Complete Linkage Method Research Articles

OPAC에서 탐색결과의 클러스터링에 관한 연구

This study evaluated the applicability of the static hierarchic clustering model to clustering query results in OPAC. Two clustering methods(Between Average Linkage(BAL) and Complete Linkage(CL)) and two similarity coefficients(Dice and Jaccard) were tested on the query results retrieved from 16 title-based keyword searchings. The precision of optimal dusters was improved more than 100% compared with title-word searching. There was no difference between similarity coefficients but clustering methods in optimal cluster effectiveness. CL method is better in precision ratio but BAL is better in recall ratio at the optimal top-level and bottom-level clusters. However the differences are not significant except higher recall ratio of BAL at the top-level duster. Small number of clusters and long chain of hierarchy for optimal cluster resulted from BAL could not be desirable and efficient.

Comparison of similarity coefficients used for cluster analysis with dominant markers in maize (Zea mays L)

The objective of this study was to evaluate whether different similarity coefficients used with dominant markers can influence the results of cluster analysis, using eighteen inbred lines of maize from two different populations, BR-105 and BR-106. These were analyzed by AFLP and RAPD markers and eight similarity coefficients were calculated: Jaccard, Sorensen-Dice, Anderberg, Ochiai, Simple-matching, Rogers and Tanimoto, Ochiai II and Russel and Rao. The similarity matrices obtained were compared by the Spearman correlation, cluster analysis with dendrograms (UPGMA, WPGMA, Single Linkage, Complete Linkage and Neighbour-Joining methods), the consensus fork index between all pairs of dendrograms, groups obtained through the Tocher optimization procedure and projection efficiency in a two-dimensional space. The results showed that for almost all methodologies and marker systems, the Jaccard, Sorensen-Dice, Anderberg and Ochiai coefficient showed close results, due to the fact that all of them exclude negative co-occurrences. Significant alterations in the results for the Simple Matching, Rogers and Tanimoto, and Ochiai II coefficients were not observed either, probably due to the fact that they all include negative co-occurrences. The Russel and Rao coefficient presented very different results from the others in almost all the cases studied and should not be used, because it excludes the negative co-occurrences in the numerator and includes them in the denominator of their expression. Due to the fact that the negative co-occurrences do not necessarily mean that the regions of the DNA are identical, the use of coefficients that do not include negative co-occurrences was suggested.

Genitores potenciais para hibridações identificados por divergência genética em feijão carioca

Noventa genótipos de feijão carioca (Phaseolus vulgaris L.) foram avaliados, em dois anos agrícolas, em Santa Maria, RS, a fim de definir quais características agromorfológicas constituem-se como melhores descritores, realizar agrupamento em função de dissimilaridade genética e de definir quais combinações híbridas mais promissoras serão obtidas para o desenvolvimento de populações segregantes. Dos 20 caracteres agromorfológicos avaliados, apenas nove (ferrugem nos legumes, acamamento, nota geral, cor do tegumento, rendimento de grãos, massa de 100 sementes, altura de inserção do primeiro legume, altura de inserção do último legume e número de sementes por legume) apresentaram maior contribuição para a divergência genética. Os genótipos de feijão carioca foram agrupados pelo método hierárquico de ligação completa. Populações segregantes, com variabilidade genética superior, podem ser obtidas com hibridações entre o genótipo ESAL 550 com genótipos do grupo 2 (LH-6, 17-4-32, R-78, H-4-5 e R-102) e/ou com genótipos do grupo 3 (FT 97-188, Cati-Taquari, CII-328, Carioca Precoce, FT 97-41, LH-11, FT 91-4067, Iapar 31, CI 102, Carioca MG, CII-54 e R-102).

Análise de crescimento de girassol em Latossolo com diferentes níveis de saturação por bases no Cerrado

Com o objetivo de avaliar o comportamento de cinco cultivares de girassol em um Latossolo Vermelho argiloso, foram realizados quatro ensaios sob quatro níveis de saturação por bases (18%, 35%, 56% e 77%) obtidos com aplicação de calcário dolomítico. Os dados semanais de crescimento foram ajustados pelas funções logística, Weibull de três parâmetros e a logística generalizada de quatro parâmetros empregando-se o procedimento NLIN do SAS, bem como os processos interativos do método DUD. A função logística forneceu os melhores ajustes aos dados de crescimento. As cultivares foram categorizadas de acordo com o porte das plantas e tendências de crescimento a partir da análise de agrupamentos (cluster analysis), usando o método das ligações completas e a métrica da distância euclidiana. Todas as cultivares tiveram o crescimento reduzido quando o nível de saturação de bases foi de 18%. A cultivar M734 apresentou bom desempenho no nível de 35% de saturação por bases. Nos demais casos, os melhores resultados foram obtidos nos níveis de saturação por bases entre 56% e 77%. As cultivares EMBRAPA 122 e M738 tiveram redução de crescimento quando a saturação foi elevada para 77%. 'EMBRAPA 122' apresentou maior crescimento inicial e menor altura ao final do ciclo vegetativo.

Emprego de análise multivariada para estratificação vertical de florestas ineqüiâneas

O objetivo do presente estudo foi propor um método de estratificação vertical da floresta ineqüiânea, empregando técnicas de análise multivariada. Os dados de altura total foram obtidos em dez parcelas de 20 x 50 m cada, remedidas em junho de 2001. As parcelas foram instaladas na Mata da Silvicultura, em 1994, município de Viçosa (20º 45' S e 42º 55' W), Estado de Minas Gerais. As árvores individuais foram organizadas em ordem crescente de altura total, separadas em classes com amplitude de 1 m, e em seguida foi elaborada uma matriz X de dados de alturas totais, em que cada variável x ij representou a altura total da i-ésima árvore classificada na j-ésima classe de altura total. Essa matriz X foi o input para as análises de agrupamento e discriminante. Utilizaram-se a distância euclidiana e o método de ligação completa. A aplicação da análise de agrupamento resultou em agrupamentos hierárquicos e seqüenciais, isto é, as classes de altura foram agrupadas seqüencialmente em ordem crescente. A análise do dendrograma permitiu estratificar o povoamento em três grupos distintos, denominados de estratos de altura total (inferior, médio e superior). A análise discriminante demonstrou que a classificação foi 100% correta. O estrato inferior reuniu 1.068 (71,63%) indivíduos arbóreos, o estrato médio 324 (21,73%) e o estrato superior 99 (6,64%). As espécies Anadenanthera macrocarpa, Apuleia leiocarpa, Pseudopiptadenia contorta e Piptadenia gonoacantha apresentaram maior densidade absoluta (DA) e distribuição regular nos três estratos. O método de estratificação empregado no presente estudo provou ser mais uma ferramenta técnica e operacionalmente viável, que pode ser empregada nas análises estruturais de florestas, principalmente para estudos fitossociológicos e para elaboração e execução de planos de manejo florestal sustentável.

Middle eocene ostracoda from Northern Somalia (paleoenvironmental appraisal)

The present study is focused on the reconstruction and interpretation of the paleoenvironmental conditions that prevailed during the deposition of the Middle Eocene deposits of Northern Somalia by means of the ostracod assemblages distinguished. Quantitative analyses of 75 ostracod species, recognized in 36 ostracod-bearing samples from the study area, has revealed six types of assemblages (a, b1, b2, b3, b+c, c), obtained by applying Q-mode cluster analysis based on the Jaccard coefficient of similarity (the complete linkage method). These types represent, from older to younger, the Thracella indica Assemblage Zone (type b1; middle Lutetian), the Paijenborchellina reticulata-Paijenborchellina pustulata Assemblage Zone (type b2; middle Lutetian), the Hornibrookella moosae Assemblage Zone (type b3; middle Lutetian), the Paragrenocythere dhofarensis somaliensis Assemblage Zone (type a; late Lutetian), the Asymmetricythere tricostata Assemblage Zone (type b+c; late Lutetian to Bartonian) and the Neocyprideis rotundata Assemblage Zone (type c; late Lutetian to Bartonian), respectively. Q-mode analysis based on the extraction of latent roots and vectors of a non-parametric matrix produced by the Spearman correlation coefficient, led to the conclusion that the most effective environmental factors in the study area are: salinity (1 st latent vector), substrate (2 nd latent vector), dissolved oxygen of water (3 rd latent vector), turbidity currents (4 th latent vector). The general paleoenvironmental conditions prevailing during the deposition of the Middle Eocene strata in Northern Somalia are discussed in details.

Methods of constructing core collections by stepwise clustering with three sampling strategies based on the genotypic values of crops

A genetic model with genotype×environment (GE) interactions for controlling systematical errors in the field can be used for predicting genotypic values by an adjusted unbiased prediction (AUP) method. Mahalanobis distance, calculated based on the genotypic values, is then applied to measure the genetic distance among accessions. The unweighted pair-group average, Ward’s and the complete linkage methods of hierarchical clustering combined with three sampling strategies are proposed to construct core collections in a procedure of stepwise clustering. A homogeneous test and t-tests are suggested for use in testing variances and means, respectively. The coincidence rate (CR%) for range and the variable rate (VR%) for the coefficient of variation are designed to evaluate the property of core collections. A worked example of constructing core collections in cotton with 21 traits was conducted. Random sampling can represent the genetic diversity structure of the initial collection. Preferred sampling can keep the accessions with special or valuable characteristics in the initial collection. Deviation sampling can retain the larger genetic variability of the initial collection. For better representation of the core collection, cluster methods should be combined with different sampling strategies. The core collections based on genotypic values retained larger genetic variability and had superior representatives than those based on phenotypic values.

Defining species guilds in the Central Hardwood Forest, USA

Tree regeneration outcomes are challenging to generalize and difficult to predict. Many tree species can establish new propagules in a variety of post-disturbance environments and many different reproductive mechanisms may be used. In order to develop conceptual models that accurately reflect reproductive potential, we need a better understanding of the similarities in regeneration ecology among species. We used information from the forest ecology literature to evaluate the reproductive attributes of sixty-two tree species in the central hardwood region of the eastern United States. Each species was classified categorically for features such as flowering, seed production and dispersal, seed dormancy, germination requirements, seedling characteristics, and vegetative reproduction. Cluster analysis (Jaccard’s similarity coefficient, complete linkage method) and ordination (homogeneity analysis) were used to separate nine groups (guilds) of species that had similar reproductive attributes. Individual attributes that had high variance in the first and second dimensions included: seed banking, seed dispersal, seedling shade tolerance, and seedbed requirements. Members of each guild had similar levels of reproductive specialization and guilds were either pioneer-like, opportunistic, or persistent. Pioneer guilds included: short-lived or fugitive species that colonize sites rapidly and are too shade intolerant to replace themselves; shade-tolerant species that colonize frequently disturbed sites; and stress-tolerant pioneers that survive on dry or nutrient-poor sites. Opportunistic guilds contained species that are remarkably versatile in their reproductive effort. The most flexible opportunists can colonize new sites, maintain seed in a seed bank, sprout from existing stems and persist as a seedling or sapling bank. Persistent guilds contain species that develop and maintain advance regeneration. These include: species with moderate understory tolerance that regenerate via cycles of dieback and resprouting; and more tolerant species that maintain seedling or sapling banks. Our regeneration guilds may provide a useful approach for more realistically representing large and diverse sets of tree species in forest ecosystem models.

Pseudocontinuity in hierarchical classifications

Geometrical representation of objects by means of tree models is based on order relations of the interdistances between the individuals that constitute the classifying set as a whole. Therefore, in this paper we show the pseudocontinuity of some classification methods that satisfy certain regularity conditions based on the conservation of the order in hierarchy construction. Finally, we prove that the required regularity conditions are not restrictive in excess and are accomplished among other methods for single linkage method, complete linkage method and UPGMA method.

Classification of Mesoscale Wind Fields in the MISTRAL Field Experiment

Abstract A two-stage classification scheme with outlier detection is proposed to find groups of wind fields. A hierarchical cluster analysis according to the complete linkage method is combined with a k-means procedure with detection and exclusion of outliers. The classification method is applied to a 1-yr dataset of 1-h mean wind observations from the MISTRAL field experiment. A small number of typical regional flow patterns is identified. An analysis of temperature observations shows that some of the 12 regional flow patterns have thermally forced wind systems. The main spatial forcing patterns are revealed by a principal component analysis of temperature observations. A comparison of the regional flow patterns and the synoptic-scale weather types of the Alpine region shows that only weak connections between the local flow and the synoptic-scale weather type exist.

Adequacy of Numerical Taxonomic Methods: Why Not be a Pheneticist?

A comparative study was conducted of the adequacy of phenetic and phylogenetic proce- dures, based on simulation experiments. Phenetic clustering methods examined included single linkage, complete linkage, average linkage, centroid sorting, median or Gower's method, Ward's method, Lance- Williams flexible BETA, and McQuitty's similarity analysis. Squared euclidean distance, product moment correlation, and cosine coefficient were used as (dis)similarity criteria, calculated both on standardized and non-standardized data. The Hennig86 and Wagner78 programs were used to calculate most parsimonious trees. The relative efficiency of these numerical procedures was evaluated using data matrices simulated under a large variety of evolutionary conditions. The efficiency was measured by the consensus of the true trees with the estimated ones, using Colless' consensus fork index (CFI). McQuitty's similarity analysis and the average linkage method based on cosine- or product moment correlation matrices for unstandardized characters were found to perform consistently better than the other phenetic clustering procedures and maximum parsimony. However, once again it was shown that none of these methods was very accurate. The majority of published phylogenies should therefore only be considered best approximations of the true tree.

Karyological Studies in Gossypium herbaceum Cultivars of Iran.

Forty-two G. herbaceum cultivars available in Iran National Gene Bank were analysed karyologically showing 2n=26 chromosome number. The cultivars varied with regard to number of SAT-chromosomes and chromosomes carrying secondary constrictions. Karyotypes were of symmetrical type having small chromosomes. Factorial analysis showed significant difference for both the cultivars as well as chromosomes.Cluster analysis using UPGMA, single linkage and complete linkage methods grouped the cultivars in six distinct clusters. Principal coordinate analysis supports the clustering results and principal component analysis showed that chromosomes 4 to 10 and long arms of the chromosomes 2, 8, 10 and 13 and short arms of the chromosomes 3, 7 and 8 are the most variable chromosomes/chromosome arms among the cultivars studied.Present work indicates genomic differences among diploid G. herbaceum cultivars of Iran which can be used in hybridization programs. Crossing of the cultivars placed in distant clusters are suggested.

Automated Classification Scheme for Wind Fields

In earlier works of several authors, wind fields or flow patterns were grouped by means of trajectory calculations or by use of a principal component analysis. A new automated classification method is proposed that makes use only of wind observations and does not require predefined circulation patterns, a priori rules, or spatial or temporal interpolation. A distance measure is defined between pairs of wind fields, represented by a set of observations. Based on the resulting distance matrix, a hierarchical cluster analysis is performed that also provides an indication for the choice of an appropriate number of clusters. The classification method is applied to a monthly record of 1-h averages of wind measurements in the region of Basel, Switzerland. Different clustering schemes are tested, and the complete linkage method is chosen as most appropriate. For the test case a relatively small number of classes (12) provides a sufficient description of the different flow patterns in this complex topography. The strong channeling of flow in the valleys of the Basel region seems to result in only a small number of distinguishable flow patterns.

Comparison of Chernoff-type face and non-graphical methods for clustering multivariate observations

During the past few years the use of Chernoff-type faces (a technique of representing points in k-dimensional space graphically) has been accelerating for discovering clusters and outliers present in a set of multivariate observations. However, their validation for clustering multivariate data seems to be few. The present study deals with evaluating the efficiency of Chernoff-type faces (Flury and Riedwyl) and four non-graphical algorithms, viz., ‘Single linkage’, ‘Complete linkage’, ‘Group average’ and Ward's methods for recovering true cluster structures. Four different data sets have been used, which were based on samples from mixtures of multivariate normal populations with varying variance-covariance matrix and different levels of separation factor. In order to account for the variability between the subjects, experiments with 30 subjects were conducted and they were asked to group all those faces together which look alike. The recovery of hierarchical cluster structure through Chernoff-type faces has been measured using the agreement score which is simply the number of subjects classifying the pair of faces together just by merely looking at the gestalt of a face. In all the four populations considered, the recovery of cluster structure through Chernoff-type faces using the perception matrix based on the classification of faces done by 30 subjects was more similar to that obtained by complete linkage method as evaluated by adjusted Rand statistic and Jaccard index. The efficacy of Chernoff-type faces in recovering true cluster structure seems to be poor when the classification is done by one subject only. The present study revealed that none of the subjects gave a true partitions in all the four populations considered. However, the recovery values (both adjusted Rand and Jaccard) had improved substantially when the clusters were obtained from perception matrix presenting the number of subjects who identified any two faces more similar in their gestalt.

Molecular evidence supports the separation of Lentinula edodes from Lentinus and related genera

Restriction polymorphisms in two regions of the ribosomal DNA (rDNA) repeat unit were examined in 18 strains of Lentinus, Neolentinus, Pleurotus, and the shiitake mushroom Lentinula edodes. The polymerase chain reaction was used to separately amplify the 18S rDNA and the region spanning the two internal transcribed spacers and the 5.8S ribosomal RNA gene. Amplified products were digested with a battery of 10 restriction endonucleases and the two data sets were subjected to cluster analysis. All strains of Lentinula edodes consistently exhibited identical restriction profiles that were distinct from those of the genera Lentinus, Neolentinus, and Pleurotus. The internal transcribed spacer region exhibited more variability than the 18S rDNA, giving distinctive profiles for two strains of Lentinus tigrinus and for one strain of Neolentinus lepideus. Similarity coefficients were clustered with the unweighted pair group method with arithmetic average, single-linkage method, and complete-linkage method. Results from cluster analysis of the two data sets were highly congruent and tree topologies were consistent irrespective of the clustering method used. The distinctiveness of the groups was further confirmed by computing for consensus trees and cophenetic correlation coefficients. Ribosomal DNA restriction polymorphisms support the placement of the strains examined in separate taxa. Key words: Lentinula, Lentinus, Neolentinus, Pleurotus, ribosomal DNA.

AN ELECTROPHORETIC COMPARISON OF CULTIVATED MEMBERS OF THE CUCURBITACEAE

Thirty-eight cultivated accessions of the diverse Cucurbitaceae were electrophoretically surveyed using 13 enzyme systems. Included were representatives from 6 of the 6 Cucurbitaceae tribes, 9 genera, and 17 species. Additionally, several cultivars or groups were included for those species possessing marked morphological diversity such as the 7 groups of Cucumis melo var. melo and 7 of the numerous cultivars representing Cucurbita pepo. Zymograms were scored for the presence or absence of bands measured in mm from the origin. Cluster analysis (complete linkage method) was used to detect affinities among the accession surveyed. Data suggest that: 1) Cucumis melo (x=12) possessed greater biochemical affinity with C. sativus (x=7) than with either C. anguira or C. metuliferus (both x=12); 2) Sechium edule and Cyclanthera pedata. both members of the tribe Sicyeae, were more closely associated with members of other tribes than with each other; 3) Some cultivars of Cucurbita pepo shared greater affinity with Cucurbita moschata than with other cultivars of C. pepo. Additional observations as well as their possible implications will be presented.

Six clustering algorithms applied to the WAIS-R: the problem of dissimilar cluster results.

Clusterings of the Wechsler Adult Intelligence Scale-Revised subtests were obtained from the application of six hierarchical clustering methods (N = 113). These sets of clusters were compared for similarities using the Rand index. The calculated indices suggested similarities of cluster group membership between the Complete Linkage and Centroid methods; Complete Linkage and Ward's methods; Centroid and Ward's methods; and Single Linkage and Average Linkage Between Groups methods. Cautious use of single clustering methods is implied, though the authors suggest some advantages of knowing specific similarities and differences. If between-method comparisons consistently reveal similar cluster membership, a choice could be made from those algorithms that tend to produce similar partitions, thereby enhancing cluster interpretation.

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.

Monte Carlo comparison of six hierarchical clustering methods on random data

There is mounting evidence to suggest that the complete linkage method does the best clustering job among all hierarchical agglomerative techniques, particularly with respect to misclassification in samples from known multivariate normal distributions. However, clustering methods are notorious for discovering clusters on random data sets also. We compare six agglomerative hierarchical methods on univariate random data from uniform and standard normal distributions and find that the complete linkage method generally is best in not discovering false clusters. The criterion is the ratio of number of within-cluster distances to number of all distances at most equal to the maximum within-cluster distance.

Using interdocument similarity information in document retrieval systems

Etude comparative de quatre methodes de constitution de grappes de documents et de leur efficacite lors de la recherche documentaire, testee sur sept collections de documents