Preliminary Clustering Research Articles

Three-step estimation of latent Markov models with covariates

A three-step approach is proposed to estimate latent Markov (LM) models for longitudinal data with and without covariates. The approach is based on a preliminary clustering of sample units on the basis of time-specific responses only, and is particularly useful when a large number of response variables are observed at each time occasion. In such a context, full maximum likelihood estimation, which is typically based on the Expectation–Maximization algorithm, may have some drawbacks, essentially due to the presence of many local maxima of the model likelihood. Moreover, this algorithm may be particularly slow to converge, and may become unstable with complex LM models. The properties of the proposed estimator are illustrated theoretically and by a simulation study in which this estimator is compared with the full likelihood estimator. How reliable standard errors for the three-step parameter estimates are obtained is also shown. The approach is applied to the analysis of a dataset about the health status of elderly people resident in certain Italian nursing homes.

Data-driven forecasting of naturally fractured reservoirs based on nonlinear autoregressive neural networks with exogenous input

In this paper we discuss the results of the modeling of naturally fractured reservoir based on the application of the nonlinear autoregressive neural network with exogenous inputs (NARX). We show that the NARX network can be efficiently applied to multivariate multi-step ahead prediction of reservoir dynamics. Predictability of the time series is studied using the Hurst exponent. We show that preliminary clustering of the time series can increase the precision of the forecasting. We evaluate the proposed approach using a real-world data set describing the dynamic behavior of a naturally fractured oilfield asset located in the coastal swamps of the Gulf of Mexico. This paper is not only intended for proposing a new model but to study carefully and thoroughly several aspects of the application of ANN models to reservoir modeling and to discuss conclusions that could be of the interest for petroleum engineers. • The NARX network is a feasible model for reservoir modeling. • It shows competitive accuracy in 1-year oil production prediction. • The use of both static and dynamic data extends the predictive capabilities of the ANN model. • Preliminary clustering can be a good means to increment the precision of the forecasting.

A Novel Clustering Routing Protocol in Wireless Sensor Network

The phenomenon of network resource-constrained often appears in wireless sensor network (WSN) because of nodes with energy-limited, so it becomes a research topic to study on high-performance routing protocol in wireless sensor network. In accordance with the characteristic of quantum particle swarm optimization (QPSO), a novel clustering routing protocol with QPSO algorithm is proposed based on the low energy adaptive clustering hierarchy (LEACH), which includes several steps, such as determining the number of cluster-head, preliminary clustering, optimization on temporary cluster by QPSO algorithm, clustering after optimization, and so on. Comparative analysis on simulation experiments show that the proposed routing protocol is superior to that of LEACH, and saves energy consumption of network excellently and balances energy of nodes so as to prolong the life of the network.

Familial clustering of mice consistent to known pedigrees enabled by the genome profiling (GP) method.

Familial clustering without any prerequisite knowledge becomes often necessary in Behavioral Science, and forensic studies in case of great disasters like Tsunami and earthquake requiring body-identification without any usable information. However, there has been no well-established method for this purpose although conventional ones such as short tandem repeats (STR) and single nucleotide polymorphism (SNP), which might be applied with toil and moil to some extent. In this situation, we could find that the universal genome distance-measuring method genome profiling (GP), which is made up of three elemental techniques; random PCR, micro-temperature gradient gel electrophoresis (μTGGE), and computer processing for normalization, can do this purpose with ease when applied to mouse families. We also confirmed that the sequencing approach based on the ccgf (commonly conserved genetic fragment appearing in the genome profile) was not completely discriminative in this case. This is the first demonstration that the familial clustering can be attained without a priori sequence information to the level of discriminating strains and sibling relationships. This method can complement the conventional approaches in preliminary familial clustering.

Authentication of Monofloral Yemeni Sidr Honey Using Ultraviolet Spectroscopy and Chemometric Analysis

This work describes a simple model developed for the authentication of monofloral Yemeni Sidr honey using UV spectroscopy together with chemometric techniques of hierarchical cluster analysis (HCA), principal component analysis (PCA), and soft independent modeling of class analogy (SIMCA). The model was constructed using 13 genuine Sidr honey samples and challenged with 25 honey samples of different botanical origins. HCA and PCA were successfully able to present a preliminary clustering pattern to segregate the genuine Sidr samples from the lower priced local polyfloral and non-Sidr samples. The SIMCA model presented a clear demarcation of the samples and was used to identify genuine Sidr honey samples as well as detect admixture with lower priced polyfloral honey by detection limits >10%. The constructed model presents a simple and efficient method of analysis and may serve as a basis for the authentication of other honey types worldwide.

Estimates of genetic parameters, and cluster and principal components analyses of breeding values related to egg production traits in a White Leghorn population

The objectives of this paper were to identify the phenotypic egg-laying patterns in a White Leghorn line mainly selected for egg production, to estimate genetic parameters of traits related to egg production and to evaluate the genetic association between these by principal components analysis to identify trait(s) that could be used as selection criteria to improve egg production. Records of 54 wk of egg production from a White Leghorn population were used. The data set contained records of the length:width ratio of eggs at 32, 37, and 40 wk of age; egg weight at 32, 37, and 40 wk of age; BW at 54 and 62 wk of age; age at first egg; early partial egg production rate from 17 to 30 wk and from 17 to 40 wk of age; late partial egg production rate from 30 to 70 wk and from 40 to 70 wk of age; and total egg production rate (TEP). The estimates of genetic parameters between these traits were estimated by the restricted maximum likelihood method. Multivariate analyses were performed: a hierarchical cluster analysis, a nonhierarchical clustering analysis by the k-means method of weekly egg production rate to describe the egg-laying patterns of hens, and a principal components analysis using the breeding values of all traits. The highest heritability estimates were obtained for BW at 54 wk of age (0.68 ± 0.07) and age at first egg (0.53 ± 0.07). It is recommended that a preliminary clustering analysis be performed to obtain the population structure that takes into account the pattern of egg production, rather than the TEP, because hens may have the same final egg production with different patterns of egg laying. Early partial production periods were not good indicators for use in improving total egg production because these traits presented an overestimated genetic correlation with TEP because of the part-whole genetic correlation component. Egg production might be improved by selecting individuals based on TEP.

The Malagasy Goblin Spiders of the New Genus Malagiella (Araneae, Oonopidae)

A new oonopid genus endemic to Madagascar, Malagiella, is proposed and its 10 constituent species newly described and illustrated: M. ambalavo (♀), M. andringitra (♂♀), M. fisheri (♂♀), M. goodmani (♀), M. nikina (♀), M. ranavalona (♀), M. ranomafana (♂♀), M. toliara (♂♀), M. valterova (♂♀), and M. vohiparara (♂♀).Malagiella is most similar to the Asian genus Camptoscaphiella in both somatic and genitalic features, but differs in some important characters. Unlike Camptoscaphiella, the Malagiella male has a palpal bulb completely fused with the tarsus and a sternum bearing stiff bristles; the female has a round copulatory opening and bifid 4th claws.The species may be recognized by the genitalia. Male palpi are fairly uniform, but differ in segment proportion and details of the embolus complex, which consists of a broad dorsal embolus connected to a laminar ventral prong. Female genitalia are more variable, especially in the curvature of the receptaculum, which spans the range from nearly straight to widely sinuous and occurs in mirror-image versions having opposite coiling. These torsional variants, with females having either left- or right-handed receptacula, probably occur in all species, although they are not evident in species with straight receptacula and (obviously) unknown in species represented by singleton female specimens.Malagiella species are very diverse somatically, ranging from dark, big, and large-eyed species to pale, small, and reduced-eyed ones. Our preliminary clustering uses mostly these somatic features, with some genitalic support, to identify three species groups: ranomafana, vohiparara, and toliara. Other genitalic characters that suggest different groupings are discussed.

Multilevel local search algorithms for modularity clustering

Modularity is a widely used quality measure for graph clusterings. Its exact maximization is NP-hard and prohibitively expensive for large graphs. Popular heuristics first perform a coarsening phase, where local search starting from singleton clusters is used to compute a preliminary clustering, and then optionally a refinement phase, where this clustering is improved by moving vertices between clusters. As a generalization, multilevel heuristics coarsen in several stages, and refine by moving entire clusters from each of these stages, not only individual vertices. This article organizes existing and new single-level and multilevel heuristics into a coherent design space, and compares them experimentally with respect to their effectiveness (achieved modularity) and runtime. For coarsening by iterated cluster joining, it turns out that the most widely used criterion for joining clusters (modularity increase) is outperformed by other simple criteria, that a recent multistep algorithm [Schuetz and Caflisch 2008] is no improvement over simple single-step coarsening for these criteria, and that the recent multilevel coarsening by iterated vertex moving [Blondel et al. 2008] is somewhat faster but slightly less effective (with refinement). The new multilevel refinement is significantly more effective than the conventional single-level refinement or no refinement, in reasonable runtime. A comparison with published benchmark results and algorithm implementations shows that multilevel local search heuristics, despite their relative simplicity, are competitive with the best algorithms in the literature.

Probability of default models of Russian banks

This paper presents results from an econometric analysis of Russian bank defaults during the period 1997–2003, focusing on the extent to which publicly available information from quarterly bank balance sheets is useful in predicting future defaults. Binary choice models are estimated to construct the probability of default model. In the first part of the paper we analyse bank survival over the financial crisis of 1998. We find that preliminary expert clustering or automatic clustering improves the predictive power of the models and incorporation of macrovariables into the models is useful. Heuristic criteria are suggested to help compare model performance from the perspectives of investors or banks supervision authorities. In the second part of the paper we use the probability of default models developed in the first part in rolling windows to analyse the Russian banking system trends after the crisis 1998.

10x10: A General-purpose Architectural Approach to Heterogeneity and Energy Efficiency

Abstract Two decades of microprocessor architecture driven by quantitative 90/10 optimization has delivered an extraordinary 1000-fold improvement in microprocessor performance, enabled by transistor scaling which improved density, speed, and energy. Recent generations of technology have produced limited benefits in transistor speed and power, so as a result the industry has turned to multicore parallelism for performance scaling. Long-range studies [1] , [2] indicate that radical approaches are needed in the coming decade – extreme parallelism, near-threshold voltage scaling (and resulting poor single-thread performance), and tolerance of extreme variability – are required to maximize energy efficiency and compute density. These changes create major new challenges in architecture and software. As a result, the performance and energy-efficiency advantages of heterogeneous architectures are increasingly attractive. However, computing has lacked an optimization paradigm in which to systematically analyze, assess, and implement heterogeneous computing. We propose a new paradigm, “10x10”, which clusters applications into a set of less frequent cases (ie. 10% cases), and creates customized architecture, implementation, and software solutions for each of these clusters, achieving significantly better energy efficiency and performance. We call this approach “10x10” because the approach is exemplified by optimizing ten different 10% cases, reflecting a shift away from the 90/10 optimization paradigm framed by Amdahl's law [3] . We describe the 10x10 approach, explain how it solves the major obstacles to widespread adoption of heterogeneous architectures, and present a preliminary 10x10 clustering, strawman architecture, and software tool chain approach.

Distributed hierarchical search for balanced energy consumption routing spanning trees in wireless sensor networks

We provide a new heuristic method approach to search for degree-balanced and small weight routing spanning trees in a network. The method is a modification of Kruskal’s minimum spanning tree search algorithm and is based on a distributed search by hierarchical clusters. It provides spanning trees with a lower maximum weighted degree, a bigger diameter, and can be used for balanced energy consumption routing in wireless sensor networks (WSN’s). The method can be naturally implemented in parallel or as a simple locally distributed algorithm. Simulations for a realistic case scenario WSN are done based on the transmission energy matrix. The simulation results show that the proposed approach can extend the functional lifetime of a WSN in terms of sensor transmission energy by 3–4 times. We also show that the results can be further improved by using a preliminary clustering of the input network.

Building networks with microarray data.

This chapter describes methods for learning gene interaction networks from high-throughput gene expression data sets. Many genes have unknown or poorly understood functions and interactions, especially in diseases such as cancer where the genome is frequently mutated. The gene interactions inferred by learning a network model from the data can form the basis of hypotheses that can be verified by subsequent biological experiments. This chapter focuses specifically on Bayesian network models, which have a level of mathematical detail greater than purely conceptual models but less than detailed differential equation models. From a network learning perspective the most severe problem with microarray data is the limited sample size, since there are usually many plausible networks for modeling the system. Since these cannot be reliably distinguished using the number of samples found in current microarray data sets, we describe robust network learning strategies for reducing the number of false interactions detected. We perform preliminary clustering using co-expression network analysis and gene shaving. Subsequently we construct Bayesian networks to obtain a global perspective of the relationships between these gene clusters. Throughout this chapter, we illustrate the concepts being expounded by referring to an ongoing example of a publicly available breast cancer data set.

Use of hybrid intelligent computing in mineral resources evaluation

Mineral resources are a formal quantification of naturally occurring materials. Estimation of resource parameters such as grade and thickness may be carried out using different methodologies. In this paper, a soft methodology, which is artificial neural network (ANN) based fuzzy modelling is presented for grade estimation and its stages are demonstrated. The neuro-fuzzy method uses preliminary clustering and finally estimates the ore grades based on radial basis neural network and interpolation. Two case studies designed for both simulated and real data sets indicate that the approach is relatively accurate and flexible. In addition, the method is suitable for modelling via limited number of data. The results and performance comparisons with conventional methods show that the computing method is efficient.

The Analytical Model of Technical Diagnosis of Telecommunication Network Devices with Preliminary Clustering of the Attribute Space

The Analytical Model of Technical Diagnosis of Telecommunication Network Devices with Preliminary Clustering of the Attribute Space

Differential patterns of evolution and distribution of the symbiotic behaviour in nostocacean cyanobacteria

Many cyanobacteria commonly identified as belonging to the genus Nostoc are well-known cyanobionts (symbionts) of a wide variety of plants and fungi. They form symbioses with bryophytes, pteridophytes, gymnosperms and angiosperms that are considerably different in the type of reciprocal interaction between the host and the cyanobiont. The phylogenetic and taxonomic relationships among cyanobionts isolated from different hosts and Nostoc strains isolated from free-living conditions are still not well understood. We compared phylogeny and morphology of symbiotic cyanobacteria originating from different host plants (genera Gunnera, Azolla, Cycas, Dioon, Encephalartos, Macrozamia and Anthoceros) with free-living Nostoc isolates originating from different habitats. After preliminary clustering with ARDRA (amplified rDNA restriction analysis), phylogeny was reconstructed on the basis of 16S rRNA gene sequences and compared with morphological characterization, obtaining several supported clusters. Two main Nostoc clusters harboured almost all cyanobionts of Gunnera, Anthoceros and of several cycads, together with free-living strains of the species Nostoc muscorum, Nostoc calcicola, Nostoc edaphicum, Nostoc ellipsosporum and strains related to Nostoc commune. We suggest that the frequent occurrence of symbiotic strains within these clusters is explained by the intensive hormogonia production that was observed in many of the strains studied. However, no evidence for discrimination between symbiotic and free-living strains, either by molecular or morphological approaches, could be found. Sequences of Azolla cyanobiont filaments, taken directly from leaf cavities, clustered tightly with sequences from the planktic cyanobacterium Cylindrospermopsis raciborskii, from the benthic Anabaena cylindrica 133 and from Anabaena oscillarioides HINDAK 1984/43, with high bootstrap values. The phylogenetic analysis showed that two distinct patterns of evolution of symbiotic behaviour might exist for the nostocacean cyanobacteria, one leading to symbioses of Nostoc species with a wide variety of plants, the other leading to the association of a unique cyanobacterial type with the water fern Azolla.

An identification approach to dynamic errors-in-variables systems with a preliminary clustering of observations

Errors-in-variables models are statistical models in which not only dependent but also independent variables are observed with error, i.e. they exhibit a symmetrical model structure in terms of noise. The application field for these models is diverse including computer vision, image reconstruction, speech and audio processing, signal processing, modal and spectral analysis, system identification, econometrics and time series analysis. This paper explores applying the errors-in-variables approach to parameter estimation of discrete-time dynamic linear systems. In particular, a framework is introduced in which a preliminary separation step is applied to group observations prior to parameter estimation. As a result, instead of one, two sets of estimates are derived simultaneously, comparing which can yield estimates for noise parameters. The proposed approach is compared to other schemes with simulation examples.

Identification of the Technical State of Telecommunication Network Devices Based on Preliminary Clustering of the Attribute Space

Identification of the Technical State of Telecommunication Network Devices Based on Preliminary Clustering of the Attribute Space

RIC

How do we find a natural clustering of a real-world point set which contains an unknown number of clusters with different shapes, and which may be contaminated by noise? As most clustering algorithms were designed with certain assumptions (Gaussianity), they often require the user to give input parameters, and are sensitive to noise. In this article, we propose a robust framework for determining a natural clustering of a given dataset, based on the minimum description length (MDL) principle. The proposed framework, robust information-theoretic clustering (RIC) , is orthogonal to any known clustering algorithm: Given a preliminary clustering, RIC purifies these clusters from noise, and adjusts the clusterings such that it simultaneously determines the most natural amount and shape (subspace) of the clusters. Our RIC method can be combined with any clustering technique ranging from K-means and K-medoids to advanced methods such as spectral clustering. In fact, RIC is even able to purify and improve an initial coarse clustering, even if we start with very simple methods. In an extension, we propose a fully automatic stand-alone clustering method and efficiency improvements. RIC scales well with the dataset size. Extensive experiments on synthetic and real-world datasets validate the proposed RIC framework.

Fuzzy Modeling for Reserve Estimation Based on Spatial Variability

This article addresses a new reserve estimation method which uses fuzzy modeling algorithms and estimates the reserve parameters based on spatial variability. The proposed fuzzy modeling approach has three stages: (1) Structure identification and preliminary clustering, (2) Variogram analysis, and (3) Clustering based rule system. A new clustering index approach and a new spatial measure function (point semimadogram) are proposed in the paper. The developed methodology uses spatial variability in each step and takes the fuzzy rules from input-output data. The model has been tested using both simulated and real data sets. The performance evaluation indicates that the new methodology can be applied in reserve estimation and similar modeling problems

PermutMatrix: a graphical environment to arrange gene expression profiles in optimal linear order

PermutMatrix is a work space designed to graphically explore gene expression data. It relies on the graphical approach introduced by Eisen and also offers several methods for the optimal reorganization of rows and columns of a numerical dataset. For example, several methods are proposed for optimal reorganization of the leaves of a hierarchical clustering tree, along with several seriation or unidimensional scaling methods that do not require any preliminary hierarchical clustering. This program, developed for MS Windows, with MS-Visual C++, has a clear and efficient graphical interface. Large datasets can be thoroughly and quickly analyzed.