Model-based Clustering Research Articles

Bridging the Gap Between Qualitative and Quantitative Assessment in Science Education Research with Machine Learning — A Case for Pretrained Language Models-Based Clustering

Science education researchers typically face a trade-off between more quantitatively oriented confirmatory testing of hypotheses, or more qualitatively oriented exploration of novel hypotheses. More recently, open-ended, constructed response items were used to combine both approaches and advance assessment of complex science-related skills and competencies. For example, research in assessing science teachers’ noticing and attention to classroom events benefitted from more open-ended response formats because teachers can present their own accounts. Then, open-ended responses are typically analyzed with some form of content analysis. However, language is noisy, ambiguous, and unsegmented and thus open-ended, constructed responses are complex to analyze. Uncovering patterns in these responses would benefit from more principled and systematic analysis tools. Consequently, computer-based methods with the help of machine learning and natural language processing were argued to be promising means to enhance assessment of noticing skills with constructed response formats. In particular, pretrained language models recently advanced the study of linguistic phenomena and thus could well advance assessment of complex constructs through constructed response items. This study examines potentials and challenges of a pretrained language model-based clustering approach to assess preservice physics teachers’ attention to classroom events as elicited through open-ended written descriptions. It was examined to what extent the clustering approach could identify meaningful patterns in the constructed responses, and in what ways textual organization of the responses could be analyzed with the clusters. Preservice physics teachers (N = 75) were instructed to describe a standardized, video-recorded teaching situation in physics. The clustering approach was used to group related sentences. Results indicate that the pretrained language model-based clustering approach yields well-interpretable, specific, and robust clusters, which could be mapped to physics-specific and more general contents. Furthermore, the clusters facilitate advanced analysis of the textual organization of the constructed responses. Hence, we argue that machine learning and natural language processing provide science education researchers means to combine exploratory capabilities of qualitative research methods with the systematicity of quantitative methods.

Clustering trends of melanoma incidence and mortality: Aworldwide assessment from 1995 to 2019.

Melanoma is a skin cancer that has higher mortality among cutaneous malignant tumours. Tumour incidence in most world regions is rising. Here, we intend to highlight these trends in the world. We used the global burden of disease dataset to assess the incidence and mortality of melanoma from 1995 to 2019 in 204 countries/territories. To identify country/territory groups with similar trends of melanoma incidence and mortality, a model-based clustering with a mixture of multivariate t-distributions was used. Australia and New Zealand had the largest incidence and mortality rates. Men in Egypt and women in Sri Lanka and Guam had the lowest incidence and mortality rates. The clustering analysis revealed 4 classes of incidence in both gender and 3 and 2 classes of mortality in male and female, respectively. All groups had a growing incidence rate in both gender, similar to the worldwide trend. In 10% of the countries/territories, central and Western Europe, the incidence increase rate was greater than in the countries/territories with the greatest incidence. The total mortality rate for men was steadily growing, although it was decreasing in Asian and African countries/territories. Overall, the mortality rate for women remained relatively steady over time. Due to the increasing trend of melanoma in the world, primary and secondary prevention of this disease, especially in areas with higher incidence and mortality, is essential. Raising awareness about the disease is helpful in prevention and early detection of melanoma.

Temperament and character differences in psychopathic and non-psychopathic antisocial adolescents

The present study tested the possibility of distinguishing between different types of antisocial adolescents based on psychopathic characteristics, and the differences between antisocial adolescents of different ages and subtypes in relation to the Cloninger's personality dimension. The sample included 101 antisocial male adolescents, divided into two age groups (71 respondents in the 13-17 age group and 30 in the 18-25 age group). After conducting model-based cluster analyses, non-psychopathic type (without pronounced signs of psychopathy) and psychopathic type (with pronounced antisocial, lifestyle and interpersonal facet) (Hare’s model) were singled out in the whole sample. Within the psychopathic type, in comparison with juveniles, older adolescents showed a significantly lower expression of character dimensions of Self-directedness and Cooperativeness, which are key in determining all personality disorders. Within the group of juvenile offenders, members of the two subtypes did not differ significantly in temperament and character, which indicates that juveniles are in the process of personality development, especially when it comes to their character, so they are more susceptible to treatment and their outcomes are uncertain. The results suggest that the inflation of psychopathic scores may occur in juveniles due to the identification of developmental features of adolescence (impulsivity, immaturity) as psychopathic, which implies the need to apply basic personality models in the assessment of antisocial adolescent.

Rapid Detection and Quantification of Adulterants in Fruit Juices Using Machine Learning Tools and Spectroscopy Data.

Fruit juice production is one of the most important sectors in the beverage industry, and its adulteration by adding cheaper juices is very common. This study presents a methodology based on the combination of machine learning models and near-infrared spectroscopy for the detection and quantification of juice-to-juice adulteration. We evaluated 100% squeezed apple, pineapple, and orange juices, which were adulterated with grape juice at different percentages (5%, 10%, 15%, 20%, 30%, 40%, and 50%). The spectroscopic data have been combined with different machine learning tools to develop predictive models for the control of the juice quality. The use of non-supervised techniques, specifically model-based clustering, revealed a grouping trend of the samples depending on the type of juice. The use of supervised techniques such as random forest and linear discriminant analysis models has allowed for the detection of the adulterated samples with an accuracy of 98% in the test set. In addition, a Boruta algorithm was applied which selected 89 variables as significant for adulterant quantification, and support vector regression achieved a regression coefficient of 0.989 and a root mean squared error of 1.683 in the test set. These results show the suitability of the machine learning tools combined with spectroscopic data as a screening method for the quality control of fruit juices. In addition, a prototype application has been developed to share the models with other users and facilitate the detection and quantification of adulteration in juices.

Clustering compositional data using Dirichlet mixture model.

A model-based clustering method for compositional data is explored in this article. Most methods for compositional data analysis require some kind of transformation. The proposed method builds a mixture model using Dirichlet distribution which works with the unit sum constraint. The mixture model uses a hard EM algorithm with some modification to overcome the problem of fast convergence with empty clusters. This work includes a rigorous simulation study to evaluate the performance of the proposed method over varied dimensions, number of clusters, and overlap. The performance of the model is also compared with other popular clustering algorithms often used for compositional data analysis (e.g. KMeans, Gaussian mixture model (GMM) Gaussian Mixture Model with Hard EM (Hard GMM), partition around medoids (PAM), Clustering Large Applications based on Randomized Search (CLARANS), Density-Based Spatial Clustering of Applications with Noise (DBSCAN) etc.) for simulated data as well as two real data problems coming from the business and marketing domain and physical science domain, respectively. The study has shown promising results exploiting different distributional patterns of compositional data.

A model-based clustering of expectation–maximization and K -means algorithms in crime hotspot analysis

Hotspot analysis of spatial attributes is a persistent research field in data mining, and applying a model-based clustering procedure is increasingly becoming popular in identifying trends and patterns in datasets on crime events occurring in space. The distributions of potential crime hotspots are parameterized as arising from Gaussian multivariate distributions, whose parameters are estimated by the expectation–maximization (E-M) algorithm, an iterative process with convergence very sensitive to initializations. In this study, a model-based clustering algorithm is explored from the E-M algorithm, initialized by K-means clustering using geodesic distance classification to estimate the model parameters and compared with the classical E-M algorithm, initialized with hierarchical clustering, to identify the distributional patterns of incidence of criminal activities. These model-based clustering algorithms were demonstrated on an open-source large dataset of violent crime activities, which occurred in West Midlands County. Training the data as a Gaussian process, the study identified 12 hotspots of Gaussian mixed models as clusters of an ellipsoidal distribution varying in shape, volume, and orientation, which are mostly found in central parts of boroughs of the study area. The proposed model-based clustering of the E-M algorithm combined with K-means clustering algorithm proved efficient as being fast and stable in convergence with low probability of uncertainty by classifications, producing same classification in some cases when compared to that of the classical E-M and K-means algorithms. The combined model-based clustering techniques applied in the hotspot analysis of criminal activities in space will not only provide insight into crime prediction and resource allocation in combating strategies but also guide researchers to adopt mechanisms for modeling large spatial attributes in data mining.

KLASTERISASI PROVINSI DI INDONESIA BERDASARKAN FAKTOR PENYEBARAN COVID-19 MENGGUNAKAN MODEL-BASED CLUSTERING t-MULTIVARIAT

The spread of Covid-19 had a significant impact in all sectors. Enforcement policies from the government that are appropriate with the conditions for the spread of the virus that are needed to prevent a bigger impact. Clusteritation by province based on data on the spread of Covid-19 is important for the government to set appropriate policies in order to prevent the spread of Covid-19. The data used include data on population density, testing rate, proportion of population 50 years and over, and proportion of population diligently hand-washing in each province. The data factors for the spread of Covid-19 tend to overlap and there are outliers in the data which causes the data not normally distributed. In this study, Model-Based Clustering t-multivariate was used for data clustering. The results show that using Integrated Completed Likelihood, two groups of optimal cluster were obtained. The second cluster has a higher risk of spreading Covid-19 than the first cluster. Keywords : Covid-19, Clustering, Model-Based Clustering t-Multivariat

Discovering Thematically Coherent Biomedical Documents Using Contextualized Bidirectional Encoder Representations from Transformers-Based Clustering.

The increasing expansion of biomedical documents has increased the number of natural language textual resources related to the current applications. Meanwhile, there has been a great interest in extracting useful information from meaningful coherent groupings of textual content documents in the last decade. However, it is challenging to discover informative representations and define relevant articles from the rapidly growing biomedical literature due to the unsupervised nature of document clustering. Moreover, empirical investigations demonstrated that traditional text clustering methods produce unsatisfactory results in terms of non-contextualized vector space representations because that neglect the semantic relationship between biomedical texts. Recently, pre-trained language models have emerged as successful in a wide range of natural language processing applications. In this paper, we propose the Gaussian Mixture Model-based efficient clustering framework that incorporates substantially pre-trained (Bidirectional Encoder Representations from Transformers for Biomedical Text Mining) BioBERT domain-specific language representations to enhance the clustering accuracy. Our proposed framework consists of main three phases. First, classic text pre-processing techniques are used biomedical document data, which crawled from the PubMed repository. Second, representative vectors are extracted from a pre-trained BioBERT language model for biomedical text mining. Third, we employ the Gaussian Mixture Model as a clustering algorithm, which allows us to assign labels for each biomedical document. In order to prove the efficiency of our proposed model, we conducted a comprehensive experimental analysis utilizing several clustering algorithms while combining diverse embedding techniques. Consequently, the experimental results show that the proposed model outperforms the benchmark models by reaching performance measures of Fowlkes mallows score, silhouette coefficient, adjusted rand index, Davies-Bouldin score of 0.7817, 0.3765, 0.4478, 1.6849, respectively. We expect the outcomes of this study will assist domain specialists in comprehending thematically cohesive documents in the healthcare field.

Spinal Cord Stimulation–Naïve Patients vs Patients With Failed Previous Experiences With Standard Spinal Cord Stimulation: Two Distinct Entities or One Population?

IntroductionNowadays, the success of spinal cord stimulation (SCS) is evaluated separately in patients who have previous experiences with standard SCS and in SCS-naïve patients. Nevertheless, it is yet to be evaluated whether both patient groups are effectively distinct patient groups. Therefore, the aims of this study are twofold: 1) Are there clusters in the data to distinguish between both patient groups? 2) Can we discriminate both patient groups based on routinely collected clinical parameters? Materials and MethodsBaseline data from the Discover study were used, in which 263 patients with persistent spinal pain syndrome type 2 were included (185 neurostimulation-naïve patients and 78 patients with previous SCS experience). Pain intensity scores for low back and leg pain, functional disability, medication use, and health-related quality of life utility scores were used in the analysis. Model-based clustering was performed on standardized data. Discriminant analysis was performed with linear and quadratic discriminant analysis, with leave-one-out cross-validation to evaluate model performance. ResultsModel-based clustering revealed two different clusters in the data. None of the clusters clearly separated SCS-naïve patients from patients with previous SCS experience. Linear discriminant analysis resulted in a leave-one-out cross-validation error rate of 30.0% to discriminate between both patient groups, based on routinely collected clinical parameters. ConclusionsClustering analysis did not result in clusters that separate SCS-naïve patients from patients with previous SCS experience. This may suggest that both patient groups should not be considered as two different patient groups when comparing them on routine clinical parameters, with potentially profound implications for research and clinical settings. Clinical Trial RegistrationThe Clinicaltrials.gov registration number for the Discover study is NCT02787265.

Identification and characterization of chickpea genotypes for early flowering and higher seed germination through molecular markers.

Chickpea is the fourth most important legume crop contributing 15.42% to the total legume production and a rich source of proteins, minerals, and vitamins. Determination of genetic diversity of wild and elite cultivars coupled with early flowering and higher seed germination lines are quintessential for variety improvement. In the present study, we have analyzed the genetic diversity, population structure, cross-species transferability, and allelic richness in 50 chickpea collections using 23 Inter simple sequence repeats (ISSR) markers. The observed parameters such as allele number varied from 3 to 16, range of allele size varied from 150 to 1600bp and polymorphic information content (PIC) range lies in between 0.15 and 0.49. Dendrogram was constructed with ISSR marker genotypic data and classified 50 chickpea germplasms into groups I and II, where the accession P 74 - 1 is in group I and the rest are in group II. Dendrogram, Principal component analysis (PCA), dissimilarity matrix, and Bayesian model-based genetic clustering of 50 chickpea germplasms revealed that P 74 - 1 and P 1883 are very diverse chickpea accessions. Based on genetic diversity analysis, 15 chickpea germplasm having been screened for early flowering and higher seed germination and found that the P 1857-1 and P 3971 have early flowering and higher seed germination percentage in comparison to P 1883 and other germplasm. These agronomic traits are essential for crop improvement and imply the potential of ISSR markers in crop improvement.

Multiple change point clustering of count processes with application to California COVID data

• Analysis of 275-day long time series of county level COVID-19 data in California state. • Multiple change point estimation in the framework of mixture modeling and model-based clustering. • Change point given by gap between logit transformed segments in negative binomial nonhomogeneous Levy process. • Three geographically meaningful clusters, each with several change points indicating the spread and decline of infection. In this paper, a model-based clustering algorithm relying on a finite mixture of negative binomial Lévy processes is proposed. The algorithm models heterogeneous stochastic count process data and automatically estimates multiple change points upon fitting the mixture model. Such change point estimation identifies time points when deviation from the standard process has occurred and serves as an important diagnostic tool for analyzing temporal data. The proposed model is applied to the COVID-positive ICU cases in the state of California with very interesting results.

Correction of multiple-blinking artifacts in photoactivated localization microscopy.

Photoactivated localization microscopy (PALM) produces an array of localization coordinates by means of photoactivatable fluorescent proteins. However, observations are subject to fluorophore multiple blinking and each protein is included in the dataset an unknown number of times at different positions, due to localization error. This causes artificial clustering to be observed in the data. We present a 'model-based correction' (MBC) workflow using calibration-free estimation of blinking dynamics and model-based clustering to produce a corrected set of localization coordinates representing the true underlying fluorophore locations with enhanced localization precision, outperforming the state of the art. The corrected data can be reliably tested for spatial randomness or analyzed by other clustering approaches, and descriptors such as the absolute number of fluorophores per cluster are now quantifiable, which we validate with simulated data and experimental data with known ground truth. Using MBC, we confirm that the adapter protein, the linker for activation of T cells, is clustered at the T cell immunological synapse.

Anderson relaxation test for intrinsic dimension selection in model-based clustering

Parsimonious finite mixture models often require the a priori selection of desired model dimensionality. For example, projection-based parsimonious models demand the dimension of the subspace for projection. Other models ask for their own structural restrictions on parameters. The subspace clustering framework is a projection-based parsimonious model for various finite mixtures, including the Gaussian variant. The existing dimension selection methods for subspace clustering are ad-hoc or potentially computationally prohibitive, creating a need for a principled, yet computationally lightweight, approach. In light of this problem, a hypothesis test-based intrinsic dimension estimation method called the Anderson Relaxation Test (ART) is introduced, and its performance is examined in both simulated and real data settings.

Comprehensive analysis of the associations between clinical factors and outcomes by machine learning, using post marketing surveillance data of cabazitaxel in patients with castration-resistant prostate cancer

BackgroundWe aimed to evaluate relationships between clinical outcomes and explanatory variables by network clustering analysis using data from a post marketing surveillance (PMS) study of castration-resistant prostate cancer (CRPC) patients.MethodsThe PMS was a prospective, multicenter, observational study of patients with metastatic, docetaxel-refractory CRPC treated with cabazitaxel in Japan after its launch in 2014. Graphical Markov (GM) model-based simulations and network clustering in ‘R’ package were conducted to identify correlations between clinical factors and outcomes. Factors shown to be associated with overall survival (OS) in the machine learning analysis were confirmed according to the clinical outcomes observed in the PMS.ResultsAmong the 660 patients analyzed, median patient age was 70.0 years, and median OS and time-to-treatment failure (TTF) were 319 and 116 days, respectively. In GM-based simulations, factors associated with OS were liver metastases, performance status (PS), TTF, and neutropenia (threshold 0.05), and liver metastases, PS, and TTF (threshold 0.01). Factors associated with TTF were OS and relative dose intensity (threshold 0.05), and OS (threshold 0.01). In network clustering in ‘R’ package, factors associated with OS were number of treatment cycles, discontinuation due to disease progression, and TTF (threshold 0.05), and liver and lung metastases, PS, discontinuation due to adverse events, and febrile neutropenia (threshold 0.01). Kaplan–Meier analysis of patient subgroups demonstrated that visceral metastases and poor PS at baseline were associated with worse OS, while neutropenia or febrile neutropenia and higher number of cabazitaxel cycles were associated with better OS.ConclusionsNeutropenia may be a predictive factor for treatment efficacy in terms of survival. Poor PS and distant metastases to the liver and lungs were shown to be associated with worse outcomes, while factors related to treatment duration were shown to positively correlate with better OS.

Minimal subphenotyping model for acute heart failure with preserved ejection fraction.

AimsApplication of the latent class analysis to acute heart failure with preserved ejection fraction (HFpEF) showed that the heterogeneous acute HFpEF patients can be classified into four distinct phenotypes with different clinical outcomes. This model‐based clustering required a total of 32 variables to be included. However, this large number of variables will impair the clinical application of this classification algorithm. This study aimed to identify the minimal number of variables for the development of optimal subphenotyping model.Methods and resultsThis study is a post hoc analysis of the PURSUIT‐HFpEF study (N = 1095), a prospective, multi‐referral centre, observational study of acute HFpEF [UMIN000021831]. We previously applied the latent class analysis to the PURSUIT‐HFpEF dataset and established the full 32‐variable model for subphenotyping. In this study, we used the Cohen's kappa statistic to investigate the minimal number of discriminatory variables needed to accurately classify the phenogroups in comparison with the full 32‐variable model. Cohen's kappa statistic of the top‐X number of discriminatory variables compared with the full 32‐variable derivation model showed that the models with ≥16 discriminatory variables showed kappa value of >0.8, suggesting that the minimal number of discriminatory variables for the optimal phenotyping model was 16. The 16‐variable model consists of C‐reactive protein, creatinine, gamma‐glutamyl transferase, brain natriuretic peptide, white blood cells, systolic blood pressure, fasting blood sugar, triglyceride, clinical scenario classification, infection‐triggered acute decompensated HF, estimated glomerular filtration rate, platelets, neutrophils, GWTG‐HF (Get With The Guidelines‐Heart Failure) risk score, chronic kidney disease, and CONUT (Controlling Nutritional Status) score. Characteristics and clinical outcomes of the four phenotypes subclassified by the minimal 16‐variable model were consistent with those by the full 32‐variable model. The four phenotypes were labelled based on their characteristics as ‘rhythm trouble’, ‘ventricular‐arterial uncoupling’, ‘low output and systemic congestion’, and ‘systemic failure’, respectively.ConclusionsThe phenotyping model with top 16 variables showed almost perfect agreement with the full 32‐variable model. The minimal model may enhance the future clinical application of this clustering algorithm.

Genome-wide survey on three local horse populations with a focus on runs of homozygosity pattern.

Purosangue Orientale Siciliano, Sanfratellano and Siciliano represent the Sicilian equine genetic resource. This study aimed to investigate the genetic diversity, population structure and the pattern of autozygosity of Sicilian horse populations using genome‐wide single‐nucleotide polymorphism (SNP) data generated with the Illumina Equine SNP70 array. The genotyping data of 17 European and Middle East populations were also included in the study. The patterns of genetic differentiation, model‐based clustering and Neighbour‐Net showed the expected positioning of Sicilian populations within the wide analysed framework and the close connections between the Purosangue Orientale Siciliano and the Arab as well as between Sanfratellano, Siciliano and Maremmano. The highest expected heterozygosity (H e) and contemporary effective population size (cNe) were reported in Siciliano (H e = 0.323, cNe = 397), and the lowest were reported in Purosangue Orientale Siciliano (H e = 0.277, cNe = 10). The analysis of the runs of homozygosity and the relative derived inbreeding revealed high internal homogeneity in Purosangue Orientale Siciliano and Arab horses, intermediate values in Maremmano and Sanfratellano and high heterogeneity in the Siciliano population. The genome‐wide SNP analysis showed the selective pressure on Purosangue Orientale Siciliano towards traits related to endurance performance. Our results underline the importance of planning adequate conservation and exploitation programmes to reduce the level of inbreeding and, therefore, the loss of genetic diversity.

Understanding tourist behaviour towards destination selection based on social media information: an evaluation using unsupervised clustering algorithms

PurposeThe study aims to cluster the travellers based on their social media interactions as well as to find the different segments with similar and dissimilar categories according to traveller's choice. The study also aims to understand the behaviour of clusters of the travellers towards destination selection and accordingly make the tour packages in order to improve tourists' satisfaction and gain viable benefits.Design/methodology/approachAgglomerative hierarchical clustering with Ward's minimum variance linkage algorithm and model-based clustering with parameterized finite Gaussian mixture models has been implemented to achieve the respective goals. The dimension reduction (DR) technique was introduced for better visualizing clustering structure obtained from a finite mixture of Gaussian densities.FindingsA total of 980 travellers have been clustered into 8 different interest groups according to their tourism destinations selection across East Asia based on individual social media feedback. For selecting the optimal number of clusters as well as the behaviour of the interested travellers groups, both these proposed methods have shown remarkable similarities. DR technique ensures the reduction in dimensionality with seven directions, of which the first two directions explained 95% of total variability.Practical implicationsTourism organizations focus on marketing efforts to promote the most attractive benefits to the clusters of travellers. By segmenting travellers of East Asia into homogeneous groups, it is feasible to choose a similar area to test different marketing techniques. Finally, it can be identified to which segments, new respondents or potential clients belong; consequently, the tourism organizations can design the tour packages.Originality/valueThe study has uniqueness in two aspects. Firstly, the study empirically revealed tourists' experience and behavioural intention to select tourism destinations and secondly, it finds quantifiable insights into the tourism phenomenon in East Asia, which helps tourism organizations to understand the buying behaviours of tourists' segments. Finally, the application of clustering algorithms to achieve the purpose of this study and the findings are very new in the literature on tourism, to understand the tourist behaviour towards destination selection based on social media reviews.

Model-based clustering of high-dimensional longitudinal data via regularization.

We propose a model-based clustering method for high-dimensional longitudinal data via regularization in this paper. This study was motivated by the Trial of Activity in Adolescent Girls (TAAG), which aimed to examine multilevel factors related to the change of physical activity by following up a cohort of 783 girls over 10 years from adolescence to early adulthood. Our goal is to identify the intrinsic grouping of subjects with similar patterns of physical activity trajectories and the most relevant predictors within each group. The previous analyses conducted clustering and variable selection in two steps, while our new method can perform the tasks simultaneously. Within each cluster, a linear mixed-effects model (LMM) is fitted with a doubly penalized likelihood to induce sparsity for parameter estimation and effect selection. The large-sample joint properties are established, allowing the dimensions of both fixed and random effects to increase at an exponential rate of the sample size, with a general class of penalty functions. Assuming subjects are drawn from a Gaussian mixture distribution, model effects and cluster labels are estimated via a coordinate descent algorithm nested inside the Expectation-Maximization (EM) algorithm. Bayesian Information Criterion (BIC) is used to determine the optimal number of clusters and the values of tuning parameters. Our numerical studies show that the new method has satisfactory performance and is able to accommodate complex data with multilevel and/or longitudinaleffects.

Global Patterns of Contemporary Welfare States

AbstractThis study proposes a novel and systematic theoretical framework to explain global welfare state policy differences. The existing scholarship examined ample welfare state variations, reforms, and transitions; however, it is typically limited to specific countries, regions, policies, or risks. In an endeavor to combine these theoretical and empirical insights, the global contemporary welfare state patterns remain vague. This study aims at bridging this gap in the literature by deploying an orderly and comprehensive three-step procedure. First, I formally design a three-stage global yet comparative conceptual framework that ensures consistency, inclusiveness, and compliance. Second, based on this framework, I assemble a unique comparative dataset for one-hundred-fifty countries, some of which appear for the first time in this literature. Third, I validate the framework using an advanced data reduction method named model-based cluster analysis. The results of this study demonstrate that global contemporary welfare states follow systematically divergent paths, revealing Proactive, Reactive, and Dual patterns.

Clinical phenotyping in sarcoidosis using cluster analysis

BackgroundMost phenotyping paradigms in sarcoidosis are based on expert opinion; however, no paradigm has been widely adopted because of the subjectivity in classification. We hypothesized that cluster analysis could be performed on common clinical variables to define more objective sarcoidosis phenotypes.MethodsWe performed a retrospective cohort study of 554 sarcoidosis cases to identify distinct phenotypes of sarcoidosis based on 29 clinical features. Model-based clustering was performed using the VarSelLCM R package and the Integrated Completed Likelihood (ICL) criteria were used to estimate number of clusters. To identify features associated with cluster membership, features were ranked based on variable importance scores from the VarSelLCM model, and additional univariate tests (Fisher’s exact test and one-way ANOVA) were performed using q-values correcting for multiple testing. The Wasfi severity score was also compared between clusters.ResultsCluster analysis resulted in 6 sarcoidosis phenotypes. Salient characteristics for each cluster are as follows: Phenotype (1) supranormal lung function and majority Scadding stage 2/3; phenotype (2) supranormal lung function and majority Scadding stage 0/1; phenotype (3) normal lung function and split Scadding stages between 0/1 and 2/3; phenotype (4) obstructive lung function and majority Scadding stage 2/3; phenotype (5) restrictive lung function and majority Scadding stage 2/3; phenotype (6) mixed obstructive and restrictive lung function and mostly Scadding stage 4. Although there were differences in the percentages, all Scadding stages were encompassed by all of the phenotypes, except for phenotype 1, in which none were Scadding stage 4. Clusters 4, 5, 6 were significantly more likely to have ever been on immunosuppressive treatment and had higher Wasfi disease severity scores.ConclusionsCluster analysis produced 6 sarcoidosis phenotypes that demonstrated less severe and severe phenotypes. Phenotypes 1, 2, 3 have less lung function abnormalities, a lower percentage on immunosuppressive treatment and lower Wasfi severity scores. Phenotypes 4, 5, 6 were characterized by lung function abnormalities, more parenchymal abnormalities, an increased percentage on immunosuppressive treatment and higher Wasfi severity scores. These data support using cluster analysis as an objective and clinically useful way to phenotype sarcoidosis subjects and to empower clinicians to identify those with more severe disease versus those who have less severe disease, independent of Scadding stage.