Model-based Clustering Research Articles

Aggregated functional data model applied on clustering and disaggregation of UK electrical load profiles

AbstractUnderstanding electrical energy demand at the consumer level plays an important role in planning the distribution of electrical networks and offering of off-peak tariffs, but observing individual consumption patterns is still expensive. On the other hand, aggregated load curves are normally available at the substation level. The proposed methodology separates substation aggregated loads into estimated mean consumption curves, called typical curves, including information given by explanatory variables. In addition, a model-based clustering approach for substations is proposed based on the similarity of their consumers’ typical curves and covariance structures. The methodology is applied to a real substation load monitoring dataset from the UK and tested in eight simulated scenarios.

Genetic Diversity and Population Structure of Common Bean (Phaseolus vulgaris L.) Landraces in the Lazio Region of Italy.

Common bean cultivation has historically been a typical component of rural economies in Italy, particularly in mountainous and hilly zones along the Apennine ridge of the central and southern regions, where the production is focused on local landraces cultivated by small-scale farmers using low-input production systems. Such landraces are at risk of genetic erosion because of the recent socioeconomic changes in rural communities. One hundred fourteen accessions belonging to 66 landraces still being grown in the Lazio region were characterized using a multidisciplinary approach. This approach included morphological (seed traits), biochemical (phaseolin and phytohemagglutinin patterns), and molecular (microsatellite loci) analyses to investigate their genetic variation, structure, and distinctiveness, which will be essential for the implementation of adequate ex situ and in situ conservation strategies. Another objective of this study was to determine the original gene pool (Andean and Mesoamerican) of the investigated landraces and to evaluate the cross-hybridization events between the two ancestral gene pools in the P. vulgaris germplasm in the Lazio region. Molecular analyses on 456 samples (four for each of the 114 accessions) revealed that the P. vulgaris germplasm in the Lazio region exhibited a high level of genetic diversity (He = 0.622) and that the Mesoamerican and Andean gene pools were clearly differentiated, with the Andean gene pool prevailing (77%) and 12% of landraces representing putative hybrids between the two gene pools. A model-based cluster analysis based on the molecular markers highlighted three main groups in agreement with the phaseolin patterns and growth habit of landraces. The combined utilisation of morphological, biochemical, and molecular data allowed for the differentiation of all landraces and the resolution of certain instances of homonymy and synonymy. Furthermore, although a high level of homozygosity was found across all landraces, 32 of the 66 examined (49%) exhibited genetic variability, indicating that the analysis based on a single or few plants per landrace, as usually carried out, may provide incomplete information.

Clustering of variables methods and measurement models for soccer players’ performances

In sports, studying player performances is a key issue since it provides a guideline for strategic choices and helps teams in the complex procedure of buying and selling of players. In this paper we aim at investigating the ability of various composite indicators to define a measurement structure for the global soccer performance. We rely on data provided by the EA Sports experts, who are the ultimate authority on soccer performance measurement: they periodically produce a set of players’ attributes that make up the broader, theoretical performance dimensions. Considering the potential of clustering techniques to confirm or disconfirm the experts’ assumptions in terms of aggregations between indicators, 29 players’ performance attributes or variables (from the FIFA19 version of the videogame, that is, sofifa) have been considered and processed with three different techniques: the Cluster of variables around latent variables (CLV), the Principal covariates regression (PCovR) and Bayesian model-based clustering (B-MBC). The three procedures yielded clusters that differed from experts’ classification. In order to identify the most appropriate measurement structure, the resulting clusters have been embedded into Structural equation models with partial least squares (PLS-SEMs) with a Higher-Order Component (that is, the overall soccer performance). The statistically derived composite indicators have been compared with those of experts’ classification. Results support the concurrent validity of composite indicators derived through the statistical methods: overall, they show that, in the lack of expert judgement, composite indicators, as well as the resulting PLS-SEM models, are a viable alternative given their greater correlation to players’ economic value and salary.

Time series clustering with an EM algorithm for mixtures of linear Gaussian state space models

In this paper, we consider the task of clustering a set of individual time series while modeling each cluster, that is, model-based time series clustering. The task requires a parametric model with sufficient flexibility to describe the dynamics in various time series. To address this problem, we propose a novel model-based time series clustering method with mixtures of linear Gaussian state space models, which have high flexibility. The proposed method uses a new expectation-maximization algorithm for the mixture model to estimate the model parameters, and determines the number of clusters using the Bayesian information criterion. Experiments on a simulated dataset demonstrate the effectiveness of the method in clustering, parameter estimation, and model selection. The method is applied to real datasets commonly used to evaluate time series clustering methods. Results showed that the proposed method produces clustering results that are as accurate or more accurate than those obtained using previous methods.

Marker-trait association analyses revealed major novel QTLs for grain yield and related traits in durum wheat.

The growing global demand for wheat for food is rising due to the influence of population growth and climate change. The dissection of complex traits by employing a genome-wide association study (GWAS) allows the identification of DNA markers associated with complex traits to improve the productivity of crops. We used GWAS with 10,045 single nucleotide polymorphism (SNP) markers to search for genomic regions associated with grain yield and related traits based on diverse panels of Ethiopian durum wheat. In Ethiopia, multi-environment trials of the genotypes were carried out at five locations. The genotyping was conducted using the 25k Illumina Wheat SNP array to explore population structure, linkage disequilibrium (LD), and marker-trait associations (MTAs). For GWAS, the multi-locus Fixed and Random Model Circulating Probability Unification (FarmCPU) model was applied. Broad-sense heritability estimates were high, ranging from 0.63 (for grain yield) to 0.97 (for thousand-kernel weight). The population structure based on principal component analysis, and model-based cluster analysis revealed two genetically distinct clusters with limited admixtures. The LD among SNPs declined within the range of 2.02-10.04 Mbp with an average of 4.28 Mbp. The GWAS scan based on the mean performance of the genotypes across the environments identified 44 significant MTAs across the chromosomes. Twenty-six of these MTAs are novel, whereas the remaining 18 were previously reported and confirmed in this study. We also identified candidate genes for the novel loci potentially regulating the traits. Hence, this study highlights the significance of the Ethiopian durum wheat gene pool for improving durum wheat globally. Furthermore, a breeding strategy focusing on accumulating favorable alleles at these loci could improve durum wheat production in the East African highlands and elsewhere.

Model-based clustering for multidimensional social networks

Abstract Social network data are relational data recorded among a group of actors, interacting in different contexts. Often, the same set of actors can be characterised by multiple social relations, captured by a multidimensional network. A common situation is that of colleagues working in the same institution, whose social interactions can be defined on professional and personal levels. In addition, individuals in a network tend to interact more frequently with similar others, naturally creating communities. Latent space models for network data are useful to recover clustering of the actors, as they allow to represent similarities between them by their positions and relative distances in an interpretable low-dimensional social space. We propose the infinite mixture latent position cluster model for multidimensional network data, which enables model-based clustering of actors interacting across multiple social dimensions. The model is based on a Bayesian non-parametric framework that allows to perform automatic inference on the clustering allocations, the number of clusters, and the latent social space. The method is tested on extensive simulated data experiments. It is also employed to investigate the presence of communities in two multidimensional workplace social networks recording relations of different types among colleagues.

Identifying urban form typologies in Seoul using a new Gaussian mixture model-based clustering framework

Seoul, the capital city of South Korea, has diverse urban forms developed through its complex history. Previous studies show limitations of strong subjectivity and difficulty in scalability in identifying typical Seoul urban forms with expert knowledge. Data-driven approach offers an opportunity to address those challenges, but previous studies often focused on direct applications of clustering algorithms to a given area with diverse methods and workflows, lacking a systematic framework. This study addressed these issues by developing a new form clustering framework to systematically identify form typologies at a large scale and demonstrated its application in Seoul. With a 500 m × 500 m grid as the basic spatial unit and twelve urban form attributes as learning features, 14 clusters were identified using the Gaussian mixture model. These clusters were further translated into form typologies following a semantic typology naming system, with representative form samples identified. The resulting typologies were then verified and validated through comparisons with previous studies. Their relationships with zoning classes were also examined, emphasizing their role in urban planning and design. Results suggest this new framework is an effective and promising way to identify urban form typologies in complex urban environments to better support urban planning and management.

Clustering spatial networks through latent mixture models

AbstractWe consider a Bayesian model-based clustering technique that directly accounts for network relations between territorial units and their position in a geographical space. This proposal is motivated by a practical problem: to design administrative structures that are intermediate between the municipality and the province within an Italian region based on the existence of a relatively (to population) high commuting flow. In our social network model, the commuting flows are explained by the distances between the municipalities, i.e., the nodes, in a 3-dimensional space, where the 2 actual geographical coordinates and the third latent variable are modelled through a Gaussian mixture.

Characterization and Application of EST-SSR Markers Developed from Transcriptome Sequences in Elymus breviaristatus (Poaceae: Triticeae).

Elymus L. is the largest genus in the Triticeae tribe. Most species in this genus are highly stress resistant, with excellent forage value. Elymus breviaristatus, a rare species endemic to the Qinghai-Tibet Plateau (QTP), is declining due to habitat fragmentation. However, genetic data for E. breviaristatus are limited, with expressed sequence tag (EST) markers being particularly rare, hampering genetic studies and protection measures. We obtained 9.06 Gb clean sequences from the transcriptome of E. breviaristatus, generating 171,522 unigenes, which were assembled and functionally annotated against five public databases. We identified 30,668 SSRs in the E. breviaristatus transcriptome, from which 103 EST-SSR primer pairs were randomly selected. Of these, 58 pairs of amplified products of the expected size, and 18 of the amplified products were polymorphic. Model-based Bayesian clustering, the unweighted pair group method with arithmetic average (UPGMA), and principal coordinate analysis (PCoA) of 179 wild E. breviaristatus in 12 populations using these EST-SSRs were generally consistent, grouping the 12 populations into two major clades. Analysis of molecular variance (AMOVA) found 70% of the genetic variation among the 12 populations and 30% within the populations, indicating a high level of genetic differentiation (or low gene exchange) among the 12 populations. The transferability of the 58 successful EST-SSR primers to 22 related hexaploid species was 86.2-98.3%. UPGMA analysis generally grouped species with similar genome types together. Here, we developed EST-SSR markers from the transcriptome of E. breviaristatus. The transferability of these markers was evaluated, and the genetic structure and diversity of E. breviaristatus were explored. Our results provide a basis for the conservation and management of this endangered species, and the obtained molecular markers represent valuable resources for the exploration of genetic relationships among species in the Elymus genus.

A generalized Bayes framework for probabilistic clustering

Summary Loss-based clustering methods, such as k-means clustering and its variants, are standard tools for finding groups in data. However, the lack of quantification of uncertainty in the estimated clusters is a disadvantage. Model-based clustering based on mixture models provides an alternative approach, but such methods face computational problems and are highly sensitive to the choice of kernel. In this article we propose a generalized Bayes framework that bridges between these paradigms through the use of Gibbs posteriors. In conducting Bayesian updating, the loglikelihood is replaced by a loss function for clustering, leading to a rich family of clustering methods. The Gibbs posterior represents a coherent updating of Bayesian beliefs without needing to specify a likelihood for the data, and can be used for characterizing uncertainty in clustering. We consider losses based on Bregman divergence and pairwise similarities, and develop efficient deterministic algorithms for point estimation along with sampling algorithms for uncertainty quantification. Several existing clustering algorithms, including k-means, can be interpreted as generalized Bayes estimators in our framework, and thus we provide a method of uncertainty quantification for these approaches, allowing, for example, calculation of the probability that a data point is well clustered.

Molecular insights into the genetic diversity and population structure of Artemisia annua L. as revealed by insertional polymorphisms.

The online version contains supplementary material available at 10.1007/s40415-022-00860-x.

Molecular and morphological analyses support recognition of Prostanthera volucris (Lamiaceae), a new species from the Central Tablelands of New South Wales

Research into the systematics of Prostanthera recently revealed close evolutionary relationship among P. phylicifolia sens. str., the critically endangered P. gilesii, and a population of uncertain identity from the Central Tablelands of New South Wales (NSW), Australia. Previous analyses were unable to establish whether genetic boundaries separated these taxa. This study assessed species boundaries among these three taxa by using a combination of single-nucleotide polymorphisms (SNPs) sampled at the population-scale and multivariate analysis of morphological characters. Ordination, model-based clustering, F-statistics, neighbour-network analysis, phylogenetic analysis, and ancestry coefficient estimates all provided support for discrete genetic differences among the three taxa. Morphological phenetic analysis recovered congruent morphological clusters and identified a suite of corresponding diagnostic characters. This congruence of molecular and morphological evidence supports the presence of three independently evolving lineages, two of which correspond with the previously described P. gilesii and P. phylicifolia sens. str. The third taxon, represented by a single population from the Central Tablelands of NSW, is here described as P. volucris R.P.O’Donnell. A detailed description, diagnostic line drawings and photographs are provided. We evaluate P. volucris as satisfying criteria to be considered Critically Endangered.

Online Robust Subspace Clustering With Application to Power Grid Monitoring

In this work, a robust subspace clustering algorithm is developed to exploit the inherent union-of-subspaces structure in the data for reconstructing missing measurements and detecting anomalies. Our focus is on processing an incessant stream of large-scale data such as synchronized phasor measurements in the power grid, which is challenging due to computational complexity, memory requirement, and missing and corrupt observations. In order to mitigate these issues, a low-rank representation (LRR) model-based subspace clustering problem is formulated that can handle missing measurements and sparse outliers in the data. Then, an efficient online algorithm is derived based on stochastic approximation. The convergence property of the algorithm is established. Strategies to maintain a representative yet compact dictionary for capturing the subspace structure are also proposed. The developed method is tested on both simulated and real phasor measurement unit (PMU) data to verify the effectiveness and is shown to significantly outperform existing algorithms based on simple low-rank structure of data.

Model-Based Clustering of Mixed Data With Sparse Dependence

Model-Based Clustering of Mixed Data With Sparse Dependence

Robust TPMB Filtering Using Sensors With Limited Sensing Range Under Nonuniform Clutter Background

This study proposes a novel solution using sensors with limited sensing range to estimate the random appearance or disappearance of multi-object trajectories with unknown object detection profiles under nonuniform clutter background. Specifically, assuming that no objects occur within the sensors’ observable area, the unknown clutter intensity involving the average false alarms per scan and nonuniform clutter density is first estimated by adopting the joint combination of maximum likelihood (ML) and model-based clustering methods. Then, the unknown object detection profile can be calculated by marginalizing the fluctuating signal-to-noise ratio (SNR) in a random manner. Further, the TPMB filter is utilized to estimate trajectories from the first principle, after which the estimated parameters of clutter intensity and the detection probability are fed to the TPMB filtering, thereby improving the performance on completeness and robustness of estimation trajectories. Simulation and experimental results demonstrated that our proposed solution exhibits excellent tracking accuracy compared with state-of-the-art solutions involving robust MS generalized LMB (R-MS-GLMB) and dynamic parameter GLMB (DP-GLMB) filters.

Towards a typology of risk preference: Four risk profiles describe two-thirds of individuals in a large sample of the U.S. population

It has been a longstanding goal of the behavioral sciences to measure and model people’s risk preferences. In this article, we adopt a novel theoretical perspective of doing so and test to what extent specific types of individuals share similar risk profiles (i.e., configurations of multidimensional risk preferences). To this end, we analyzed data of a U.S. sample (N = 3,123) in a comprehensive and rigorous way, resulting in a twofold contribution. First, based on data from the Domain-Specific Risk-Taking scale (DOSPERT) and using a cross-validation procedure, we established a multidimensional trait space including general and domain-specific dimensions of risk preference. Second, we employed model-based cluster analyses in this multidimensional trait space, finding that 66% of participants can be described well with four basic risk profiles. In sum, the typological perspective proposed in this article has important implications for current theories of risk preference and the measurement of individual differences therein.

Allergic Polysensitization Clusters: Newly Recognized Severity Marker in Urban Asthmatic Adults

Introduction: While reliable, quantitative in vitro testing for sensitivity to aeroallergens has been available for decades, such information has largely been ignored in clustering analyses of asthma. Our aim is to explore allergic polysensitization as a possible marker of asthma severity and, as such, to be considered as an integral marker in future asthma clustering analyses. Methods: We constructed a database of sensitizations to the 25 aeroallergens in our geographic area (zone 1, Northeastern US) using the ImmunoCAP® in vitro assay. We used the Scikit-Learn® machine learning library for model-based clustering to identify allergic polysensitization clusters. Clusters were compared for differences in common office-based clinical markers of asthma. Results: The database consisted of 509 patients. Unbiased machine learning identified ten clusters of increasing allergic polysensitization of varying sizes (n = 1–339) characterized by significant increases in mean serum immunoglobulin E (p < 0.001), peripheral blood eosinophil count (p < 0.001), and D_LCO (p = 0.02). There was a significant decline in mean age at presentation (p < 0.001), FEV₁/FVC (p = 0.01), and FEF_25-75 (p = 0.002) with increasing allergic polysensitization. Finally, we identified two divergent paths for the poly-atopic march, one driven by perennial and the other by seasonal allergens. Conclusion: This pilot study showed that allergic polysensitization, using readily available qualitative and quantitative in vitro sensitization data, largely ignored in cluster analyses to date, may add further clinical precision in cluster analyses of asthma. We suggest the methods used here can be applied and tested using larger databases and aeroallergens present in diverse geographic regions.

Model-based clustering via mixtures of unrestricted skew normal factor analyzers with complete and incomplete data

Model-based clustering via mixtures of unrestricted skew normal factor analyzers with complete and incomplete data

DECT-CLUST: Dual-Energy CT Image Clustering and Application to Head and Neck Squamous Cell Carcinoma Segmentation.

Dual-energy computed tomography (DECT) is an advanced CT computed tomography scanning technique enabling material characterization not possible with conventional CT scans. It allows the reconstruction of energy decay curves at each 3D image voxel, representing varied image attenuation at different effective scanning energy levels. In this paper, we develop novel unsupervised learning techniques based on mixture models and functional data analysis models to the clustering of DECT images. We design functional mixture models that integrate spatial image context in mixture weights, with mixture component densities being constructed upon the DECT energy decay curves as functional observations. We develop dedicated expectation-maximization algorithms for the maximum likelihood estimation of the model parameters. To our knowledge, this is the first article to develop statistical functional data analysis and model-based clustering techniques to take advantage of the full spectral information provided by DECT. We evaluate the application of DECT to head and neck squamous cell carcinoma. Current image-based evaluation of these tumors in clinical practice is largely qualitative, based on a visual assessment of tumor anatomic extent and basic one- or two-dimensional tumor size measurements. We evaluate our methods on 91 head and neck cancer DECT scans and compare our unsupervised clustering results to tumor contours traced manually by radiologists, as well as to several baseline algorithms. Given the inter-rater variability even among experts at delineating head and neck tumors, and given the potential importance of tissue reactions surrounding the tumor itself, our proposed methodology has the potential to add value in downstream machine learning applications for clinical outcome prediction based on DECT data in head and neck cancer.

Poisson hurdle model-based method for clustering microbiome features.

High-throughput sequencing technologies have greatly facilitated microbiome research and have generated a large volume of microbiome data with the potential to answer key questions regarding microbiome assembly, structure and function. Cluster analysis aims to group features that behave similarly across treatments, and such grouping helps to highlight the functional relationships among features and may provide biological insights into microbiome networks. However, clustering microbiome data are challenging due to the sparsity and high dimensionality. We propose a model-based clustering method based on Poisson hurdle models for sparse microbiome count data. We describe an expectation-maximization algorithm and a modified version using simulated annealing to conduct the cluster analysis. Moreover, we provide algorithms for initialization and choosing the number of clusters. Simulation results demonstrate that our proposed methods provide better clustering results than alternative methods under a variety of settings. We also apply the proposed method to a sorghum rhizosphere microbiome dataset that results in interesting biological findings. R package is freely available for download at https://cran.r-project.org/package=PHclust. Supplementary data are available at Bioinformatics online.