Mixture Model Research Articles

Signal processing and data analysis for GCaMP filtering and OASIS algorithm

This paper presents a comprehensive exploration of Calcium Imaging, a robust technique in neuroscience for monitoring neuronal activity. Utilizing GCaMP fluorescence indicators, the study focuses on the mouse primary visual cortex, aiming to decipher various cellular processes. The research highlights the significance of calcium ions in cellular processes and introduces the GCaMP indicator and elucidates the data preprocessing technique involving high-pass filtering and Fourier transformation, as well as the employment of the Online Active Set method to Infer Spike (OASIS) for spike inference. Results showcase the efficacy of the approach in filtering low-frequency noise and inferring neuronal spikes. Future prospects include exploring alternative noise reduction techniques and comparing OASIS with Gaussian mixture models

Sustainable Applications of Satellite Video Technology in Transportation Land Planning and Management

The accurate perception and prediction of traffic parameters like vehicles is essential to transportation land planning and management. Video satellites launched in recent years have brought promising opportunities into this field, providing a wide perspective and high frame frequency for extracting moving vehicles. However, detecting moving vehicles remains a challenge due to their small size, which diminishes shape and texture details, often causing them to blend with noise or other objects. To address this issue, we propose an effective method for moving vehicle detection in video satellites by integrating road maps. Experiments conducted on videos sampled from Jilin-1 and Skysat satellites show that our approach achieves F-scores of 0.98 and 0.87, respectively, which are superior to the three traditional methods, Gaussian mixture model (GMM), improved frame difference (IFD), and visual background extractor (ViBe). Our method can be used for accurate parameter estimation in real traffic, which paves the way for the application of video satellites in transportation land planning and management.

Association between gynecological disorders and insomnia and depression trajectories: a longitudinal study of middle-aged women.

Insomnia and depression often receive inadequate attention regarding their association with common menopausal gynecological disorders (GDs), and there is a lack of longitudinal epidemiological evidence. Furthermore, the specific disorders that exhibit the strongest correlation with depression, as well as the potential mediating role of insomnia, remain poorly understood. Using data from the Study of Women's Health Across the Nation (SWAN) spanning 1996 to 2008, this study analyzed a sample of 2217 racially diverse premenopausal women (aged 42 to 53 at baseline). Longitudinal trajectory analysis, employing latent class mixture models (LCMM), was used to identify optimal patterns for insomnia and depression. Logistic regression explored associations between pelvic pain, pelvic prolapse or relaxation, abnormal bleeding, and insomnia/depression trajectories. A causal mediation model investigated whether insomnia mediated the link between gynecological disorders and depression. The analysis included 2217 participants for insomnia trajectories and 1767 for depression trajectories. Insomnia and depression showed similar patterns, with a single high and low trajectory and minimal fluctuations. Logistic regression revealed a significant positive correlation between pelvic pain, abnormal bleeding, number of GDs, and an increased risk of high insomnia and depression trajectories. Insomnia trajectories mediated 23.6%, 14.3%, and 11.9% of the association between pelvic pain, abnormal bleeding, and number of GDs, respectively, with depression trajectories. This study found the significant associations between pelvic pain, abnormal bleeding, and comorbidity with an elevated risk of insomnia and depression during the menopausal transition.

A study of particle motion in the presence of clusters

The motivation for this study came from the task of analyzing the kinetic behavior of single molecules in a living cell based on Single Molecule Localization Microscopy. Given measurements of both the motion of clusters and molecules, the main task consists in detecting if a molecule belongs to a cluster. While the exact size of the clusters is usually unknown, upper bounds are available. In this study, we simulate the cluster movement by a Brownian motion and those of the particles by a Gaussian mixture model with two modes depending on the position of the particle within or outside a cluster. We propose various variational models to detect if a particle lies within a cluster based on the Wasserstein and maximum mean discrepancy distances between measures. We compare the performance of the proposed models for simulated data.

CAT-Posterior Mean Site Frequencies Improves Phylogenetic Modeling Under Maximum Likelihood and Resolves Tardigrada as the Sister of Arthropoda Plus Onychophora.

Tardigrada, the water bears, are microscopic animals with walking appendages that are members of Ecdysozoa, the clade of molting animals that also includes Nematoda (round worms), Nematomorpha (horsehair worms), Priapulida (penis worms), Kinorhyncha (mud dragons), Loricifera (loricated animals), Arthropoda (insects, spiders, centipedes, crustaceans, and their allies), and Onychophora (velvet worms). The phylogenetic relationships within Ecdysozoa are still unclear, with analyses of molecular and morphological data yielding incongruent results. Accounting for across-site compositional heterogeneity using mixture models that partition sites in frequency categories, CATegories (CAT)-based models, has been shown to improve fit in Bayesian analyses. However, CAT-based models such as CAT-Poisson or CAT-GTR (where CAT is combined with a General Time Reversible matrix to account for replacement rate heterogeneity) have proven difficult to implement in maximum likelihood. Here, we use CAT-posterior mean site frequencies (CAT-PMSF), a new method to export dataset-specific mixture models (CAT-Poisson and CAT-GTR) parameterized using Bayesian methods to maximum likelihood software. We developed new maximum likelihood-based model adequacy tests using parametric bootstrap and show that CAT-PMSF describes across-site compositional heterogeneity better than other across-site compositionally heterogeneous models currently implemented in maximum likelihood software. CAT-PMSF suggests that tardigrades are members of Panarthropoda, a lineage also including Arthropoda and Onychophora. Within Panarthropoda, our results favor Tardigrada as sister to Onychophora plus Arthropoda (the Lobopodia hypothesis). Our results illustrate the power of CAT-PMSF to model across-site compositionally heterogeneous datasets in the maximum likelihood framework and clarify the relationships between the Tardigrada and the Ecdysozoa.

PhiSiCal-Checkup: A Bayesian framework to validate amino acid conformations within experimental protein structures

As structural biology and drug discovery depend on high-quality protein structures, assessment tools are essential. We describe a new method for validating amino-acid conformations: "PhiSiCal ([Formula: see text]al) Checkup." Twenty new joint probability distributions in the form of statistical mixture models explain the empirical distributions of dihedral angles [Formula: see text] of canonical amino acids in experimental protein structures. Marginal and conditional probability distributions for subsets of dihedral angles are derived from these joint mixture models. Together, these distributions are employed to measure rapidly the information-theoretic "favorability" of any proposed experimental protein structure. The inferred statistical models and measures overcome several shortcomings and afford improvements over the current state of the art in amino-acid conformation verification. Experimental comparisons are made against current protein conformation verification software. In a number of examples, we pick up outliers that are invisible to current methods. We also calculate, as part of verification, the sensitivity of favorability to small changes in a proposed structure accounting for the precision of coordinates. In some cases a near neighbor of a proposed amino-acid conformation may be either less or more favorable. This raises the question, is the current reliance on fixed "thresholds" for validation a good thing? PhiSiCal-Checkup is freely available for online and offline (open-source) use from https://lcb.infotech.monash.edu.au/phisical/checkup.

Unveiling Risk Profiles: A Latent Profile Analysis of 21st-Century Skills, Resistance to Change, and Cognitive Flexibility.

Limited research utilized a person-centered approach in examining 21st-century skills. This study used latent profile analysis to explore the relationships between resistance to change, cognitive flexibility, and 21st-century skills, including creativity, cooperativity, critical thinking, communication, and problem-solving. A convenience sampling approach was used to recruit 502 individuals (342 females; meanage =21.12±2.34 years) via an online survey. The results showed that creativity, cooperativity, critical thinking, communication, problem-solving, and resistance to change were determined as profile indicators. The cognitive flexibility of the subclasses was also predicted. Furthermore, the results revealed four distinct profiles: (i) average level across all skills (5.8%); (ii) low risk for most skills, moderate for cooperativity (18.1%); (iii) high risk for all skills (55.2%); and (iv) moderate risk for most skills, moderate for cooperativity (20.9%). The study highlights the importance of cognitive flexibility in identifying risk profiles. Using mixture modeling provides a fresh perspective for researchers, potentially aiding in targeted interventions for at-risk university students to enhance their 21st-century skills.

Statistical Inference With Anchored Bayesian Mixture of Regressions Models: An Illustrative Study of Allometric Data

Statistical Inference With Anchored Bayesian Mixture of Regressions Models: An Illustrative Study of Allometric Data

Deep learning model to identify and validate hypotension endotypes in surgical and critically ill patients.

Hypotension is associated with organ injury and death in surgical and critically ill patients. In clinical practice, treating hypotension remains challenging because it can be caused by various underlying haemodynamic alterations. We aimed to identify and independently validate endotypes of hypotension in big datasets of surgical and critically ill patients using unsupervised deep learning. We developed an unsupervised deep learning algorithm, specifically a deep learning autoencoder model combined with a Gaussian mixture model, to identify endotypes of hypotension based on stroke volume index, heart rate, systemic vascular resistance index, and stroke volume variation observed during episodes of hypotension. The algorithm was developed with data from 871 surgical patients who had 6962 hypotensive events and validated in two independent datasets, one including 1000 surgical patients who had 7904 hypotensive events and another including 1000 critically ill patients who had 53 821 hypotensive events. We defined hypotension as a mean arterial pressure <65 mm Hg for at least 1 min. In the development dataset, we identified four hypotension endotypes. Based on their physiological and clinical characteristics, we labelled them as: vasodilation, hypovolaemia, myocardial depression, and bradycardia. The same four hypotension endotypes were identified in the two independent validation datasets of surgical and critically ill patients. Unsupervised deep learning identified four endotypes of hypotension in surgical and critically ill patients: vasodilation, hypovolaemia, myocardial depression, and bradycardia. The algorithm provides the probability of each endotype for each hypotensive data point. Identifying hypotensive endotypes could guide clinicians to causal treatments for hypotension.

Gaussian Mixture Models with Concave Penalized Fusion

Gaussian Mixture Models with Concave Penalized Fusion

Enhancing Automatic Speech Recognition Accuracy Using a Gaussian Mixture Model (GMM)

Enhancing Automatic Speech Recognition Accuracy Using a Gaussian Mixture Model (GMM)

Apache Spark based distributed clustering for big data analytic with application to 3D road network

The vast amount of data stored nowadays has turned big data analytics into a very promising research field. Clustering is an essential step in data analysis, widely used for classification, collecting statistics, and acquiring insights in specific domains of knowledge. However, the most of existing algorithms based on Lloyd-Forgy’s method, have an enormously huge average-case complexity while clustering data sets with a large number of features, which may be superpolynomial time (NP-hard) and are severely constrained in terms of speed, productivity, and adaptability. Aiming to improve Lloyd-Forgy’s clustering performance, K-means++ algorithms, a variety of algorithm-level optimizations which is not been well-studied, is discussed along with very promising gaussian mixture model (GMM) and soft clustering based Fuzzy C-means (FCM). Further, for fast and distributed data processing and to leverage the benefits of big data platforms, such as Apache Spark, Spark-based clustering methods are applied on three-dimensional (3D) road network data set which is collected from UCI repository. However, Spark-based clustering research is still in infancy. The distributed computation tests are conducted by allocating two core processors and one databricks unit (DBU) with 15 GB memory and measuring execution times, as well as root mean square error (RMSE), mean absolute error (MAE), clustering accuracy, and silhouette values. The results are promising and provide new research directions in the field of spark-based clustering on big data.

Social functioning longitudinal trajectory and its predictors in young and middle-aged postoperative pituitary tumor patients: A growth mixture model.

Pituitary adenoma and its treatments can significantly impact the social function of young and middle-aged patients. Monitoring social functioning trajectories over time provides a more holistic understanding of patients' quality of life. We aimed to study the social functional trajectories and their predictors to achieve individualized and comprehensive management. A longitudinal prospective survey was conducted from January 2021 to December 2022 in Nanjing, China. The convenience sampling method was adopted to recruit the patients with pituitary adenoma who underwent transsphenoidal surgery at the neurosurgery department in a tertiary-grade A hospital. The general information questionnaire, social disability screening scale and postoperative pituitary adenoma symptom distress scale were applied in different stages. Growth mixture modeling and the Logistic regression were utilized to identify the trajectory subgroups and their predictors of the social function. Of 225 patients, 33 (14.67%) experienced "progressively recovered function", while 192 (85.33%) exhibited "persistently impaired social function". Factors associated with persistently impaired social function included fertility status (OR=0.167, p=0.001), visual loss (OR=0.154, p=0.012), and electrolyte disturbances at 6th month (OR=0.047, p=0.003). Conversely, patients with high monthly income (OR=0.116, p=0.000), good family support (OR=3.890, p=0.006), and preoperative headache (OR=4.129, p=0.006) were more likely to show progressively recovered function after surgery. Patients with pituitary adenoma still experience social dysfunction six months after surgery. Fertility status, family support, economic status, and symptom distress were significant predictors. This study emphasizes the crucial role of symptom distress and family support, providing a basis for follow-up and intervention strategies.

Source Identification of Organic Carbon in Mountainous Reservoirs Sediments Using Stable Isotopes and n‐Alkanes

AbstractThis study focuses on the sources of organic carbon (OC) in surface sediments of reservoirs in the mountainous regions, as well as their dynamics and contribution mechanisms in the carbon cycle. Spatial variations in organic carbon, n‐alkanes, δ13C, and δ15N were analyzed, and a Bayesian isotope mixing model was applied to quantify the relative contributions of different OC sources. The results indicate that the concentration range of organic carbon in sediments is 0.88%–3.72%. The average concentration of long‐chain n‐alkanes is 3.69 μg/g, accounting for 71.4%, indicating that the main source of organic carbon is allochthonous organic carbon. In addition, the Bayesian mixture model results of carbon and nitrogen isotopes also indicate that allochthonous organic carbon is the main contributor. Specifically, sewage (33.1%), C₃ plants (27.1%), and soil organic carbon (19.9%) were identified as the dominant sources. This research highlights the influence of human activities, such as urbanization and agriculture, on OC dynamics and underscores the role of reservoirs in regulating OC transport. The findings provide critical insights into the mechanisms of OC sequestration in agricultural watersheds and offer valuable guidance for water resource management and ecological protection strategies in mountainous environments.

Human-interpretable clustering of short text using large language models.

Clustering short text is a difficult problem, owing to the low word co-occurrence between short text documents. This work shows that large language models (LLMs) can overcome the limitations of traditional clustering approaches by generating embeddings that capture the semantic nuances of short text. In this study, clusters are found in the embedding space using Gaussian mixture modelling. The resulting clusters are found to be more distinctive and more human-interpretable than clusters produced using the popular methods of doc2vec and latent Dirichlet allocation. The success of the clustering approach is quantified using human reviewers and through the use of a generative LLM. The generative LLM shows good agreement with the human reviewers and is suggested as a means to bridge the 'validation gap' which often exists between cluster production and cluster interpretation. The comparison between LLM coding and human coding reveals intrinsic biases in each, challenging the conventional reliance on human coding as the definitive standard for cluster validation.

Moving target detection based on improved Gaussian mixture model in dynamic and complex environments

AbstractRecently, background modeling has garnered significant attention for motion target detection in vision and image applications. However, most methods do not achieve satisfactory results because of the influence of background dynamics and other factors. The Gaussian mixture model (GMM) background modeling method is a popular and powerful motion background modeling technology owing to its ability to balance robustness and real‐time constraints in various practical environments. However, when the background is complex and the target moves slowly, the traditional GMM cannot accurately detect the target and is prone to misjudging the moving background as a moving target. To address the interference from complex backgrounds, this study proposes a target detection method that combines an adaptive GMM with an improved three‐frame difference method, along with an algorithm that combines grayscale statistics with an improved Phong illumination model for illumination compensation and shadow removal. The experimental results demonstrate that the improved method has better robustness, improves target detection accuracy, and reduces noise and background interference.

Augmentation of predictive turbulence modeling applied to low pressure turbines using adaptive field inversion

Based on the framework of field inversion and machine learning, an adaptive modification for Reynolds-Averaged Navier–Stokes-based turbulence models is proposed for the simulation of low-pressure turbine cascades involving flow separation. This method adjusts the results of turbulence by modifying the source terms correspondingly at different spatial locations. First, the specific regions for modification are obtained by Gaussian mixture models adaptively according to the baseline results and the spatial distribution of the correction term is inferred by an ensemble-based inversion method with effective utilization of high-fidelity data. Then a corrective model form of the flow quantities calculated with the baseline model is established by the Gradient Boosting Decision Tree model and used for the simulation of T106 cascade cases. The results demonstrate that with the adaptive modified turbulence model, reduced deficiency on predicting the load distribution can be obtained. The modified model can also predict a more accurate separation onset by damping eddy viscosity in separated region for case out of the training set. With correction added to the turbulence model solely in a specific region, the computational cost can be reduced compared with full-field inversion, and the method can be possibly applied in simulating the three-dimensional flow considering rotation effects.

Exponential utility maximization in small/large financial markets

Obtaining a utility-maximizing optimal portfolio in a closed form is a challenging issue when the return vector follows a more general distribution than the normal one. In this paper, for markets based on finitely many assets, a closed-form expression is given for optimal portfolios that maximize an exponential utility function when the return vector follows normal mean-variance mixture models. Especially, the used approach expresses the closed-form solution in terms of the Laplace transformation of the mixing distribution of the normal mean-variance mixture model and no distributional assumptions on the mixing distribution are made. Also considered are large financial markets based on normal mean-variance mixture models, and it is shown that the optimal exponential utilities in small markets converge to the optimal exponential utility in the large financial market. This shows, in particular, that to reach the best utility level investors need to diversify their investments to include infinitely many assets into their portfolio, and with portfolios based on only finitely many assets they will never be able to reach the optimum level of utility.

Bayesian Dependent Mixture Models: A Predictive Comparison and Survey

Bayesian Dependent Mixture Models: A Predictive Comparison and Survey

Choroid plexus morphology in schizophrenia and early-stage psychosis: A cross-sectional study.

The choroid plexus is an important structure within the ventricular system. Schizophrenia has been associated with morphological changes to the choroid plexus but the presence and extent of alterations at different illness stages is unclear. We examined choroid plexus volumes in participants at clinical high-risk for psychosis (N=110), participants with first-episode psychosis (N=37), participants with schizophrenia (N=28), clinical (N=38) and non-clinical controls (N=75). Automated segmentation (Gaussian mixture model) was used to estimate choroid plexus volumes from T1 magnetic resonance (MR) images. We then conducted a linear model and Bayes factor analysis to investigate group differences. In addition, the relationship between choroid plexus volumes and clinical characteristics was assessed. Schizophrenia patients were characterized by increased choroid plexus and ventricular volume while first-episode psychosis and clinical high-risk for psychosis participants showed no differences in choroid plexus volumes. However, choroid plexus volumes in schizophrenia patients did not significantly differ from controls when controlling for ventricular volume. Finally, choroid plexus volumes were not associated with clinical characteristics in the clinical high-risk group. Our findings suggest that morphological alterations are not specific to the choroid plexus in schizophrenia and early-stage psychosis. Previously reported choroid plexus abnormalities in schizophrenia patients could be explained by changes in ventricular volume.