Model Data Research Articles

Enhancing Federated Learning Evaluation: Exploring Instance-Level Insights with SQUARES in Image Classification Models

Federated Learning (FL) presents a novel approach within the domain of Machine Learning (ML)—enabling the training of ML models in a distributed manner. This paradigmatic shift ensures the preservation of user privacy by conducting the training process at the edge, where data remains localized on individual devices. This stands in contrast to the conventional centralized paradigm—where users transmit data to a central server for processing. The FL framework operates within a more heterogeneous environment characterized by diverse data distributions across clients, a departure from the engineer-analyzed datasets typical of centralized paradigms. Whereas, evaluation of FL models typically relies on statistical techniques such as accuracy, recall, precision, log loss, and the confusion matrix, alongside visualization methods. While these techniques provide an overview of the model's performance and data utilization—they may lack granularity when comparing models with disparate characteristics. However, the SQUARES technique, offers a more nuanced evaluation of model performance at the instance level. This approach facilitates the examination of data biases, outlier detection, and model behavior during training on individual samples. Therefore, this study presents the development and evaluation of an FL image classification model across various scenarios—utilizing the SQUARES prototype. In addressing these more intricate scenarios, we aim to augment traditional visualizations and metrics, thereby uncovering insights and nuances that may elude standard evaluation methods prevalent in ML benchmarks.

A flexible time-varying coefficient rate model for panel count data.

Panel count regression is often required in recurrent event studies, where the interest is to model the event rate. Existing rate models are unable to handle time-varying covariate effects due to theoretical and computational difficulties. Mean models provide a viable alternative but are subject to the constraints of the monotonicity assumption, which tends to be violated when covariates fluctuate over time. In this paper, we present a new semiparametric rate model for panel count data along with related theoretical results. For model fitting, we present an efficient EM algorithm with three different methods for variance estimation. The algorithm allows us to sidestep the challenges of numerical integration and difficulties with the iterative convex minorant algorithm. We showed that the estimators are consistent and asymptotically normally distributed. Simulation studies confirmed an excellent finite sample performance. To illustrate, we analyzed data from a real clinical study of behavioral risk factors for sexually transmitted infections.

Modeling rehabilitation dataset to implement effective AI assistive systems

The demand for automated systems monitoring and supporting patients in their home-based recovery programs is substantial. While emerging technologies have been proposed as potential solutions to enhance at-home patient care, limited systems are in place due to their challenges in offering real-time monitoring and corrective feedback. Most proposed methodologies provide an overall measure or score for the executed movement. The proposed work involves the adaptation of an existing published dataset to cater to the needs of a system capable of remotely assisting patients, effectively acting as a virtual physical therapist able to provide corrective feedback. A dataset containing a set of three simple exercises for shoulder rehabilitation was processed. Each movement was meticulously annotated for temporal and categorical motion domains to monitor exercise execution in terms of the range of motion completeness and to evaluate compensatory movement patterns. This work carries substantial significance by offering a standardized and easily accessible model for human movement data, thus fostering the advancement of digital assistive systems designed to support home-based rehabilitation programs.

FACTORS AFFECTING FINANCIAL DISTRESS IN THE CONSUMER INDUSTRY SECTOR DURING THE COVID-19 PANDEMIC

The COVID-19 pandemic has led to changes in financial performance and a decline in global economic conditions. This study aims to obtain empirical evidence about the influence of Corporate Social Responsibility (CSR), firm size, and firm age on corporate financial distress in the cyclical and non-cyclical consumer sectors during the COVID-19 pandemic. The research design used in this study is descriptive research with quantitative methods. The research sample was selected using purposive sampling, which contains 80 companies as sample. Hypothesis testing is done using a regression analysis model of panel data. The data processing for this study using some application such as Microsoft Excel and EViews 10. The results revealed that Corporate Social Responsibility has no influence on financial distress, the size of a company has a significant positive influence over financial distresses, and the age of the company has significant negative influence upon financial distress. This means that both the size and age of the company are interrelated indicators that should be considered because they can affect the risk of financial distress.

A hierarchical signal detection model with unequal variance for binary responses.

Gaussian signal detection models with equal variance are commonly used in simple yes-no detection and discrimination tasks whereas more flexible models with unequal variance require additional information. Here, a hierarchical Bayesian model with equal variance is extended to an unequal-variance model by exploiting variability of hit and false-alarm rates in a random sample of participants. This hierarchical model is investigated analytically, in simulations and in applications to existing data sets. The results suggest that signal variance and other parameters can be accurately estimated if plausible assumptions are met. It is concluded that the model provides a promising alternative to the ubiquitous equal-variance model for binary data.

Implications of Lairage and Environmental Enrichment on Behavioral Responses and Skin Lesions in Finishing Pigs in a Slaughterhouse.

Different resources, such as environmental enrichment, are being evaluated in order to minimize animal stress and promote better conditions during the life cycle of animals, as consumers are increasingly concerned about animal welfare issues. Lairage represents an important stage in the swine production chain and is directly related to animal welfare. The aim of this study was to evaluate the effect of lairage time in the slaughterhouse and environmental enrichment on the level of skin lesions and behavioral responses in pigs. A total of 648 finishing pigs of both sexes were assessed before and after lairage at the slaughterhouse with a five-point scale (0 = none, to 4 = ≥16 superficial lesions or >10 deep lesions). After lairage (after slaughter), lesions were also classified according to their source (mounting, fighting, and handling). Pigs were distributed into two treatments groups during lairage: with environmental enrichment (EE) on the pen, with hanging metal chains, and with no enrichment (NE). Behavior was monitored during the first four hours of lairage. Proportional odds, mixed linear model for longitudinal data, and non-parametric Wilcoxon signed rank tests were used to analyze the relation between treatments, skin lesions, and behavior. The simple metal chains did not affect skin lesion score or pigs' behavior (p > 0.05), whereas lairage duration influenced standing (SA), sitting (S), lying (L), idleness (I), and drinking water (D) (p < 0.001). The main source of skin lesions was handling, which did not differ between treatments (EE and NE) (p > 0.05). It was observed that the severity of the lesions (highest scores of 3, 4, and 5) increased in the different anatomical regions of the pigs when compared before and after slaughter, with the exception of the frontal area, which was the same (p = 0.7547). Lairage time has a proportional relation with skin lesions, and hanging chains at the slaughterhouse pens was not enough to reduce the number of lesions and to change pig behavior.

APLIKASI MODEL TREND UNTUK MEMPREDIKSI PRODUKSI KELAPA SAWIT PROVINSI KALIMANTAN SELATAN

The plantation sector has a large and extensive reach, one type of plantation crop that is widely cultivated in Indonesia is oil palm. Business in the oil palm plantation sector is also one of the leading potential areas in South Kalimantan Province. One of the provinces producing the most palm oil in Indonesia is South Kalimantan Province. This research is a literature study regarding predictions of palm oil production. The aim of this research is to determine the appropriate prediction model for data on the amount of palm oil (CPO) production in the South Kalimantan Province Region and to explain the results of palm oil predictions in the South Kalimantan Province Region using a trend model. The application of the trend model used in this research takes into account the error value obtained with the smallest MAPE (Mean Absolute Percentage Error). The nature of the data regarding the amount of palm oil production in the South Kalimantan Province Region from 2001 to 2021 tends to continue to increase so that the data can be predicted using trend methods, namely linear, quadratic and exponential. For the prediction results of palm oil production using the best trend model with the smallest MAPE value, namely 16% with the linear trend model: ... The predicted amounts of palm oil production from 2022 to 2027 which tend to continue to increase are: 1.336.359, 1.397.196, 1.458.032, 1.518.869, 1.579.705 and 1.640.542 tons.

Statistical Inference Using Partially Accelerated Life Test Model for the Weibull Population Mean Unified Hybrid Censored Data

In this paper, estimates of the parameters and the acceleration factor of a new distribution known as the Weibull Population Mean Distribution (WPMD) are explored. We propose the use of the model suggested by Abd-Elrahman [A better alternative to the generalized Bilal distribution: A new model and applications, Int. J. Reliab. Qual. Saf. Eng. 30(6) (2023) 2350027, doi: 10.1142/S0218539323500274] for estimating these parameters. To accomplish this, we employ the maximum likelihood estimation method via a Constant-Stress Partially Accelerated Life Test (CSPALT) model. In this study, we have utilized Unified Hybrid (UH) censored data obtained from the WPMD. Confidence intervals for the model parameters and the acceleration factor are constructed based on the approximate Fisher information matrix. The average and Mean Squared Error (MSE) of estimates are computed to evaluate the performance of the proposed method under the CSPALT model. Additionally, a real data set is provided for illustrative purposes, and a numerical analysis of the UH Censoring Scheme (UHCS) is carried out. Finally, concluding remarks are presented.

The cycle of violence: effects of violence experience, behavior, and attitudes on adolescents’ peer rejection networks

Previous research on adolescent peer networks has mainly focused on friendship networks and their association with violence, but very limited research is available on peer rejection networks. This lack of knowledge hinders the effectiveness of preventing peer rejection and its negative effects. Based on the theory of the cycle of violence, the present study examines the extent to which parental physical abuse experiences, aggressive behavior toward peers and acceptance of violence are related to peer rejection networks at school. Social network analysis with a stochastic actor-oriented model of longitudinal data collected from high school students (Wave 1, n = 620; Wave 2, n = 590) confirms that adolescents who frequently use aggression toward their peers are more likely to be rejected, especially if those adolescents have experienced abuse. Similarly, peers are more likely to reject adolescents with high levels of accepting violence. The results also show that aggression toward peers generally tends to decrease over time but not for adolescents who reject a larger number of students. For students who reject many peers, aggression frequency increases. From the perspective of resilience theory, peer rejection, when combined with aggression toward peers and acceptance of violence, particularly in female adolescents, creates a significant risk factor for socio-emotional development. Therefore, tackling violence attitudes, experiences and behavior in the school environment and at home is crucial in overcoming the cycle of violence.

On-the-fly training architecture for a time-series neural network on ADITYA/ADITYA-U data

Plasma disruption prediction using the data drive technique is popular for fusion experiments. Auto-encoder LSTM neural network is implemented on ADITYA/ADITYA-U data to predict the plasma disruption. There is a generic way of processing time series from a dataset that facilitates efficient generalization on new time-series data using a transfer learning technique. This technique enables task-specific parameter improvement with residual correction on univariate time series. This provides an on-the-fly architecture to expand the training model for new data without retraining the whole model. It also provides layer updates based on linearization learning for new input data This architecture enables us to update real-time saved models with new features of data that provide performance improvement during prediction. The demo is tested on ADITYA/ADITYA-U data which save time and run time updates to improve the accuracy of data. The improved results are derived by following the trend while predicting the result after adding new features via this architecture to the existing saved model. This technique allows automated on-the-fly training architecture at time-to-time basis or explicit need of retraining with new data.

A Compression and Encryption Based Heart Disease Diagnosis with Deep Learning through ECG Signals

Electrocardiogram (ECG) monitoring models are commonly employed for diagnosing heart diseases. Since ECG signals are normally acquired for a longer time duration with high resolution, there is a need to compress the ECG signals for transmission and storage. So, a novel compression technique is essential in transmitting the signals to the telemedicine center to monitor and analyses the data. In addition, the protection of ECG signals poses a challenging issue, which encryption techniques can resolve. The existing Encryption-Then-Compression (ETC) models for multimedia data fail to properly maintain the trade- off between compression performance and signal quality. In this view, this study presents a new ETC with a diagnosis model for ECG data, called the ETC-ECG model. The proposed model involves four major processes, namely, pre-processing, encryption, compression, and classification. Once the ECG data of the patient are gathered, Discrete Wavelet Transform (DWT) with a Thresholding mechanism is used for noise removal. In addition, the chaotic map-based encryption technique is applied to encrypt the data. Moreover, the Burrows-Wheeler Transform (BWT) approach is employed for the compression of the encrypted data. Finally, a Deep Neural Network (DNN) is applied to the decrypted data to diagnose heart disease. The detailed experimental analysis takes place to ensure the effective performance of the presented model to assure data security, compression, and classification performance for ECG data

Even low levels of cannibalism can bias population estimates for Pacific hake

Abstract By incorporating trophic interactions and temperature-dependent bioenergetics, multi-species models such as CEATTLE (climate-enhanced age-based model with temperature-specific trophic linkages and energetics) are a step towards ecosystem-based stock assessment and management of high-value commercial species such as Pacific hake (Merluccius productus). Hake are generalist predators and previous studies in the California Current Ecosystem have determined that their diet consists of ∼30% cannibalism. We used CEATTLE to include cannibalism in a model of hake population dynamics and re-examined hake diet data to determine the proportion by age that can attributed to cannibalism. The proportion was highly variable, ranging between 0 and 80% of stomach contents by weight. When included in the CEATTLE model, the estimated spawning biomass, total biomass, and recruitment increased by 15, 23, and 58%, on average, relative to the single-species model, due to the estimation of time- and age-varying predation mortality, primarily for age-1 hake. The effects of cannibalism varied over time, with further increases in total biomass and recruitment resulting from the age structure of the population following large cohorts in 1980 and 1984. Results from the cannibalism model could be used to inform the estimation of time- and age-varying mortality in the single-species assessment and as a pathway for including ecosystem information in management through environmental and trophic drivers of variability in mortality.

Deep bed filtration and formation damage by particles with distributed properties

Current models for deep bed filtration describe particles with uniform properties. Yet, the sizes, densities, and mineral composition of particles vary significantly in the same injection well. The aim of this work is to provide an effective mathematical model for water injection of particles with distributed properties and formation damage prediction. We average the set of traditional population balance equations for single-property particles and obtain one upscaled equation. The upscaled equation for particle retention rate contains a non-linear function of suspended concentration, which we call the 'suspension function'. We derive analytical solutions for the upscaled equation for linear (coreflood) and radial (well injectivity) flows. Then we treat lab coreflood data to determine the model suspension function and provide a model for well injectivity prediction. The retention profile for the flow of uniform particles has an exponential form. Frequently reported in the literature, hyper-exponential forms have been hypothetically explained by multiple particle properties. The inverse solution allows revealing the individual filtration coefficients for binary mixtures from total breakthrough concentrations during coreflood. Treatment of the data from lab experiments reveals individual filtration coefficients that belong to common intervals. For the first time, deep bed filtration of particles with distributed properties is upscaled and presented using a single equation that reflects the particle property distribution. This equation provides an effective mathematical model for tuning lab coreflood data, determines the model function, and uses it for injectivity decline prediction.

Modeling brain sex in the limbic system as phenotype for female-prevalent mental disorders

BackgroundSex differences exist in the prevalence and clinical manifestation of several mental disorders, suggesting that sex-specific brain phenotypes may play key roles. Previous research used machine learning models to classify sex from imaging data of the whole brain and studied the association of class probabilities with mental health, potentially overlooking regional specific characteristics.MethodsWe here investigated if a regionally constrained model of brain volumetric imaging data may provide estimates that are more sensitive to mental health than whole brain-based estimates. Given its known role in emotional processing and mood disorders, we focused on the limbic system. Using two different cohorts of healthy subjects, the Human Connectome Project and the Queensland Twin IMaging, we investigated sex differences and heritability of brain volumes of limbic structures compared to non-limbic structures, and subsequently applied regionally constrained machine learning models trained solely on limbic or non-limbic features. To investigate the biological underpinnings of such models, we assessed the heritability of the obtained sex class probability estimates, and we investigated the association with major depression diagnosis in an independent clinical sample. All analyses were performed both with and without controlling for estimated total intracranial volume (eTIV).ResultsLimbic structures show greater sex differences and are more heritable compared to non-limbic structures in both analyses, with and without eTIV control. Consequently, machine learning models performed well at classifying sex based solely on limbic structures and achieved performance as high as those on non-limbic or whole brain data, despite the much smaller number of features in the limbic system. The resulting class probabilities were heritable, suggesting potentially meaningful underlying biological information. Applied to an independent population with major depressive disorder, we found that depression is associated with male–female class probabilities, with largest effects obtained using the limbic model. This association was significant for models not controlling for eTIV whereas in those controlling for eTIV the associations did not pass significance correction.ConclusionsOverall, our results highlight the potential utility of regionally constrained models of brain sex to better understand the link between sex differences in the brain and mental disorders.

Combination of clinical and spectral-CT iodine concentration for predicting liver metastasis in gastric cancer: a preliminary study.

This study aimed to determine the diagnostic efficacy of various indicators and models for the prediction of gastric cancer with liver metastasis. Clinical and spectral computed tomography (CT) data from 80 patients with gastric adenocarcinoma who underwent surgical resection were retrospectively analyzed. Patients were divided into metastatic and non-metastatic groups based on whether or not to occur liver metastasis, and the region of interest (ROI) was measured manually on each phase iodine map at the largest level of the tumor. Iodine concentration (IC), normalized iodine concentration (nIC), and clinical data of the primary gastric lesions were analyzed. Logistic regression analysis was used to construct the clinical indicator (CI) and clinical indicator-spectral CT iodine concentration (CI-Spectral CT-IC) Models, which contained all of the parameters with statistically significant differences between the groups. Receiver operating characteristic (ROC) curves were constructed to evaluate the accuracy of the models. The metastatic group showed significantly higher levels of Cancer antigen125 (CA125), carcinoembryonic antigen (CEA), IC, and nIC in the arterial phase, venous phase, and delayed phase than the non-metastatic group (all p < 0.05). Normalized iodine concentration Venous Phase (nICVP) exhibited a favorable performance among all IC and nIC parameters for forecasting gastric cancer with liver metastasis (area under the curve (AUC), 0.846). The combination model of clinical data with significant differences and nICVP showed the best diagnostic accuracy for predicting liver metastasis from gastric cancer, with an AUC of 0.897. nICVP showed the best diagnostic efficacy for predicting gastric cancer with liver metastasis. Clinical Indicators-normalized ICVP model can improve the prediction accuracy for this condition.

Anomaly Detection for Industrial Time Series Data Based on Correlation Analysis and CNN-BiLSTM with Self-attention

This paper aims to propose an anomaly detection model for industrial time series data based on correlation analysis and CNN-BiLSTM with self-attention to solve the problem of abnormal data detection in the field of industrial data analysis. Industrial data anomaly detection is an important task in the industrial field, which can help people to timely understand the production operation status and real-time record and perception of the operating environment. This paper introduces two key technologies: correlation analysis and CNN-BiLSTM with self-attention, and how to combine them to build an effective anomaly detection model for industrial time series data. Through experimental evaluation, this paper proves the effectiveness and superiority of the proposed model in industrial data anomaly detection tasks.

A Single-Buyer Model of Imbalance Cost Pass-Through Pricing Forecasting in the Malaysian Electricity Supply Industry

The imbalance cost pass-through (ICPT) is a flexible component of the incentive-based regulation (IBR) that empowers power producers to adjust tariffs in response to variable fuel prices, thereby enhancing the economic resilience of electricity generation. In Malaysia, the Energy Commission has conducted biannual reviews of fuel and other generation costs. Any cost savings or increases identified during these reviews will be passed on to customers in the form of rebates or surcharges. Meanwhile, if an increment in the ICPT price signal can be provided to electricity providers and consumers, early preparation for operation budgeting can be realised, and energy management program development can be properly prepared. Due to this reason, this study proposes ICPT price forecasting for the electricity market in Peninsular Malaysia that will benefit the stakeholders. The study aims to construct an ICPT-related baseline model for the peninsular generation data by employing three forecasting methods. The forecasting performance is analysed using the mean absolute percentage error (MAPE). In light of our findings, the ARIMA method is one of the most accurate forecasting methods for fuel prices compared to the moving average (MA) and LSSVM methods. The observed price differences between the ARIMA and LSSVM models for ICPT are minimal. The ICPT price for July–December 2022 and January–June 2023 is MYR 0.21/kWh for the ARIMA and MYR 0.18/kWh for LSSVM, which are close to the actual TNB’s ICPT tariff. As for forecasting, the ICPT price is expected to drop in the next announcement. The findings of this study may have a positive impact on the sustainability of the energy sector in Malaysia.

Exploring risky health behaviors and vulnerability to sexually transmitted diseases among transnational undocumented labor migrants from Bangladesh: a qualitative study

BackgroundIn Bangladesh, remittances constitute a substantial portion of the country’s foreign exchange earnings and serve as a primary source of income. However, a considerable number of Bangladeshi citizens reside overseas without proper documentation, exposing them to significant challenges such as limited access to healthcare and socioeconomic opportunities. Moreover, their irregular migration status often results in engaging in risky health behaviors that further exacerbate their vulnerability. Hence, this study aimed to investigate the risky health behavior and HIV/STI susceptibility of Bangladeshi irregular international migrants residing across the globe with undocumented status.MethodsUsing a qualitative Interpretative Phenomenological Approach (IPA), 25 illegal migrants were interviewed who are currently living illegally or returned to their home country. The author used a thematic approach to code and analyze the data, combining an integrated data-driven inductive approach with a deductive approach. Concurrent processing and coding were facilitated by employing the Granheim model in data analysis.ResultsThe study identified four risky health behaviors among irregular Bangladeshi migrants: hazardous living conditions, risky jobs, suicidal ideation, and tobacco consumption. Additionally, the authors found some HIV/STI risk behavior among them including engaging in unprotected sex, consuming alcohol and drugs during sexual activity, and having limited access to medical facilities.ConclusionsThe findings of this study can be used by health professional, governments, policymakers, NGOs, and concerned agencies to develop welfare strategies and initiatives for vulnerable undocumented migrant workers.

The modeling of the Organic Molecules Rejection using the Bootstrap Aggregated Neural Networks for the evaluation of the Forward Osmosis Process performance

&lt;p&gt;The forward osmosis process is currently more studied to be a replacement for another consuming-energy process, for this, many works show up the rejection of different molecules, energy consumption, and modeling of different objectives related to FO process. Our study consists to model the rejection of organic molecules (neutral and ionic) by FO process; however, this paper is the simultaneous applications of the single neural network based on quantitative- structure properties relationship (QSPR-SNN) and the bootstrap aggregated neural network (BANN) to predict the rejection of 53 OM. According to the results obtained, the coefficient correlation "R" is used to evaluate the performance of each model for the unseen data, the QSPR-BANN gives R value equal to 0.9909 higher than the value of the SNN which is 0.9401, the Root Mean Square Error of the QSPR-BANN is less than that of the QSPR-SNN with values equal to 0.5764% and 1.2826% respectively.&lt;/p&gt;

Approximate inferences for Bayesian hierarchical generalised linear regression models

SummaryGeneralised linear mixed regression models are fundamental in statistics. Modelling random effects that are shared by individuals allows for correlation among those individuals. There are many methods and statistical packages available for analysing data using these models. Most require some form of numerical or analytic approximation because the likelihood function generally involves intractable integrals over the latents. The Bayesian approach avoids this issue by iteratively sampling the full conditional distributions for various blocks of parameters and latent random effects. Depending on the choice of the prior, some full conditionals are recognisable while others are not. In this paper we develop a novel normal approximation for the random effects full conditional, establish its asymptotic correctness and evaluate how well it performs. We make the case for hierarchical binomial and Poisson regression models with canonical link functions, for hierarchical gamma regression models with log link and for other cases. We also develop what we term a sufficient reduction (SR) approach to the Markov Chain Monte Carlo algorithm that allows for making inferences about all model parameters by replacing the full conditional for the latent variables with a considerably reduced dimensional function of the latents. We expect that this approximation could be quite useful in situations where there are a very large number of latent effects, which may be occurring in an increasingly ‘Big Data’ world. In the sequel, we compare our methods with INLA, which is a particularly popular method and which has been shown to be excellent in terms of speed and accuracy across a variety of settings. Our methods appear to be comparable to theirs in terms of accuracy, while INLA was faster, for the settings we considered. In addition, we note that our methods and those of others that involve Gibbs sampling trivially handle parameters that are functions of multiple parameters, while INLA approximations do not. Our primary illustration is for a three‐level hierarchical binomial regression model for data on health outcomes for patients who are clustered within physicians who are clustered within particular hospitals or hospital systems.