Model Comparison Methods Research Articles

Should Zeros Count? Modeling the Galaxy–Globular Cluster Scaling Relation with(out) Zero-inflated Count Models

The scaling relation between the size of a galaxy’s globular cluster (GC) population (N GC) and the galaxy’s stellar mass (M *) is usually described with a continuous, linear model, but in reality it is a count relationship that should be modeled as such. For massive galaxies, a negative binomial (NB) model has been shown to describe the data well, but it is unclear how the scaling relation behaves at low galaxy masses where a substantial portion of galaxies have N GC = 0. In this work, we test the utility of Poisson and NB models for describing the low-mass end of the N GC−M * scaling relation. We introduce the use of zero-inflated versions of these models, which allow for larger zero populations (e.g., galaxies without GCs) than would otherwise be predicted. We evaluate our models with a variety of predictive model comparison methods, including predictive intervals, the leave-one-out cross-validation criterion, and posterior predictive comparisons. We find that the NB model is consistent with our data, but the naive Poisson is not. Moreover, we find that zero inflation of the models is not necessary to describe the population of low-mass galaxies that lack GCs, suggesting that a single formation and evolutionary process acts over all galaxy masses. Under the NB model, there does not appear to be anything unique about the lack of GCs in many low-mass galaxies; they are simply the low-mass extension of the larger N GC−M * scaling relation.

Bayesian regression modeling and inference of energy efficiency data: the effect of collinearity and sensitivity analysis

The majority of research predicted heating demand using linear regression models, but they did not give current building features enough context. Model problems such as Multicollinearity need to be checked and appropriate features must be chosen based on their significance to produce accurate load predictions and inferences. Numerous building energy efficiency features correlate with each other and with heating load in the energy efficiency dataset. The standard Ordinary Least Square regression has a problem when the dataset shows Multicollinearity. Bayesian supervised machine learning is a popular method for parameter estimation and inference when frequentist statistical assumptions fail. The prediction of the heating load as the energy efficiency output with Bayesian inference in multiple regression with a collinearity problem needs careful data analysis. The parameter estimates and hypothesis tests were significantly impacted by the Multicollinearity problem that occurred among the features in the building energy efficiency dataset. This study demonstrated several shrinkage and informative priors on likelihood in the Bayesian framework as alternative solutions or remedies to reduce the collinearity problem in multiple regression analysis. This manuscript tried to model the standard Ordinary Least Square regression and four distinct Bayesian regression models with several prior distributions using the Hamiltonian Monte Carlo algorithm in Bayesian Regression Modeling using Stan and the package used to fit linear models. Several model comparison and assessment methods were used to select the best-fit regression model for the dataset. The Bayesian regression model with weakly informative prior is the best-fitted model compared to the standard Ordinary Least Squares regression and other Bayesian regression models with shrinkage priors for collinear energy efficiency data. The numerical findings of collinearity were checked using variance inflation factor, estimates of regression coefficient and standard errors, and sensitivity of priors and likelihoods. It is suggested that applied research in science, engineering, agriculture, health, and other disciplines needs to check the Multicollinearity effect for regression modeling for better estimation and inference.

Deep Learning for Pneumonia Detection in Chest X-ray Images: A Comprehensive Survey.

This paper addresses the significant problem of identifying the relevant background and contextual literature related to deep learning (DL) as an evolving technology in order to provide a comprehensive analysis of the application of DL to the specific problem of pneumonia detection via chest X-ray (CXR) imaging, which is the most common and cost-effective imaging technique available worldwide for pneumonia diagnosis. This paper in particular addresses the key period associated with COVID-19, 2020-2023, to explain, analyze, and systematically evaluate the limitations of approaches and determine their relative levels of effectiveness. The context in which DL is applied as both an aid to and an automated substitute for existing expert radiography professionals, who often have limited availability, is elaborated in detail. The rationale for the undertaken research is provided, along with a justification of the resources adopted and their relevance. This explanatory text and the subsequent analyses are intended to provide sufficient detail of the problem being addressed, existing solutions, and the limitations of these, ranging in detail from the specific to the more general. Indeed, our analysis and evaluation agree with the generally held view that the use of transformers, specifically, vision transformers (ViTs), is the most promising technique for obtaining further effective results in the area of pneumonia detection using CXR images. However, ViTs require extensive further research to address several limitations, specifically the following: biased CXR datasets, data and code availability, the ease with which a model can be explained, systematic methods of accurate model comparison, the notion of class imbalance in CXR datasets, and the possibility of adversarial attacks, the latter of which remains an area of fundamental research.

Analyzing Motorcycle Accident Frequency Using Generalized Poisson Distributions

Motorcycle accidents in East Java are more common than accidents in other modes of transportation. In addition to the many motorcycle users today, human, environmental, and road factors are considered the highest causes of these accidents. The study's goal is to find the best model of the Generalized Poisson Family Distribution (GPR), namely Lagrangian Poisson Regression (LPR) and to construct a model that will quantify the frequency of motorcycle accidents in East Java. Akaike Information Criterion (AIC) and Schwarz Bayesian Criterion (SBC) criteria are the model comparison methods used in this research. The selection was also made based on the model's exponential coefficient with a 95% CI to further deepen the selection results obtained. In addition, the paired samples test was performed to determine the degree of dissimilarity between the outcomes produced by the developed model and the actual data. The best performance model is applied to identify the characteristics or factors highly involved in motorcycle accidents. The research uses secondary data from related agencies, namely the East Java Regional Police, especially the traffic accident unit, and East Java BPS, for 38 cities and districts in 2021. The numerical optimization method used is the iteratively reweighted least squares (IRLS) algorithm, assisted by R Studio software. The study findings show that LPR is the most efficient and exact approach for modeling the frequency of motorcycle accidents. Meanwhile, the percentage of teenagers (X1), the frequency of motorized vehicles (X3), and the average annual rainfall (X5) have a considerable impact on accident occurrence. This research has an important contribution, especially in the field of transportation modeling and designing appropriate strategies to reduce the frequency of motorcycle accidents.

Refrigerant leak detection in industrial vapor compression refrigeration systems using machine learning

Efficient detection of refrigerant leakage is of utmost importance for industrial refrigeration systems due to its potential to cause substantial impacts on system performance and the environment. Existing research on fault detection and diagnosis in refrigeration systems primarily revolves around solutions based on experimental or laboratory data. However, in the industrial use case, achieving accurate and early detection poses significant challenges. This paper reports on the development of a novel refrigerant leak detection method for industrial vapor compression refrigeration systems. Our method leverages real-world data obtained from operational installations, enabling us to assess its reliability and applicability. The proposed data-driven approach involves predicting the fault-free liquid level in the installation receiver and comparing the actual and predicted levels. In this work, we place emphasis on features and model selection. Dedicated metrics combined with a model comparison method are proposed to evaluate and compare the performance of commonly used regression models with two sets of features to determine the most effective one. Furthermore, we provide insights into the results obtained from the deployment of the proposed method in real-world industrial installations.

Ant backbone phylogeny resolved by modelling compositional heterogeneity among sites in genomic data

Ants are the most ubiquitous and ecologically dominant arthropods on Earth, and understanding their phylogeny is crucial for deciphering their character evolution, species diversification, and biogeography. Although recent genomic data have shown promise in clarifying intrafamilial relationships across the tree of ants, inconsistencies between molecular datasets have also emerged. Here I re-examine the most comprehensive published Sanger-sequencing and genome-scale datasets of ants using model comparison methods that model among-site compositional heterogeneity to understand the sources of conflict in phylogenetic studies. My results under the best-fitting model, selected on the basis of Bayesian cross-validation and posterior predictive model checking, identify contentious nodes in ant phylogeny whose resolution is modelling-dependent. I show that the Bayesian infinite mixture CAT model outperforms empirical finite mixture models (C20, C40 and C60) and that, under the best-fitting CAT-GTR + G4 model, the enigmatic Martialis heureka is sister to all ants except Leptanillinae, rejecting the more popular hypothesis supported under worse-fitting models, that place it as sister to Leptanillinae. These analyses resolve a lasting controversy in ant phylogeny and highlight the significance of model comparison and adequate modelling of among-site compositional heterogeneity in reconstructing the deep phylogeny of insects.

Factor models for Chinese A-shares

We compare the performance of commonly employed asset pricing models on a large, liquid, but mostly segmented Chinese A-shares equity market. When restricting ourselves to factor models developed for the U.S. equity market, the q-factor model performs well. However, it is outperformed by a modified Fama-French six-factor model and by a four-factor asset pricing model tailored to the Chinese A-shares market. A data-driven method results in a seven-factor model, however the ranking of asset pricing models changes when we incorporate transaction costs. Both direct and data-driven model comparison methods now lead to a three-factor model comprising a market, size, and earnings-based value factor.

Estimating and testing skewness in a stochastic volatility model

In this paper we propose a novel approach to estimating and testing skewness in a stochastic volatility (SV) model. Our key idea is to replace a normal return error in the standard SV model with a split normal error. We show that this simple variation in the model brings about two large computational advantages. First, the stochastic volatility process can be simulated fast and efficiently using a one-block Gibbs sampling technique. Second, more importantly, this is the first to provide a marginal likelihood calculation method to formally test the coexistence of stochastic volatility and skewness in return errors within a Bayesian framework. We demonstrate the efficiency and reliability of our posterior sampling and model comparison methods through a simulation study. The simulation results show that neglecting skewness leads to inaccurate estimates on both the volatility process and conditional expected returns. Our empirical applications to daily stock return data provide a strong evidence of negative skewness.

Discussion of “Specifying prior distributions in reliability applications”

Discussion of “Specifying prior distributions in reliability applications”

Fitting the grain orientation distribution of a polycrystalline material conditioned on a Laguerre tessellation

The description of distributions related to grain microstructure helps physicists to understand the processes in materials and their properties. This paper presents a general statistical methodology for the analysis of crystallographic orientations of grains in a 3D Laguerre tessellation dataset which represents the microstructure of a polycrystalline material. We introduce complex stochastic models which may substitute expensive laboratory experiments: conditional on the Laguerre tessellation, we suggest interaction models for the distribution of cubic crystal lattice orientations, where the interaction is between pairs of orientations for neighbouring grains in the tessellation. We discuss parameter estimation and model comparison methods based on maximum pseudolikelihood as well as graphical procedures for model checking using simulations. Our methodology is applied for analysing a dataset representing a nickel–titanium shape memory alloy.

Modified Generalized Exponential Distribution

In this article, a new distribution having three called Modified Generalized Exponential Distribution is proposed. Important statistical properties of the proposed model like survival function, hazard rate function, the probability density function (PDF), the cumulative distribution function (CDF), quantile function, skewness, and kurtosis are discussed here. Least Square Estimation (LSE), Cramer- Von Mises (CVM) and Maximum Likelihood estimators (MLE) methods are used for estimation of parameters using R programming software. A data set is discussed and performed the goodness-of-fit to assess the applications of the proposed distribution. Various methods of model comparison and model validation are also used. The proposed model Modified Generalized Exponential Distribution is more applicable as compared to some existing probability model.

Benchmark of embedding-based methods for accurate and transferable prediction of drug response.

Prediction of therapy response has been a major challenge in cancer precision medicine due to the extensive tumor heterogeneity. Recently, several deep learning methods have been developed to predict drug response by utilizing various omics data. Most of them train models by using the drug-response screening data generated from cell lines and then use these models to predict response in cancer patient data. In this study, we focus on and evaluate deep learning methods using transcriptome data for the long-standing question of personalized drug-response prediction. We developed an embedding-based approach for drug-response prediction and benchmarked similar methods for their performance. For all methods, we used pretreatment transcriptome data to train models and then conducted a comprehensive evaluation and comparison of the models using cross-panels, cross-datasets and target genes. We further validated the methods using three independent datasets assessing multiple compounds for their predictive capability of drug response, survival outcome and cell line status. As a result, the methods building on gene embeddings had an overall competitive performance with reduced overfitting when we applied evaluation parameters for model fitting as well as the correlation with clinical outcomes in the validation data. We further developed an ensemble model to combine the results from the three most competitive methods for an overall prediction. Finally, we developed DrVAEN (https://bioinfo.uth.edu/drvaen), a user-friendly and easy-accessible web-server that hosts all these methods for drug-response prediction and model comparison for broad use in cancer research, method evaluation and drug development.

ПРАВОВЫЕ ВОПРОСЫ ВНЕДРЕНИЯ СОВРЕМЕННЫХ ИНФОРМАЦИОННЫХ СИСТЕМ В ОБРАЗОВАТЕЛЬНЫЙ ПРОЦЕСС (на примере творческих вузов)

2023 has become a significant year in the development of the IT industry in the Russian Federation. Within the Federal project “Digital Public Administration” of the national programme “Digital Economy of the Russian Federation”, activities are being carried out to digitally transform state and municipal services. First of all, the Superservice programme was launched. The Superservice University Admission system has certain advantages. It is one of a whole list of proposed implementation programmes in various areas of the economy and public administration, developed by domestic companies. At the same time, programmes positioning themselves as artificial intelligence programmes are being actively introduced into the environment. Especially it concerns the Chat GPT and the Midjourney programme. Various research methods are used in the study of neural network data, including: data mining methods, statistical analysis, data visualization and machine learning, model comparison methods, comparison of different models based on their performance and other criteria, etc. In addition, a large number of legal issues arise when implementing these systems. This applies to a greater extent to the introduction of neural networks as part of artificial intelligence. These are such issues as copyright, protection and security issues, liability for errors and damage, etc. This article is devoted to the problems of implementation.

Comparing stochastic volatility specifications for large Bayesian VARs

Large Bayesian vector autoregressions with various forms of stochastic volatility have become increasingly popular in empirical macroeconomics. One main difficulty for practitioners is to choose the most suitable stochastic volatility specification for their particular application. We develop Bayesian model comparison methods–based on marginal likelihood estimators that combine conditional Monte Carlo and adaptive importance sampling–to choose among a variety of stochastic volatility specifications. The proposed methods can also be used to select an appropriate shrinkage prior on the VAR coefficients, which is a critical component for avoiding over-fitting in high-dimensional settings. Using US quarterly data of different dimensions, we find that both the Cholesky stochastic volatility and factor stochastic volatility outperform the common stochastic volatility specification. Their superior performance, however, can mostly be attributed to the more flexible priors that accommodate cross-variable shrinkage.

Statistical comparison of DEC and DEC+J is identical to comparison of two ClaSSE submodels, and is therefore valid

AbstractAimStatistical model comparison has become common in historical biogeography, enabled by the R package BioGeoBEARS, which implements several models in a common framework, allowing models to be compared with standard likelihood‐based methods of statistical model comparison. Ree and Sanmartín (Journal of Biogeography, 45, 741–749, 2018) critiqued the comparison of Dispersal–Extinction–Cladogenesis (DEC) and a modification of it, DEC+J, which adds the process of jump dispersal at speciation. DEC+J provides highly significant improvements in model fit on most (although not all) datasets. They claim that the comparison is statistically invalid for a variety of reasons. I analyse the key claims made by the critique.LocationSimulated data.TaxonSimulated data.MethodsLikelihood calculations are checked by comparison between programs and by‐hand calculations, and by summing likelihoods across all possible datasets. Model adequacy of DEC versus DEC+J is checked by a simulation/inference experiment.ResultsMistakes in the critique's example likelihood calculations are demonstrated. DEC+J fits better on datasets because the DEC model is statistically inadequate in the common situation when most species have geographical ranges of single areas; the DEC model requires long residence times of multi‐area ranges, and when these are not observed, a model that does produce such data patterns, such as DEC+J, prevails. More fundamentally, statistical comparison of DEC and DEC+J produces identical log‐likelihood differences to statistical comparison of two submodels of ClaSSE where extinction rates are fixed to 0.Main ConclusionsDEC fails a basic model adequacy check for understandable reasons, while DEC+J does not. As Ree and Sanmartín recommend ClaSSE models as valid for comparison, the comparison of DEC and DEC+J is statistically valid according to their own criteria.

Common EOFs: a tool for multi-model comparison and evaluation

With the increase in the volume of climate model simulations for past, present and future climate, from various institutions across the globe, there is a need for efficient and robust methods for model comparison and/or evaluation. This manuscript discusses common empirical orthogonal function analysis with a step-wise algorithm, which can be used for the above objective. The method looks for simultaneous diagonalisation of several covariance matrices in a step-wise fashion ensuring thus simultaneous monotonic decrease of the eigenvalues in all groups, and allowing therefore for dimension reduction. The method is applied to a number of tropospheric and stratospheric fields from the main four reanalysis products, and also to several historical climate model simulations from CMIP6, the Coupled Model Intercomparison Project (Phase 6). Monthly means as well as winter daily gridded data are considered over the Northern Hemisphere. The method shows consistency between mass fields as well as mid-tropospheric and stratospheric fields of the reanalyses, but also reveals significant differences in the 2 m surface-air temperature in terms of explained variance. CMIP6 models, on the other hand, show differences reflected in the percentage of explained variance of the leading common EOFs with inter-group variation ranging from 5–10% in the troposphere to about 25% in the stratosphere. Higher order statistics within the leading common modes of variability, in addition to further merits of the method are also discussed.

Grouping affective psychoses in early intervention: Justification for specific treatment guidelines

The concept of affective psychosis regroups psychotic disorders with mood syndrome. Previous studies provided evidence to support a dichotomy between affective and non-affective psychoses although questions remain regarding the utility and validity of such a category to develop clinical guidelines.The aim of this study is to explore similarities and differences within affective psychoses to question whether strategies would apply to all the diagnoses falling under this umbrella term.Using Bayesian model comparison methods, we explored the homogeneity of the characteristics of first-episode affective patients (N = 77) treated in a specialized 3-year early intervention in psychosis programme.Our analysis revealed affective psychoses display many similarities regarding socio-demographic variables, the course of positive and manic symptoms over three years, and outcome at discharge. Our results did not support the heterogeneous model. However, despite no significant differences in the course of symptoms with the major depressive disorder group, the schizoaffective disorder group displayed a more severe clinical picture at the beginning of the programme and a poorer functional outcome than the two other groups.Absence of clear boundaries and the several similarities within affective psychoses suggest they can usefully be grouped to define treatment strategies that are easily legible by clinicians.

A New Generalized-X Family of Distributions: Applications, Characterization and a Mixture of Random Effect Models

The researchers in applied statistics are recently highly motivated to introduce new generalizations of distributions due to the limitations of the classical univariate distributions. In this study, we propose a new family called new generalized-X family of distributions. A special sub-model called new generalized-Weibull distribution is studied in detail. Some basic statistical properties are discussed in depth. The performance of the new proposed model is assessed graphically and numerically. It is compared with the five well-known competing models. The proposed model is the best in its performance based on the model adequacy and discrimination techniques. The analysis is done for the real data and the maximum likelihood estimation technique is used for the estimation of the model parameters. Furthermore, a simulation study is conducted to evaluate the performance of the maximum likelihood estimators. Additionally, we discuss a mixture of random effect models which are capable of dealing with the overdispersion and correlation in the data. The models are compared for their best fit of the data with these important features. The graphical and model comparison methods implied a good improvement in the combined model.

Dynamic Causal Modeling of the Prefrontal/Amygdala Network During Processing of Emotional Faces.

Introduction: The importance of the amygdala/medial orbitofrontal cortex (OFC) network during processing of emotional stimuli, emotional faces in particular, is well established. This premise is supported by converging evidence from animal models, human neuroanatomical results, and neuroimaging studies. However, there is missing evidence from human brain connectivity studies that the OFC and no other prefrontal brain areas such as the dorsolateral prefrontal cortex (DLPFC) or ventrolateral prefrontal cortex (VLPFC) are responsible for amygdala regulation in the functional context of emotional face stimuli. Methods: Dynamic causal modeling of ultrahigh-field functional magnetic resonance imaging data acquired at 7 Tesla in 38 healthy subjects and a well-established paradigm for emotional face processing were used to assess the central role of the OFC to provide empirical validation for the assumed network architecture. Results: Using Bayesian model selection, it is demonstrated that indeed the OFC, and not the VLPFC and the DLPFC, downregulates amygdala activation during the emotion discrimination task. In addition, Bayesian model averaging group results were rigorously tested using bootstrapping, further corroborating these findings and providing an estimator for robustness and optimal sample sizes. Discussion: While it is true that VLPFC and DLPFC are relevant for the processing of emotional faces and are connected to the OFC, the OFC appears to be a central hub for the prefrontal/amygdala interaction. Impact statement Using dynamic causal modeling (DCM), abnormal effective connectivity in the orbitofrontal cortex (OFC)/amygdala network has been repeatedly observed in the pathophysiology of psychiatric disorders. However, it has to be considered that these findings are all based on the a priori assumption of the OFC being the central area for prefrontal control regulating amygdala activation. This is particularly important, as DCM results conditionally depend on the underlying model space used for model selection. Using Bayesian model comparison methods, it is shown that the OFC (and not the dorsolateral prefrontal cortex or ventrolateral prefrontal cortex) engages in amygdala downregulation in the context emotional face processing.

Overview of the Jack Rabbit II (JR II) field experiments and summary of the methods used in the dispersion model comparisons

This paper serves as an introduction to the special Atmospheric Environment issue on comparisons of 17 widely-used dense-gas dispersion models using observations from the Jack Rabbit II (JR II) chlorine release experiments, which were led by the U.S. Department of Homeland Security Science & Technology (DHS S&T) Chemical Security Analysis Center (CSAC) and a collaborative team of interagency partners, and conducted at the U.S. Army Dugway Proving Ground in 2015 and 2016. Nine of the 17 models that were compared were from outside the U.S. Other papers in the Special Issue describe specific aspects (e.g., observations of emissions, meteorology, and concentrations) of the JR II field experiments, results of individual dispersion model simulations for the models and comparisons of model predictions with JR II observations and other models. This paper provides a general overview of the field experiments and a brief summary of the model comparison goals and methods.