been investigated by analytical methods to work out, whether there exists physiological plasticity with respect to different ecological habitats within this plant family. The plants were collected at natural habitats and divided into 4 groups according to their ecological behaviour. The mineral ions, organic acids, carbohydrates and free amino acids have been estimated. Additionally, the component of the mucilage which is responsible for the "hard cell saps" (H ofmeister 1940) has been analysed. Cluster analysis has been applicated to the data using "UPGMA" and "single linkage clustering method". By use of the presented physiological data the plant species could not be assigned to the 4 ecological groups. Water soluble pectin, the main component of the mucilage, is functioning as anion and therefore part of the ion-balance. The quantity of water soluble pectin shows no significant differences at the 4 types of habitats, whereas arnrnoniumoxalate soluble - mostly cell wall bound - pectin is accumulated at dry sites. An interesting connection between cyclitols, uronic acids and fructans has been estimated and their possible role in adaption to different habitats is discussed. D-Bomesitol is proposed to act as "compatible solute" in these species. The family represents an uniform "physiotype".

Classification of the family Plectonemertidae (Nemertea): a phenetic comparison

A phenetic classification based on overall morphological similarity between the species in the family Plectonemertidae (genera Plectonemertes, Campbellonemertes, Potamonemertes, Leptonemertes, Katechonemertes, Argonemertes, Anliponemertes, and Acteonemertes) was undertaken and the result compared with a cladistic and an evolutionary classification. Similarity between species was computed by Gower's general coefficient of similarity and various techniques were used to find patterns in the similarity matrix: single-linkage, average-linkage, and complete-linkage clustering, together with principal coordinate analysis. Although the explicit aim of phenetics is not to estimate the phylogeny, the classification based on overall similarity still portrays phylogeny better than an intuitive assessment of morphological similarity, as judged by the cladistic analysis. The classification does not support the previously proposed hypothesis that the two freshwater genera Campbellonemertes and Potamonemertes have descended from a terrestrial ancestor.

Single linkage versus average linkage clustering in machine cells formation applications

Machine cells formation is the process of identifying families of similar parts (part-families) and forming the associated machine cells such that one or more part-families can be processed within a single cell. Among approaches to the machine cells formation problem the similarity coefficient method (SCM) has the capability of incorporating manufacturing data such as production volume, sequence of operations, and processing times into the machine cells formation process. Single Linkage Clustering (SLINK) and Average Linkage Clustering (ALC) are two widely used clustering techniques based on SCM. While SLINK is simple and has minimal computational requirement, it may generate machine cells in which a large number of machines are far apart in terms of similarity. ALC, on the other hand, overcomes this problem but requires more computations. This paper discusses the application of these two techniques to the machine cells formation problem.

EFFICIENCY OF HIERARCHIC AGGLOMERATIVE CLUSTERING USING THE ICL DISTRIBUTED ARRAY PROCESSOR

The implementation of hierarchic agglomerative methods of cluster anlaysis for large datasets is very demanding of computational resources when implemented on conventional computers. The ICL Distributed Array Processor (DAP) allows many of the scanning and matching operations required in clustering to be carried out in parallel. Experiments are described using the single linkage and Ward's hierarchical agglomerative clustering methods on both real and simulated datasets. Clustering runs on the DAP are compared with the most efficient algorithms currently available implemented on an IBM 3083 BX. The DAP is found to be 2.9–7.9 times as fast as the IBM, the exact degree of speed‐up depending on the size of the dataset, the clustering method, and the serial clustering algorithm that is used. An analysis of the cycle times of the two machines is presented which suggests that further, very substantial speed‐ups could be obtained from array processors of this type if they were to be based on more powerful processing elements.

Machine grouping — Expert systems: Comparison between single linkage and average linkage clustering techniques in forming machine cells

This paper compares the single linkage and average linkage clustering techniques when applied to the machine cells formation problem and points out the advantages and disadvantages of each technique.

Numerical Classification of some Thermus Isolates from Icelandic Hot Springs

A numerical classification was performed on a collection of 51 Thermus isolates from Icelandic hot pools, six reference strains including T. aquaticus YT-1, “ T. thermophilus ” HB8 and T. ruber , and on seven strains respresenting clusters produced by a previous numerical classification of Thermus isolates from New Zealand hot pools. Unweighted average linking (UPGMA) and single linkage clustering methods were applied to similarity matrices derived from simple matching (S SM ) and Jaccard similarity (S J ) coefficients. Differences were observed between the phenograms derived from the SSM and S J coefficients indicating that some of the clusters formed were derived from a significant component of negative matches. Test error was estimated at 6.4%. In the UPGMA/S SM phenogram four clusters were formed. A majority of the Icelandic isolates did not cluster with non-Icelandic isolates. Analysis of variance showed there to be a relationship between the composition of the clusters and the pH of the isolate source, but not to the isolate source temperature. Cluster composition did not appear to reflect the geographical sources within Iceland.

Complete linkage as a multiple stopping rule for single linkage clustering

Two commonly used clustering criteria are single linkage, which maximizes the minimum distance between clusters, and complete linkage, which minimizes the maximum distance within a cluster. By synthesizing these criteria, partitions of objects are sought which maximize a combined measure of the minimum distance between clusters and the maximum distance within a cluster. Each combined measure is shown to select a partition in the single linkage hierarchy. Therefore, in effect, complete linkage is used to provide a stopping rule for single linkage. An algorithm is outlined which uses the distance between each pair of objects twice only. To illustrate the method, an example is given using 23 Glamorganshire soil profiles.

Phylogenetic Systematics of Pontederiaceae

Six different methods of cladogram construction were applied to a data set embracing 34 species and subspecies of Pontederiaceae. Nineteen Wagner trees were constructed using dif- ferent combinations of character subsets and outgroups. Although 12 trees were constructed using the largest character cliques uncovered by character compatibility analysis, these were highly un- resolved. Four additional trees were constructed using phenetic methods. Trees resulting from the various procedures were compared using several optimality criteria. Wagner trees based on all 42 characters or on the subset of 27 reproductive characters were most successful according to these criteria. Branch-swapping trials, based upon common features found in the Wagner and other trees produced another 10 trees that were as successful as the Wagner trees. The number of about equally successful trees is partly a function of homoplasy. Two-thirds of the character state changes on the trees are homoplasious. The primitive breeding system of the Pontederiaceae is not clearly resolved, but heterostyly is probably a synapomorphy of only one lineage of the family. A vicariance expla- nation of the biogeography of the family, based on a Gondwanan distribution, seems less parsi- monious than an origin in South America, followed by several eastward dispersals. While Mono- choria and Pontederia are apparently monophyletic under different analyses, Eichhornia and Heteranthera, as presently circumscribed, are probably paraphyletic. We had two main goals in undertaking this study. The first was to examine the evolution- ary history of tristyly and several other system- atic features in the small, aquatic, monocoty- ledonous family Pontederiaceae. The other was to assess the performance of several different numerical cladistic procedures using character distributions in this family as test data. Grow- ing discussion of the theoretical bases of biological classification has spawned a bewil- dering array of individual methods. No com- parative studies have evaluated all proposed methods, but several of the most popular, in- cluding single linkage clustering, UPGMA (av- erage linkage) clustering, character compatibil- ity analysis, and Wagner tree construction

Isolation and Preliminary Taxonomic Studies of Thermus Strains Isolated from Yellowstone National Park, USA

Forty-eight strains of Thermus isolated from hot springs in Yellowstone National Park, Wyoming, USA, and eight reference strains were subjected to a numerical taxonomic analysis using Gower's coefficient (SG) with single and average linkage clustering. Two major groups were distinguished, which could be differentiated by colony morphology, ability to reduce nitrate and proteolytic activity. Cluster 1 contained Thermus aquaticus YT-1, the type strain of the species, and cluster 2 contained authentic strains of 'T. flavus' and 'T. thermophilus'. T. ruber was recovered as a single member cluster. The mol % G + C of DNA from representative strains from each cluster was 64.4 to 66.8 for cluster 1, 62.2 to 67.1 for cluster 2 and 62.5 for T. ruber.

Abstracts

The first part of this paper reports a comparative study of the document classifications produced by the use of the single linkage, complete linkage, group average, and Ward clustering methods. Studies of cluster membership and of the effectiveness of cluster searches support previous findings that suggest that the single linkage classifications are rather different from those produced by the other three methods. These latter methods all produce large numbers of small clusters containing just pairs of documents. This finding motivates the work reported in the second part of the paper, which considers the use of clusters consisting of a document together with that document with which it is most similar. A comparison of the use of such clusters with conventional best match searches using seven documents test collections suggest that the two types of search are of comparable effectiveness, but they retrieve noticeably different sets of relevant documents.

A Numerical Classification of Some Thermus Isolates

Summary: A numerical classification was performed on a collection of 45 Thermus isolates recovered from New Zealand hot pools and on six type strains including T. aquaticus and ‘T. thermophilus’. Unweighted average linking (UPGMA) and single linkage clustering methods were applied to similarity matrices derived from simple matching (SSM ) and Jaccard similarity (SJ ) coefficients. Differences were observed between phenograms derived from SSM and SJ coefficients, indicating that some of the clusters formed were derived from a significant component of negative matches. Test error was estimated at 2∙9%. In the UPGMA/SSM phenogram, seven clusters were formed. A majority of the New Zealand isolates did not cluster with non-New Zealand isolates. Analysis of variance showed that there was a relationship between the composition of the clusters and the temperature and pH of the source of the isolate. Chi-squared testing showed that, within New Zealand, the geographical source of the isolate had no bearing on the clusters formed.

Detection of diffuse clusters in noise background

The paper presents a method for the detection of a two-dimensional extended pattern of points in a flat background. The method is based on the statistical properties of a single-link clustering algorithm, which allows to select the most significant clusters. Its calibration is based on simulated data. An application on real X-ray astronomical is also shown.

Investigation of proportional link linkage clustering methods

Proportional link linkage (PLL) clustering methods are a parametric family of monotone invariant agglomerative hierarchical clustering methods. This family includes the single, minimedian, and complete linkage clustering methods as special cases; its members are used in psychological and ecological applications. Since the literature on clustering space distortion is oriented to quantitative input data, we adapt its basic concepts to input data with only ordinal significance and analyze the space distortion properties of PLL methods. To enable PLL methods to be used when the numbern of objects being clustered is large, we describe an efficient PLL algorithm that operates inO(n 2 logn) time andO(n 2) space.

Phylogeny of the Vibrionaceae, and Recommendation for Two New Genera, Listonella and Shewanella

Ribonucleotide sequences were determined for 5S rRNAs purified from 31 type cultures and freshly isolated strains representing species of the eubacterial family Vibrionaceae. These were compared with published 5S rRNA sequences for 11 other species of rRNA superfamily I (Vibrionaceae, Enterobacteriaceae) for the purpose of constructing a phylogenetic taxonomy of the Vibrionaceae. Sequence data were compiled and analyzed, using single, average, and complete linkage clustering methods, resulting in the generation of evolutionary trees. Results of this study indicate a good correlation between the taxonomy of the Vibrionaceae established using conventional and molecular genetic data and evolutionary relationships deduced from 5S rRNA sequences. However, several species, including V. anguillarum, V. pelagius, V. damsela, V. fischeri, V. logei, V. psychroerythrus, V. marinus, and Plesiomonas shigelloides require reconsideration of their taxonomic status, based on the new information obtained in this study. Furthermore, the genus Aeromonas appears to be sufficiently distantly related to the Vibrionaceae to warrant exclusion from the family Vibrionaceae. We propose a restructuring of the family Vibrionaceae along phylogenetic lines, and establishment of two new genera Listonella and Shewanella, and a new family Aeromonadaceae.

A note on the use of nearest neighbors for implementing single linkage document classifications

AbstractBest match search algorithms provide an efficient means of identifying the sets of nearest neighbors for each of the documents in a collection. These sets contain much of the important similarity data contained in a full interdocument similarity matrix and may be used for the generation of hierarchic document classifications, such as those arising from the use of the single linkage clustering method. Cluster based retrieval experiments based upon such classifications are shown to give results that are comparable in effectiveness with those obtained using the full similarity matrix.

HIERARCHIC AGGLOMERATIVE CLUSTERING METHODS FOR AUTOMATIC DOCUMENT CLASSIFICATION

This paper considers the classifications produced by application of the single linkage, complete linkage, group average and Ward clustering methods to the Keen and Cranfield document test collections. Experiments were carried out to study the structure of the hierarchies produced by the different methods, the extent to which the methods distort the input similarity matrices during the generation of a classification, and the retrieval effectiveness obtainable in cluster based retrieval. The results would suggest that the single linkage method, which has been used extensively in previous work on document clustering, is not the most effective procedure of those tested, although it should be emphasized that the experiments have used only small document test collections.

An Evaluation of the Stability of Single-Linkage Clusters

A stability of clusters can be defined by the tolerance for a small disturbance in the data. An estimating expression of the stability is shown in reference to the case of single-linkage algorithm. As one of the applications of this stability, the method of deciding on the number of single-linkage clusters is proposed. Some numerical examples are showrn in order to compare the method with other conventional methods.

Variability among isolates of Pseudomonas syringae pv. savastanoi from the phylloplane of the olive.

Leaves of three or four different ages were taken from olive plants quarterly in 1974-1980. One thousand and fifty isolates of Pseudomonas syringae pv. savastanoi from the phylloplane were tested for virulence to the olive and subjected to numerical phenetic analysis using 60 unit characters. The data were analysed using unweighted average linkage (UPGMA) and single linkage clustering on the simple matching (SSM) and pattern (SP) coefficients. The isolates obtained from leaves of a given age at a given time of the year shared higher percentage similarity (S) values between themselves than with the others. Cluster composition was only marginally affected by different coefficients and methods of clustering. UPGMA analysis on the SSM coefficient recovered 92% of the isolates in 10 major clusters at 75% S. Of the isolates from leaves of the same age collected at the same time of the year, 81-99% fell in the same cluster. Conversely, 91-97% of the isolates in five of the major clusters were from leaves of the same type. Of the isolates in the other major clusters, 95-98% were from two different sources but most of the isolates from leaves of one type segregated into discrete subclusters at 85-90% S. Hypothetical median organisms (HMOs) were constructed to represent all the isolates obtained from the leaves of each type each year. The resulting relationships between the HMOs confirmed those described above for the individual isolates.