Cluster Assignment Research Articles

Deep Incomplete Multi-view Clustering via Multi-level Imputation and Contrastive Alignment

Deep incomplete multi-view clustering (DIMVC) aims to enhance clustering performance by capturing consistent information from incomplete multiple views using deep models. Most existing DIMVC methods typically employ imputation-based strategies to handle missing views before clustering. However, they often assume complete data availability across all views, overlook potential low-quality views, and perform imputation at a single data level, leading to challenges in accurately inferring missing data. To address these issues, we propose a novel imputation-based approach called Multi-level Imputation and Contrastive Alignment (MICA) to simultaneously improve imputation quality and boost clustering performance. Specifically, MICA employs an individual deep model for each view, which unifies view feature learning and cluster assignment prediction. It leverages the learned features from available instances to construct an adaptive cross-view graph for reliable view selection. Guided by these reliable views, MICA performs multi-level (feature-level, data-level, and reconstruction-level) imputation to preserve topological structures across levels and ensure accurate missing feature inference. The complete features are then used for discriminative cluster assignment learning. Additionally, an instance- and cluster-level contrastive alignment is conducted on the cluster assignments to further enhance semantic consistency across views. Experimental results show the effectiveness and superior performance of the proposed MICA method.

Ensemble Predictability of Week 3/4 Precipitation and Temperature over the United States via Cluster Analysis of the Large-Scale Circulation

Abstract Forecasting the week 3/4 period presents many challenges, resulting in a need for improvements to forecast skill. At this distance from initial conditions, numerical models struggle to present skillful forecasts of temperature, precipitation, and associated extremes. One approach to address this is to utilize more predictable large-scale circulation regimes to make forecasts of temperature and precipitation anomalies, using the association between the regimes and surface weather obtained from reanalysis products. This study explores the utility of k-means cluster analysis on geopotential heights and their ability to make skillful regime predictions in the week 3/4 period. Using 14-day running means of European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis v5 (ERA5) 500-hPa geopotential heights for the wintertime December–February (DJF) period, circulation regimes are identified using k-means clustering. Each period is assigned a cluster number, allowing the compositing of any reanalysis or observation variable to form cluster maps. Maps of 500-hPa height, 2-m temperature, precipitation, and storm-track anomalies are some of the variables composited. The utility of these relationships in a dynamical forecast setting is tested via Global Ensemble Forecast System v12 (GEFSv12) hindcasts and real-time ensemble suite forecasts. Week 3/4 deterministic and probabilistic experimental forecasts are then derived from cluster assignments using several methods. We find, via a conditional skill analysis, forecasts strongly correlated with a cluster exhibit greater skill for both dynamical model and cluster-derived forecasts. Our preliminary results represent a step forward to aid forecasters make more skillful assessments of the circulation regime and its associated surface weather for this challenging forecast time scale. Significance Statement Our paper links the local statistics of 14-day precipitation and temperature within the United States to very broad preferred patterns of winds and pressure over the wide Pacific–North American region. Such patterns, known as “circulation regimes,” provide the background that heavily influences local surface conditions. We find that forecasts for which midatmospheric anomalies are closely aligned with one or more circulation regimes are better at predicting the U.S. weather probabilities than other week 3–4 forecasts. A tool based on this finding identifies forecasts for which confidence is higher. Future work will be designed to further exploit the links between circulation regimes and weather to improve the accuracy of week 3–4 forecasts.

MvWECM: Multi-view Weighted Evidential [formula omitted]-Means clustering

Traditional multi-view clustering algorithms, designed to produce hard or fuzzy partitions, often neglect the inherent ambiguity and uncertainty in the cluster assignment of objects. This oversight may lead to performance degradation. To address these issues, this paper introduces a novel multi-view clustering method, termed MvWECM, capable of generating credal partitions within the framework of belief functions. The objective function of MvWECM is introduced considering the uncertainty in the cluster structure included in the multi-view dataset. We take into account inter-view conflict to effectively leverage coherent information across different views. Moreover, the effectiveness is heightened through the incorporation of adaptive view weights, which are customized to modulate their smoothness in accordance with their entropy. The optimization method to get the optimal credal membership and class prototypes is derived. The view wights can be also provided as a by-product. Experimental results on several real-word datasets demonstrate the effectiveness and superiority of MvWECM by comparing with some state-of-the-art methods.

Robust extraction of pneumonia-associated clinical states from electronic health records

Mining of electronic health records (EHR) promises to automate the identification of comprehensive disease phenotypes. However, the realization of this promise is hindered by the unavailability of generalizable ground-truth information, data incompleteness and heterogeneity, and the lack of generalization to multiple cohorts. We present here a data-driven approach to identify clinical states that we implement for 585 critical care patients with suspected pneumonia recruited by the SCRIPT study, which we compare to and integrate with 9,918 pneumonia patients from the MIMIC-IV dataset. We extract and curate from their structured EHRs a primary set of clinical features (53 and 59 features for SCRIPT and MIMIC-IV, respectively), including disease severity scores, vital signs, and so on, at various degrees of completeness. We aggregate irregular time series into daily frequency, resulting in 12,495 and 94,684 patient-day pairs for SCRIPT and MIMIC, respectively. We define a "common-sense" ground truth that we then use in a semisupervised pipeline to optimize choices for data preprocessing, and reduce the feature space to four principal components. We describe and validate an ensemble-based clustering method that enables us to robustly identify five clinical states, and use a Gaussian mixture model to quantify uncertainty in cluster assignment. Demonstrating the clinical relevance of the identified states, we find that three states are strongly associated with disease outcomes (dying vs. recovering), while the other two reflect disease etiology. The outcome associated clinical states provide significantly increased discrimination of mortality rates over standard approaches.

Machine-learning phenotyping of patients with functional mitral regurgitation undergoing transcatheter edge-to-edge repair - the MITRA-AI study

Abstract Background Severe functional mitral regurgitation (FMR) may benefit from transcatheter mitral valve repair (TMVR), but selection of patients remains to be optimised. Objectives The aim of this study was to use Machine-Learning (ML) approaches to uncover concealed connections between clinical, echocardiographic, and hemodynamic data associated with patients' outcomes. Methods Consecutive patients undergoing TMVR from 2009 to 2020 were included in the MITRA-AI registry. Eleven relevant clinical and echocardiographic variables were selected in a clustering-based model. The primary endpoint was a composite of cardiovascular death or heart failure hospitalisation at one year, while its single components were secondary endpoints. External validation was performed on the Mitrascore dataset and comparison between clustering and Mitrascore was performed with Net Reclassification Index (NRI). Moreover cluster assignment was performed also on a cohort of patients with moderate-to-severe MR treated with medical therapy alone. Results 822 patients were included in the derivation cohort. The composite primary endpoint occurred in 250 (30%) patients. Four clusters with decreasing risk of the primary endpoint were identified (42%, 37%, 25% and 20% from cluster 1 to cluster 4, respectively). Clusters were combined into a high-risk (clusters 1 and 2) and a low-risk phenotype (clusters 3 and 4). High-risk phenotype patients had larger LVs (> 107 ml/m2), lower LVEF (< 35%) and more prevalent ischemic aetiology (51% to 60% vs. 30% to 40%) compared to low-risk phenotype patients. Moreover, within the high-risk group, patients with diabetes mellitus and with advanced age were at increased risk. Within the low-risk group, ischemic aetiology increased the risk of cardiovascular death, while permanent atrial fibrillation amplified that of HF hospitalizations. In the Mitrascore validation cohort of 1119 patients, incidence of the primary end point varied consistently (48%, 52%, 35% and 42%). Clustering achieved a NRI in 27% of the patients in the derivation cohort and 7% in the validation group compared to Mitrascore. In the validation group of 207 patients treated medically, the incidence of the primary endpoint decreased from the first to the fourth cluster (53%, 42%, 30% and 30%). Conclusions A ML analysis identified meaningful clinical phenotypic presentations in FMR undergoing TMVR, with significant differences in terms of cardiovascular death and heart failure hospitalizations, confirmed in an external validation cohort. ML phenotyping also discriminated prognosis of medically treated patients with severe MR.Central Figure

Identifying clinical phenotype clusters for coronary artery disease: a novel approach using clustering by unsupervised machine learning

Abstract Background Guideline recommendations for the prevention of cardiovascular (CV) events in patients with coronary artery disease (CAD) are one-size-fits-all. Clinically identifiable phenotypes needing specific considerations might exist. Purpose To identify phenotypes in patients with CAD based on clinical characteristics and their relationship with the risk of new CV events and treatment benefits. Methods Unsupervised machine learning through latent class analysis (LCA) was performed on patients with CAD included in the UCC-SMART cohort (n = 5,888) and SWEDEHEART registry (n = 67,428), to identify patient clusters (i.e., phenotypes) based on clinical characteristics: age, sex, body-mass-index, diastolic blood pressure, history of diabetes, myocardial infarction, atrial fibrillation, polyvascular disease, current smoking, non-high density lipoprotein cholesterol, estimated glomerular filtration rate, and C-reactive protein. Number of clusters was determined using the Bayesian information criterion. Event free survival and hazard ratios (HR) for major adverse cardiovascular events (MACE) as primary endpoint, including non-fatal myocardial infarction, non-fatal stroke, and cardiovascular death, were estimated per cluster. The LCA were performed in each respective cohort and cross-validated in the other cohort to assess impact of variations in clustering variables on cluster reproducibility. In UCC-SMART, bootstrapping was performed the assess stability of the LCA model and impact on cluster assignment. Results In UCC-SMART four distinct phenotypes within the CAD population could be distinguished: Phenotype A (16%) of mostly males with isolated CAD, phenotype B (30%) of older patients with polyvascular disease, Phenotype C (28%) of only women with limited traditional risk factors and Phenotype D (27%) young, current smokers with a premature event (Figure 1). Cross-validation in SWEDEHEART identified roughly the same four phenotypes, i.e., phenotype A (39%) with 100% males and 3.8% polyvascular disease, phenotype B (8%) with average age of 71 years and 30% polyvascular disease, phenotype C (19%) with 100% women, and phenotype D (34%) 54 years and 25% current smokers. Notably, although age was comparable between phenotype A and C, MACE-free survival was worse for the female-only cluster. Similarly, although patients where on average 9 years younger in phenotype D when compared to phenotype A, MACE-free survival was worse (Figure 2). No significant changes in clustering models were observed in the bootstrap samples. Conclusion Phenotypes can be distinguished within patients with CAD. These phenotypes are related to differences in the risk of recurrent events. These findings suggest that specific guideline recommendations for these groups could be considered.Cluster interpretation in SMARTSurvival along cluster lines

Computational algorithm based on health and lifestyle traits to categorize lifemetabotypes in the NUTRiMDEA cohort

Classifying individuals based on metabotypes and lifestyle phenotypes using exploratory factor analyses, cluster definition, and machine-learning algorithms is promising for precision chronic disease prevention and management. This study analyzed data from the NUTRiMDEA online cohort (baseline: n = 17332 and 62 questions) to develop a clustering tool based on 32 accessible questions using machine-learning strategies. Participants ranged from 18 to over 70 years old, with 64.1% female and 35.5% male. Five clusters were identified, combining metabolic, lifestyle, and personal data: Cluster 1 (“Westernized Millennial”, n = 967) included healthy young individuals with fair lifestyle habits; Cluster 2 (“Healthy”, n = 10616) consisted of healthy adults; Cluster 3 (“Mediterranean Young Adult”, n = 2013) represented healthy young adults with a healthy lifestyle and showed the highest adherence to the Mediterranean diet; Cluster 4 (“Pre-morbid”, n = 600) was characterized by healthy adults with declined mood; Cluster 5 (“Pro-morbid”, n = 312) comprised older individuals (47% >55 years) with poorer lifestyle habits, worse health, and a lower health-related quality of life. A computational algorithm was elicited, which allowed quick cluster assignment based on responses (“lifemetabotypes”). This machine-learning approach facilitates personalized interventions and precision lifestyle recommendations, supporting online methods for targeted health maintenance and chronic disease prevention.

Self-supervised learning from images: No negative pairs, no cluster-balancing

Learning with self-derived targets provides a non-contrastive method for unsupervised image representation learning, where the variety in targets is crucial. Recent work has achieved good performance by learning with targets obtained via cluster-balancing. However, the equal-cluster-size constraint becomes too restrictive for handling data with imbalanced categories or coming in small batches. In this paper, we propose a new clustering-based approach for non-contrastive image representation learning with no need for a particular architecture design or extra memory bank and no explicit constraints on cluster size. A key formulation is to learn embedding consistency and variable decorrelation in the cluster space by tweaking the batch-wise cross-correlation matrix towards an identity one. With this identitization loss incorporated, predicted cluster assignments of two randomly augmented views of the same image serve as targets for each other. We carried out comprehensive experimental studies of linear classification with learned representations of benchmark image datasets. Our results show that the proposed approach significantly outperforms state-of-the-art approaches and is more robust to class imbalance than those with cluster balancing.

Towards the construction of regional marine radiocarbon calibration curves: an unsupervised machine learning approach

Abstract. Radiocarbon may serve as a powerful dating tool in palaeoceanography, but its accuracy is limited by the need to calibrate radiocarbon dates to calendar ages. A key problem is that marine radiocarbon dates must be corrected for past offsets from either the contemporary atmosphere (i.e. “reservoir age” offsets) or a modelled estimate of the global average surface ocean (i.e. delta-R offsets). This presents a challenge because the spatial distribution of reservoir ages and delta-R offsets can vary significantly, particularly over periods of major marine hydrographic and/or carbon cycle change such as the last deglaciation. Modern reservoir age and delta-R estimates therefore have limited applicability. While forward modelling of past R-age variability has been proposed as a means of resolving this problem, this requires accurate a priori knowledge of past global radiocarbon budget closure (i.e. production, and cycling), which we currently lack. In this context, the construction of empirical regional marine calibration curves could provide a way forward. However, the spatial reach of such calibrations and their robustness subject to (uncertain) temporal changes in climate and ocean circulation would need to be tested. Here, we use unsupervised machine learning techniques to define distinct regions of the surface ocean that exhibit coherent behaviour in terms of their radiocarbon age offsets from the contemporary atmosphere (R ages), regardless of the causes of R-age variability. We apply multiple algorithms (k-means, k-medoids, and hierarchical clustering) to outputs from two different numerical models spanning a range of climate states, forcings, and timescales of adjustment. Comparisons between the cluster assignments across model runs confirm some robust regional patterns that likely stem from constraints imposed by large-scale ocean and atmospheric physics. At the coarsest scale, regions of coherent R-age variability correspond to the major ocean basins. By further dividing basin-scale shape-based clusters into amplitude-based subclusters, we recover regional associations, such as increased high-latitude R ages, or the propagation of R-age anomalies from regions of deep mixing in the Southern Ocean to upwelling sites in the eastern equatorial Pacific, which cohere with known modern oceanographic processes. We show that the medoids (i.e. the most representative locations) for these regional sub-clusters provide significantly better approximations of simulated local R-age variability than constant offsets from the global surface average. This remains true when cluster assignments obtained from one model simulation are applied to simulated R-age time series from another. Further, model-based clusters are found to be broadly consistent with existing reservoir age reconstructions that span the last ∼30 kyr. We therefore propose that machine learning provides a promising approach to the problem of defining regions for which empirical marine radiocarbon calibration curves may eventually be generated.

Fractional Derivative to Symmetrically Extend the Memory of Fuzzy C-Means

The fuzzy C-means (FCM) clustering algorithm is a widely used unsupervised learning method known for its ability to identify natural groupings within datasets. While effective in many cases, FCM faces challenges such as sensitivity to initial cluster assignments, slow convergence, and difficulty in handling non-linear and overlapping clusters. Aimed at these limitations, this paper introduces a novel fractional fuzzy C-means (Frac-FCM) algorithm, which incorporates fractional derivatives into the FCM framework. By capturing non-local dependencies and long memory effects, fractional derivatives offer a more flexible and precise representation of data relationships, making the method more suitable for complex datasets. Additionally, a genetic algorithm (GA) is employed to optimize a new least-squares objective function that emphasizes the geometric properties of clusters, particularly focusing on the Fukuyama–Sugeno and Xie–Beni indices, thereby enhancing the balance between cluster compactness and separation. Furthermore, the Frac-FCM algorithm is evaluated on several benchmark datasets, including Iris, Seed, and Statlog, and compared against traditional methods like K-means, SOM, GMM, and FCM. The results indicate that Frac-FCM consistently outperforms these methods in terms of the Silhouette and Dunn indices. For instance, Frac-FCM achieves higher Silhouette scores of most cases, indicating more distinct and well-separated clusters. Dunn’s index further shows that Frac-FCM generates clusters that are better separated, surpassing the performance of traditional methods. These findings highlight the robustness and superior clustering performance of Frac-FCM. The Friedman test was employed to enhance and validate the effectiveness of Frac-FCM.

Contrastive clustering based on generalized bias-variance decomposition

Contrastive learning demonstrates remarkable generalization performance but lacks theoretical understanding, while contrastive clustering achieves promising performance but exhibits some shortcomings. We first introduce a generalized bias-variance decomposition to study contrastive learning, then present the concept of the conformal field, which unifies instance-level contrastive loss and cluster-level de-redundancy loss (Barlow Twins). Finally, we integrate the conformal field and self-labeling to propose the outstanding contrastive clustering model D3CF. D3CF consists of two novel stages: the pre-training stage simultaneously performs instance-level contrastive learning and multi-view cluster-level redundancy reduction, bringing positive samples together and separating negative samples in the row and column space of the augmented feature matrix; to alleviate the adverse effects caused by false-negative pairs and misclustered assignments in the pre-training stage, the boosting stage enhances contrastive learning from single-positive pairs to multiple-positive pairs by leveraging cross-sample similarities, while utilizing pseudo-labels with high confidence criteria for self-labeling to correct clustering assignments. Extensive experiments on six image benchmark datasets and two text benchmarks demonstrate D3CF’s superior performance and validate the effectiveness of its components. Particularly on CIFAR-10, ImageNet-10, and STL-10, D3CF achieves average accuracies of 89.5%, 97%, and 91%, improving NMI by 5.2%, 4.8%, and 2.1%, and ARI by 7%, 7.3%, and 7.3% over the closest baseline.

12596 Characterizing Glycemic Variability In Novel Subphenotypes Of Individuals At Elevated Risk For Type 2 Diabetes Using Continuous Glucose Monitoring

Abstract Disclosure: A.J. Kretowski: None. P. Konopka: None. A. Janucik: None. A. Citko: None. A. Paszko: None. M. Klak: None. A. Golonko: None. A. Szklaruk: None. L. Szczerbinski: None. Background: The heterogeneity among individuals at elevated risk for Type 2 Diabetes (T2D) recently has led to the identification of six subphenotypes, each with distinct risks for developing T2D, its complications, and mortality. Our study leveraged Continuous Glucose Monitoring (CGM) to offer a detailed characterization of these subphenotypes, aiming to uncover differences in glycemic variability and control. Methods: CGM data from 616 individuals without diabetes, from the Polish Registry of Diabetes (PolReD) study, classified into six clusters, were analyzed. The PolReD used Freestyle Libre 2 (Abbott) as the CGM device. CGM-derived measures of glucose control and glucose variability were calculated using the iglu R package. We utilized Analysis of Variance (ANOVA) and Analysis of Covariance (ANCOVA), adjusting for age and sex, to identify significant differences in CGM parameters across the clusters. Results: Our analysis identified significant differences in several CGM-derived parameters among the clusters, notably in the Glycemic Risk Assessment Diabetes Equation (GRADE), Standard Deviation of weekly glucose values (SDw), Mean Amplitude of Glycemic Excursions (MAGE), Maximum glucose value, Glucose range, and Mean Absolute Glucose (MAG). Clusters 3 and 5, characterized by beta-cell dysfunction and high insulin resistance respectively, and associated with the highest risk of T2D, exhibited the greatest glucose variability, with elevated risk for glycemic excursions. In contrast, Cluster 4, indicative of a metabolically healthy obesity profile with a low risk of T2D, demonstrated the most stable glycemic control and the lowest variability among the clusters. Conclusions: This study shows that subtypes of patients at elevated risk for T2D exhibit distinct patterns in CGM-derived parameters, indicating variations in glucose control and stability that extend beyond traditional fasting and post-challenge glucose assessments. Notably, we found that subtypes facing the highest T2D risk display increased glycemic variability, despite their differing underlying mechanisms of dysglycemia. These findings highlight the utility of CGM as a tool for identifying individuals at an increased risk of T2D, reducing the need for extensive phenotyping typically required for cluster assignment. Presentation: 6/1/2024

Label distribution-driven multi-view representation learning

In multi-view representation learning (MVRL), the challenge of category uncertainty is significant. Existing methods excel at deriving shared representations across multiple views, but often neglect the uncertainty associated with cluster assignments from each view, thereby leading to increased ambiguity in the category determination. Additionally, methods like kernel-based or neural network-based approaches, while revealing nonlinear relationships, lack attention to category uncertainty. To address these limitations, this paper proposes a method leveraging the uncertainty of label distributions to enhance MVRL. Specifically, our approach combines uncertainty reduction based on label distribution with view representation learning to improve clustering accuracy and robustness. It initially computes the within-view representation of the sample and semantic labels. Then, we introduce a novel constraint based on either variance or information entropy to mitigate class uncertainty, thereby improving the discriminative power of the learned representations. Extensive experiments conducted on diverse multi-view datasets demonstrate that our method consistently outperforms existing approaches, producing more accurate and reliable class assignments. The experimental results highlight the effectiveness of our method in enhancing MVRL by reducing category uncertainty and improving overall classification performance. This method is not only very interpretable but also enhances the model’s ability to learn multi-view consistent information.

Deep attributed graph clustering with feature consistency contrastive and topology enhanced network

Deep attributed graph clustering has attracted considerable interest lately due to its capability to uncover meaningful latent knowledge from heterogeneous spaces, thereby improving our comprehension of real-world systems. However, ensuring the consistency of the clustering assignments generated from topological and attribute information remains a key issue, which is one of the reasons for the low performance of clustering. To tackle these issues, a novel deep clustering approach with Feature Consistency Contrastive and Topology Enhanced Network (FCC-TEN) is proposed, which consists of GAT and AE that can mine the topological and attributed information and achieve consistency contrastive learning to improve clustering performance. First, a Fusion Graph Convolutional Auto-encoder module is proposed to fuse the attribute information captured by each layer of the AE and enrich topological information for improving the feature extraction capability of AE. Then, using a Feature Consistency Contrastive module to uncover consistency information of the GAT and AE through contrastive learning at the feature and label level. Finally, clustering results are obtained directly by the clustering assignment obtained at the label level. Comprehensive testing on five improved datasets shows that our method provides advanced clustering performance. Moreover, visual analyses of the clustering results corroborate a gradual refinement of the clustering structure, proving the validity of our approach.

Iterative Multiview Subspace Learning for Unpaired Multiview Clustering.

In real applications, several unpredictable or uncertain factors could result in unpaired multiview data, i.e., the observed samples between views cannot be matched. Since joint clustering among views is more effective than individual clustering in each view, we investigate unpaired multiview clustering (UMC), which is a valuable but insufficiently studied problem. Due to lack of matched samples between views, we could fail to build the connection between views. Therefore, we aim to learn the latent subspace shared by views. However, existing multiview subspace learning methods usually rely on the matched samples between views. To address this issue, we propose an iterative multiview subspace learning strategy [iterative unpaired multiview clustering (IUMC)], aiming to learn a complete and consistent subspace representation among views for UMC. Moreover, based on IUMC, we design two effective UMC methods: 1) Iterative unpaired multiview clustering via covariance matrix alignment (IUMC-CA) that further aligns the covariance matrix of subspace representations and then performs clustering on the subspace and 2) iterative unpaired multiview clustering via one-stage clustering assignments (IUMC-CY) that performs one-stage multiview clustering (MVC) by replacing the subspace representations with clustering assignments. Extensive experiments show the excellent performance of our methods for UMC, compared with the state-of-the-art methods. Also, the clustering performance of observed samples in each view can be considerably improved by those observed samples from the other views. In addition, our methods have good applicability in incomplete MVC.

Mobility trajectories in multiple sclerosis: A comparative study of timed 25-foot walk and a patient-reported outcome measure

Background: Loss of mobility is common in persons with multiple sclerosis (PwMS), but little is known about this impairment from the patient’s perspective. Objective: The aim is to model longitudinal variation in a mobility patient-reported outcome (PRO) and compare trajectories to those observed for Timed 25-Foot Walk (T25FW) in a retrospective cohort. Methods: Latent-class growth analysis was applied to 47,508 measures of Performance Scales© Mobility PRO (PS-Mobility) over ~4 years for 8524 PwMS. For 7347 PwMS, there were 41,988 T25FW measures during this period. Repeated measures correlation and concordance of trajectory assignment were evaluated. Results: At the group level, PS-Mobility and T25FW linearly worsened and repeated-measures correlation was moderate. Eight latent classes with varying shapes that worsened described PS-Mobility variation, compared to six latent classes for T25FW that differed by intercept. The agreement between PS-Mobility and T25FW cluster assignment was modest. A higher proportion of individuals who were Black/African American, older, Medicaid beneficiaries, living in deprived neighborhoods, had longer disease duration, had progressive disease, and ever smokers were assigned to more impaired clusters. Discussion: Cross-sectionally, PS-Mobility and T25FW were highly correlated, but longitudinally correlation was modest to moderate, underscoring the importance of considering both objective and subjective perspectives in evaluating mobility changes in PwMS.

Role of patatin-like phospholipase domain-containing 3 gene for decreasing kidney function in recently diagnosed diabetes mellitus

AimsWe examined the association of the G allele in the single-nucleotide polymorphism (SNP) rs738409 in the third exon of patatin-like phospholipase domain-containing 3 gene (PNPLA3) gene, with chronic kidney disease in diabetes endotypes. MethodsParticipants with recent-onset diabetes (n=707) from the prospective German Diabetes Study (GDS) underwent cluster assignment, detailed phenotyping, genotyping and magnetic resonance spectroscopy to quantify hepatocellular lipid content (HCL). ResultsSevere insulin-resistant diabetes (SIRD) had the lowest glomerular filtration rates (eGFR) and highest HCL compared to severe insulin-deficient, moderate obesity-related, moderate age-related and severe autoimmune diabetes endotypes (all p<0.05). HCL was negatively associated with eGFR (r=-0.287, p<0.01) across all groups. Stratification by G-allele carrier status did not reveal any association between HCL and eGFR among the endotypes. However, the proportion of G-allele carriers increased from 44% for eGFR >60 ml/min to 52% for eGFR <60 ml/min (p<0.05). ConclusionsThe PNPLA3 polymorphism may contribute to declining kidney function independently of liver lipids.

Cross-Modal Clustering With Deep Correlated Information Bottleneck Method.

Cross-modal clustering (CMC) intends to improve the clustering accuracy (ACC) by exploiting the correlations across modalities. Although recent research has made impressive advances, it remains a challenge to sufficiently capture the correlations across modalities due to the high-dimensional nonlinear characteristics of individual modalities and the conflicts in heterogeneous modalities. In addition, the meaningless modality-private information in each modality might become dominant in the process of correlation mining, which also interferes with the clustering performance. To tackle these challenges, we devise a novel deep correlated information bottleneck (DCIB) method, which aims at exploring the correlation information between multiple modalities while eliminating the modality-private information in each modality in an end-to-end manner. Specifically, DCIB treats the CMC task as a two-stage data compression procedure, in which the modality-private information in each modality is eliminated under the guidance of the shared representation of multiple modalities. Meanwhile, the correlations between multiple modalities are preserved from the aspects of feature distributions and clustering assignments simultaneously. Finally, the objective of DCIB is formulated as an objective function based on a mutual information measurement, in which a variational optimization approach is proposed to ensure its convergence. Experimental results on four cross-modal datasets validate the superiority of the DCIB. Code is released at https://github.com/Xiaoqiang-Yan/DCIB.

Assessing clustering methods using Shannon's entropy

Unsupervised clustering techniques are crucial for effectively partitioning datasets into meaningful subgroups. In this paper, we introduce a novel clustering fuzziness metric based on Shannon's entropy, which quantifies the level of uncertainty in clustering assignments. This metric is employed to develop a statistical test for evaluating clustering algorithms and to propose parametric corrections to enhance their performance. We provide empirical evidence showing that our approach significantly improves clustering quality compared to well-known algorithms such as Fuzzy K-Means (FKM) and Fanny. For instance, applying our correction method led to a notable decrease in Jensen-Shannon Divergence (JSD) values, indicating enhanced clustering accuracy. Our methodology is demonstrated through comprehensive experiments on both simulated and real-world datasets. Additionally, our approach proves effective in determining the optimal number of clusters, showing superiority over traditional methods. These results highlight the practical benefits of our metric and correction techniques for improving clustering performance.

Bayesian clustering of spatially distributed compositional data with application to the Great Barrier Reef.

The relative abundance of groups of species is often used in ecological surveys to estimate community composition, a metric that reflects patterns of commonness and rarity of biological assemblages. The focus of this paper is measurements of the abundances of four benthic groups (that live on the seafloor) at several reefs on Australia's Great Barrier Reef (GBR) gathered between 2012 and 2017. In this paper we develop a statistical model to find clusters of locations with similar composition. We examine the changes in clusters during a period impacted by an unprecedented sequence of extreme environmental disturbances. To achieve this, we propose a model that incorporates the geographical location of the data, accounting for the possibility that nearby reefs are similar in composition. This is accomplished with a Dirichlet mixture model and a Potts distribution on the cluster assignments. Non-availability of the normalised Potts distribution makes Bayesian inference a doubly-intractable task. To circumvent this additional inferential challenge, an approximate exchange algorithm is specified. The analysis of the 2012 data, collected before the weather disturbances, reveals four clusters. The four groups highlight the primary habitat patterns in the 2012 GBR, each with distinct ecological characteristics: (1) areas with above-average soft coral abundance, (2) sand-dominated regions commonly found in the central part, (3) southern reefs with a more balanced distribution of species, and (4) habitats dominated by algae and hard corals. Compared to subsequent surveys conducted after disturbances, there is evidence of a decline in the number of clusters and a simplification of reef composition at the regional scale.