Mixture Model Research Articles

S5: New insights from deep spectroscopic observations of the tidal tails of the globular clusters NGC 1261 and NGC 1904

As globular clusters (GCs) orbit the Milky Way, their stars are tidally stripped and form tidal tails that follow the orbit of the cluster around the Galaxy. The morphology of these tails is complex and shows correlations with the phase of orbit and the orbital angular velocity, especially for GCs on eccentric orbits. Here we focus on two GCs, NGC 1261 and NGC 1904, that were potentially accreted alongside Gaia-Enceladus and that have shown signatures of having, in addition to tidal tails, structures formed by distributions of extra-tidal stars that are misaligned with the general direction of the clusters' respective orbits. To provide an explanation for the formation of these structures, we made use of spectroscopic measurements from the Southern Stellar Stream Spectroscopic Survey ($S^5$) as well as proper motion measurements from $Gaia$'s third data release (DR3), and applied a Bayesian mixture modelling approach to isolate high-probability member stars. We recovered extra-tidal features surrounding each cluster matching findings from previous work. We then conducted N-body simulations and compared the expected spatial distribution and variation in the dynamical parameters along the orbit with those of our potential member sample. Furthermore, we used Dark Energy Camera (DECam) photometry to inspect the distribution of the member stars in the colour-magnitude diagram (CMD). We find that potential members agree reasonably with the N-body simulations, and that the majority follow a simple stellar population distribution in the CMD, which is characteristic of GCs. We link the extra-tidal features with their orbital properties and find that the presence of the tails agrees well with the theory of stellar stream formation through tidal disruption. In the case of NGC 1904, we clearly detect the tidal debris escaping the inner and outer Lagrange points, which are expected to be prominent when at or close to the apocentre of its orbit. Our analysis allows for further exploration of other GCs in the Milky Way that exhibit similar extra-tidal features.

Mixtures of discrete decomposable graphical models

Mixtures of discrete decomposable graphical models

Effect of thyroid peroxidase antibody titers trajectories during pregnancy and postpartum on postpartum thyroid dysfunction.

To identify risk factors contributing to the development of postpartum hypothyroidism in women newly diagnosed with subclinical hypothyroidism (SCH) during the first trimester of pregnancy (T1). Additionally, this study aimed to explore the impact of thyroid peroxidase antibody (TPOAb) titers trajectories throughout pregnancy and postpartum. Thyroid hormone levels and thyroid autoantibody titers were collected from T1 to the 12th month postpartum. Logistic regression analysis was employed to identify independent risk factors for hypothyroidism at the 12th month postpartum and to develop a prediction model. Model performance was assessed through discrimination, calibration, and clinical applicability, with internal validation using the bootstrap resampling method. Growth Mixture Modeling was applied to delineate the trajectory of TPOAb titers during pregnancy and postpartum, and logistic regression analysis was conducted to investigate the influence of these trajectories on the occurrence of postpartum hypothyroidism. At the 12th month postpartum, hypothyroidism was either newly diagnosed or persisted in 76 of 209 cases (36.36%). Several significant risk factors for postpartum hypothyroidism were identified, including multiparity, positive TPOAb in T1, positive TPOAb and thyroglobulin antibody in T1, serum thyroid-stimulating hormone levels at SCH diagnosis in T1, and the final dose of levothyroxine in the third trimester. A prediction model was constructed and presented with a nomogram. Furthermore, a higher trajectory of serum TPOAb titer during pregnancy and postpartum emerged as a predictive factor for hypothyroidism at the 12th month postpartum. Women with elevated TPOAb titers during pregnancy and postpartum necessitate ongoing and vigilant monitoring of thyroid function, even after childbirth.

Abnormal Sound Detection of Wind Turbine Gearboxes Based on Improved MobileFaceNet and Feature Fusion

To solve problems such as the unstable detection performance of the sound anomaly detection of wind turbine gearboxes when only normal data are used for training, and the poor detection performance caused by the poor classification of samples with high similarity, this paper proposes a self-supervised wind turbine gearbox sound anomaly detection algorithm that fuses time-domain features and Mel spectrograms, improves the MobileFaceNet (MFN) model, and combines the Gaussian Mixture Model (GMM). This method compensates for the abnormal information lost in Mel spectrogram features through feature fusion and introduces a style attention mechanism (SRM) in MFN to enhance the expression of features, improving the accuracy and stability of the abnormal sound detection model. For the wind turbine gearbox sound dataset of a certain wind farm in Guangyuan, the average AUC of the sound data at five measuring point positions of the wind turbine gearbox using the method proposed in this paper, STgram-MFN-SRM, reached 96.16%. Compared with the traditional anomaly detection methods LogMel-MFN, STgram-MFN, STgram-Resnet50, and STgram-MFN-SRM(CE), the average AUC of sound detection at the five measuring point positions increased by 5.19%, 4.73%, 11.06%, and 2.88%, respectively. Therefore, the method proposed in this paper effectively improves the performance of the sound anomaly detection model of wind turbine gearboxes and has important engineering value for the healthy operation and maintenance of wind turbines.

The EX-Lindley distribution with applications to renewable energy sources data

This study introduces a novel extension of the X-Lindley model, called the EX-Lindley model. The EX-Lindley model is a mixture model that combines the characteristics of the gamma model with the weighted Lindley model. The EX-Lindley model is analyzed to determine various statistical properties, including ordinary moments, inverse moments, moment generating function, incomplete moments, conditional moments, mean deviation, Lorenz, mean residual life, mean inactivity time, order statistics, and extropy. The maximum likelihood method is employed to estimate model parameters using complete data. Moreover, the simulation study is used to compare and evaluate the attributes of the estimations for this generalized model. Finally, two examples of datasets on renewable energy sources are utilized to show the importance of the new model compared to some well-known statistical models.

Calling for Equity-focused Quantitative Methodology in Discipline-based Education Research: An Introduction to Latent Class Analysis.

Mixture modeling is a latent variable (i.e., a variable that cannot be measured directly) approach to quantitatively represent unobserved subpopulations within an overall population. It includes a range of cross-sectional (such as latent class [LCA] or latent profile analysis) and longitudinal (such as latent transition analysis) analyses and is often referred to as a "person-centered" approach to quantitative data. This research methods paper describes one type of mixture modeling, LCA, and provides examples of how this method can be applied to discipline-based education research in biology and other science, technology, engineering, and math (STEM) disciplines. This paper briefly introduces LCA, explores the affordances LCA provides for equity-focused STEM education research, highlights some of its limitations, and provides suggestions for researchers interested in exploring LCA as a method of analysis. We encourage discipline-based education researchers to consider how statistical analyses may conflict with their equity-minded research agendas while also introducing LCA as a method of leveraging the affordances of quantitative data to pursue research goals aligned with equity, inclusion, access, and justice agendas.

Deep Single Image Defocus Deblurring via Gaussian Kernel Mixture Learning.

This paper proposes an end-to-end deep learning approach for removing defocus blur from a single defocused image. Defocus blur is a common issue in digital photography that poses a challenge due to its spatially-varying and large blurring effect. The proposed approach addresses this challenge by employing a pixel-wise Gaussian kernel mixture (GKM) model to accurately yet compactly parameterize spatially-varying defocus point spread functions (PSFs), which is motivated by the isotropy in defocus PSFs. We further propose a grouped GKM (GGKM) model that decouples the coefficients in GKM, so as to improve the modeling accuracy with an economic manner. Afterward, a deep neural network called GGKMNet is then developed by unrolling a fixed-point iteration process of GGKM-based image deblurring, which avoids the efficiency issues in existing unrolling DNNs. Using a lightweight scale-recurrent architecture with a coarse-to-fine estimation scheme to predict the coefficients in GGKM, the GGKMNet can efficiently recover an all-in-focus image from a defocused one. Such advantages are demonstrated with extensive experiments on five benchmark datasets, where the GGKMNet outperforms existing defocus deblurring methods in restoration quality, as well as showing advantages in terms of model complexity and computational efficiency.

Changes in Skeletal Muscle Mass in the First 3 Months Following Gastrointestinal Cancer Surgery: A Prospective Study.

Surgical resection is the primary treatment for gastrointestinal (GI) cancers, but postoperative skeletal muscle loss (SML) is common and linked to poor prognosis. This study aims to identify patterns of muscle change, examine its association with quality of life (QoL), and explore predictors of SML in the first 3 months. A prospective cohort study was conducted on patients newly diagnosed with GI cancer and undergoing surgery in China between September 2021 and May 2022. Skeletal muscle mass (SMM) and QoL were assessed at admission, 7 days, 1 month, and 3 months post-surgery. Demographic, clinical data, and biomarkers were collected. Missing data were imputed using multiple imputation. Data were analyzed using growth mixture modelling, bivariate analyses, and logistic regression. A total of 483 patients completed baseline assessment. Of the 242 patients with complete muscle assessments, 92% experienced SML. Three distinct patterns of muscle change were identified: 57% had normal preoperative SMM with mild postoperative SML, 16% had low preoperative SMM with moderate SML, and 27% had normal preoperative mass but severe postoperative SML. Moderate/severe SML was associated with more postoperative complications, poorer health, and higher symptom burden. Independent predictors included advanced age, preoperative sarcopenia, advanced cancer stage, and low prognostic nutrition index (PNI ≤ 45). The results did not change when using imputed values. Although SML is prevalent, patterns of muscle change are heterogeneous among patients. Advanced age, preoperative sarcopenia, advanced cancer stage, and cancer-related inflammation are predictors for moderate/severe SML, highlighting the need for early detection and management.

Towards Effective Clustered Federated Learning: A Peer-to-Peer Framework With Adaptive Neighbor Matching

In federated learning (FL), clients may have diverse objectives, and merging all clients' knowledge into one global model will cause negative transfer to local performance. Thus, clustered FL is proposed to group similar clients into clusters and maintain several global models. In the literature, centralized clustered FL algorithms require the assumption of the number of clusters and hence are not effective enough to explore the latent relationships among clients. In this paper, without assuming the number of clusters, we propose a peer-to-peer (P2P) FL algorithm named PANM. In PANM, clients communicate with peers to adaptively form an effective clustered topology. Specifically, we present two novel metrics for measuring client similarity and a two-stage neighbor matching algorithm based Monte Carlo method and Expectation Maximization under the Gaussian Mixture Model assumption. We have conducted theoretical analyses of PANM on the probability of neighbor estimation and the error gap to the clustered optimum. We have also implemented extensive experiments under both synthetic and real-world clustered heterogeneity. Theoretical analysis and empirical experiments show that the proposed algorithm is superior to the P2P FL counterparts, and it achieves better performance than the centralized cluster FL method. PANM is effective even under extremely low communication budgets.

Investigation of Individual Pulse Emission Behaviors from Pulsar J1741–0840

We have carried out a detailed study of individual pulse emission from the pulsar J1741−0840 (B1738−08) observed using the Parkes and Effelsberg radio telescopes at the L band. The pulsar exhibits four emission components, which are not well resolved by employing multicomponent Gaussian fitting. The radio emission originates at a height of approximately 1000 km, with the viewing geometry characterized by inclination and impact angles roughly estimated at 81° and 3°, respectively. Fluctuation spectral analysis of single pulse behavior reveals two prominent periodicities, around 32 and five rotation periods. The longer periodic modulation feature is linked to nulling behavior across the entire emission window, with an updated nulling fraction of 23% ± 2% derived from pulse energy distribution via Gaussian mixture modeling. In addition to quasiperiodic nulling, the pulsar also exhibits the presence of subpulse drifting in the trailing component, with the shorter periodic feature in the fluctuation spectra related to the phenomenon of subpulse drifting, and the longitudinal separation estimated to be about 5°. Both periodic modulations show significant temporal evolution with time-dependent fluctuation power. The implications of understanding the radio emission mechanisms are discussed.

Functional Concurrent Regression Mixture Models Using Spiked Ewens-Pitman Attraction Priors.

Functional concurrent, or varying-coefficient, regression models are a form of functional data analysis methods in which functional covariates and outcomes are collected concurrently. Two active areas of research for this class of models are identifying influential functional covariates and clustering their relations across observations. In various applications, researchers have applied and developed methods to address these objectives separately. However, no approach currently performs both tasks simultaneously. In this paper, we propose a fully Bayesian functional concurrent regression mixture model that simultaneously performs functional variable selection and clustering for subject-specific trajectories. Our approach introduces a novel spiked Ewens-Pitman attraction prior that identifies and clusters subjects' trajectories marginally for each functional covariate while using similarities in subjects' auxiliary covariate patterns to inform clustering allocation. Using simulated data, we evaluate the clustering, variable selection, and parameter estimation performance of our approach and compare its performance with alternative spiked processes. We then apply our method to functional data collected in a novel, smartphone-based smoking cessation intervention study to investigate individual-level dynamic relations between smoking behaviors and potential risk factors.

Prevalence, Trajectory, and Predictors of Post-stroke Fatigue in Older Adults

ObjectiveTo explore the prevalence, trajectories, and predictors of Post-stroke fatigue (PSF) in older adults following a first ischemic stroke. DesignA longitudinal observational cohort study. SettingTwo hospitals. ParticipantsA total of 381 patients aged ≥ 65 years with their first ischemic stroke were included. The mean (SD) age was 71.1 (4.27) years, with 96 patients (25.2%) being women and 285 (74.8%) being men. InterventionsNot applicable. Main Outcome MeasuresPatients were assessed using the Fatigue Severity Scale (FSS) at admission, 3 months, and 12 months. Growth mixture models (GMM) were used to identify distinct fatigue trajectories, and baseline variables were analyzed to determine their association with these trajectories. ResultsThe prevalence of clinical fatigue was 39.11%, 33.33%, and 22.31% at admission, 3 months, and 12 months, respectively. Five distinct fatigue trajectories were identified: persistently low fatigue (class 1, 49.1%), persistently high fatigue (class 2, 21.5%), initial high but early decreasing fatigue (class 3, 15.0%), initial high but late decreasing fatigue (class 4, 8.7%), and increasing-then-decreasing fatigue (class 5, 5.8%). Multinomial logistic regression analysis revealed that several factors were significantly associated with high and persistent fatigue (class 2), including older age, lower social support, decreased physical activity, higher depression and anxiety scores, cognitive impairment, and greater stroke severity. ConclusionThese findings indicate significant variability in the progression of fatigue among stroke survivors. Further research is necessary to determine the outcomes linked to these fatigue trajectory subgroups and to identify the most effective treatment strategies tailored to each specific subgroup.

Design of an effective multiple objects tracking framework for dynamic video scenes

Nowadays, the applications corresponding to video surveillance systems are getting popular due to their wide range of deployment in various places such as schools, roads, and airports. Despite the continuous evolution and increasing deployment of object-tracking features in video surveillance applications, the loopholes still need to be solved due to the limited functionalities of video-tracking systems. The existing video surveillance systems pose high processing overhead due to the larger size of video files. However, the traditional literature report quite sophisticated schemes which might successfully retain higher object detection accuracy from the video scenes but needs more effectiveness regarding computational complexity under limited computing resources. The study thereby identifies the scope of enhancement in traditional object-tracking functions. Further, it introduces a novel, cost-effective tracking model based on Gaussian mixture model (GMM) and Kalman filter (KF) that can accurately identify numerous mobile objects from a dynamic video scene and ensures computing efficiency. The study's outcome shows that the proposed strategic modelling offers better tracking performance for dynamic objects with cost-effective computation compared to the popular baseline approaches.

A latent variable mixture model for composition-on-composition regression with application to chemical recycling

A latent variable mixture model for composition-on-composition regression with application to chemical recycling

Clustering based strategic 3D deployment and trajectory optimization of UAVs with A-star algorithm for enhanced disaster response

A broad spectrum of communication and information technologies is currently being investigated for their potential applications in disaster management. A high level of situational awareness, combined with a prompt and accurate response, is essential for the preservation of life during catastrophe scenarios. This study presents a novel communication strategy employing Unmanned Aerial Vehicles (UAVs) as aerial base stations for providing connectivity to the affected area. The system takes advantage of the flexibility and quick deployment characteristics of UAVs. The main focus is to determine the optimal UAV deployment along with trajectory planning to ensure connectivity in areas where conventional base stations are inaccessible. The proposed system employs two types of UAVs: cluster UAVs which act as stationary base stations and relay UAVs acting as mobile base stations. A three-step strategy is proposed to find the suitable location of cluster UAVs, optimize their height and power, and find the optimal trajectory of the relay UAVs to maximize the percentage of users served. Gaussian Mixture Model (GMM) clustering is employed to determine the optimal horizontal location of cluster UAVs. An optimization problem is framed for finding out the optimal height and power for cluster UAVs. Heuristic-based A-star algorithm is used to find out the trajectory of the relay UAVs which can efficiently minimize the overall path length while avoiding obstacles. The simulation results confirm the effectiveness of the proposed approach and demonstrate the performance enhancement by comparing it with the benchmark schemes.

Diachronous Cretaceous Closure of the Bangong‐Nujiang‐Shyok Ocean (Westernmost Central Tibet)

AbstractPrecisely reconstructing the diachroneity of continental collision along various Paleo‐, Meso‐, and Neo‐Tethyan branches has proved to be challenging, which limited our understanding of the dynamics of terrane suturing and associated topographic inheritance in the Tibetan Plateau. In this study, we present stratigraphic and sedimentological evidence to support, for the first time, the presence of a late Early Cretaceous remnant deep‐water basin active at least until 110 Ma in the westernmost Bangong‐Nujiang suture in central Tibet. This basin was confined by shallow‐marine deposition in the north, east, and southeast, as indicated by obitolinid biostratigraphy and maximum depositional ages for six lithostratigraphic units, while the corresponding ocean had already closed in the east by Aptian‐Albian time. Detrital zircon U‐Pb age data (n = 965), forward numerical mixture modeling, and in situ εHf(t) values point at syn‐collisional magmatic rocks along strike in the east as the main sediment source, with additional variable input from uplifted Jurassic‐lowermost Cretaceous turbidites. Our paleogeographic reconstruction indicates that the closure of the Bangong‐Nujiang Ocean progressed markedly diachronously from east to west, culminating in the final closure of the Shyok oceanic branch during the Late Cretaceous (probably by 95 and no later than 85 Ma). This westward‐younging ocean closure corresponds to a shorter duration of continental convergence in the west, resulting in less crustal shortening and tectonic uplift than in the east. This may have played a pivotal role in shaping Cenozoic topography, characterized by higher elevations in the east compared to the west along the Bangong‐Nujiang‐suture valley.

Research on controlling measures of snowdrifts around Arctic ground-based buildings through shape optimization

Elevated structures in the Arctic region have proven effective in mitigating surrounding snowdrifts by enhancing the airflow beneath, and serve as a preferred architectural form. However, ground-based buildings still occupy the dominant position due to the constraints posed by construction costs and foundation conditions. One of the effective measures to reduce and prevent snowdrifts around such buildings is to modify their aerodynamic shapes. Therefore, this research endeavors to explore measures for controlling snowdrifts around ground-based buildings through shape optimization. Initially, the predictive accuracy of a refined Mixture model for simulating snowdrifts around ground-based structures was checked against wind tunnel test results. Based on the validated numerical model and representative Arctic meteorological conditions, the snowdrift controlling effects by changing the overall and local building shapes were investigated separately. Overall, the snow reduction and prevention effects can be effectively achieved for ground-based buildings with smoother sidewalls, such as those featuring a circular plane. Conversely, for traditional rectangular ground-based buildings, a larger downwind aspect ratio and a more inclined windward surface can significantly diminish peak snow depth. Additionally, the measures by adding windward side chamfering can effectively manage the locations of snow erosion areas, thereby enabling flexible placement of entrances and exits.

High-temporal-resolution ERT characterization for vegetation effects on soil hydrological response under wet-dry cycles

Characterization of vegetation effect on soil response is essential for comprehending site-specific hydrological processes. Traditional research often relies on sensors or remote sensing data to examine the hydrological properties of vegetation zones, yet these methods are limited by either measurement sparsity or spatial inaccuracy. Therefore, this paper is the first to propose a data-driven approach that incorporates high-temporal-resolution electrical resistivity tomography (ERT) to quantify soil hydrological response. Time-lapse ERT is deployed on a vegetated slope site in Foshan, China, during a discontinuous rainfall induced by Typhoon Haikui. A total of 97 ERT measurements were collected with an average time interval of 2.7hours. The Gaussian Mixture Model (GMM) is applied to quantify the level of response and objectively classify impact zones based on features extracted directly from the ERT data. The resistivity-moisture content correlation is established based on on-site sensor data to characterize infiltration and evapotranspiration across wet-dry conditions. The findings are compared with the Normalized Difference Vegetation Index (NDVI), a common indicator for vegetation quantification, to reveal potential spatial errors in remote sensing data. In addition, this study provides discussions on the potential applications and future directions. This paper showcases significant spatio-temporal advantages over existing studies, providing a more detailed and accurate characterization of superficial soil hydrological response.

VILOCA: sequencing quality-aware viral haplotype reconstruction and mutation calling for short-read and long-read data.

RNA viruses exist as large heterogeneous populations within their host. The structure and diversity of virus populations affects disease progression and treatment outcomes. Next-generation sequencing allows detailed viral population analysis, but inferring diversity from error-prone reads is challenging. Here, we present VILOCA (VIral LOcal haplotype reconstruction and mutation CAlling for short and long read data), a method for mutation calling and reconstruction of local haplotypes from short- and long-read viral sequencing data. Local haplotypes refer to genomic regions that have approximately the length of the input reads. VILOCA recovers local haplotypes by using a Dirichlet process mixture model to cluster reads around their unobserved haplotypes and leveraging quality scores of the sequencing reads. We assessed the performance of VILOCA in terms of mutation calling and haplotype reconstruction accuracy on simulated and experimental Illumina, PacBio and Oxford Nanopore data. On simulated and experimental Illumina data, VILOCA performed better or similar to existing methods. On the simulated long-read data, VILOCA is able to recover on average [Formula: see text] of the ground truth mutations with perfect precision compared to only [Formula: see text] recall and [Formula: see text] precision of the second-best method. In summary, VILOCA provides significantly improved accuracy in mutation and haplotype calling, especially for long-read sequencing data, and therefore facilitates the comprehensive characterization of heterogeneous within-host viral populations.

Unsupervised object-based spectral unmixing for subpixel mapping

Subpixel mapping (SPM) addresses the widespread mixed pixel problem in remote sensing images by predicting the spatial distribution of land cover within mixed pixels. However, conventional pixel-based spectral unmixing, a key pre-processing step for SPM, neglects valuable spatial contextual information and struggles with spectral variability, ultimately undermining SPM accuracy. Additionally, while extensively utilized, supervised spectral unmixing is labor-intensive and user-unfriendly. To address these issues, this paper proposes a fully automatic, unsupervised object-based SPM (UO-SPM) model that exploits object-scale information to reduce spectral unmixing errors and subsequently enhance SPM. Given that mixed pixels are typically located at the edges of objects (i.e., the inner part of objects is characterized by pure pixels), segmentation and morphological erosion are employed to identify pure pixels within objects and mixed pixels at the edges. More accurate endmembers are extracted from the identified pure pixels for the secondary spectral unmixing of the remaining mixed pixels. Experimental results on 10 study sites demonstrate that the proposed unsupervised object (UO)-based analysis is an effective model for enhancing both spectral unmixing and SPM. Specifically, the spectral unmixing results of UO show an average increase of 3.65 % and 1.09 % in correlation coefficient (R) compared to Fuzzy-C means (FCM) and linear spectral mixture model (LSMM)-derived coarse proportions, respectively. Moreover, the UO-derived results of four SPM methods (i.e., Hopfield neural network (HNN), Markov random field (MRF), pixel swapping (PSA) and radial basis function interpolation (RBF)) exhibit an average increase of 5.89 % and 3.04 % in overall accuracy (OA) across the four SPM methods and 10 study sites compared to the FCM and LSMM-based results, respectively. Moreover, the proportions of both mixed and pure pixels are more accurately predicted. The advantage of UO-SPM is more evident when the size of land cover objects is larger, benefiting from more accurate identification of objects.