Binary Response Research Articles

A framework for block-wise missing data in multi-omics.

High-throughput technologies have generated vast amounts of omic data. It is a consensus that the integration of diverse omics sources improves predictive models and biomarker discovery. However, managing multiple omics data poses challenges such as data heterogeneity, noise, high-dimensionality and missing data, especially in block-wise patterns. This study addresses the challenges of high dimensionality and block-wise missing data through a regularization and constrained-based approach. The methodology is implemented in the R package bwm for binary and continuous response variables, and applied to breast cancer and exposome multi-omics datasets, achieving strong performance even in scenarios with missing data present in all omics. In binary classification task, our proposed model achieves accuracy in the range of 86% to 92%, and F1 in the range of 68% to 79%. And, in regression task the correlation between true and predicted responses is in the range of 72% to 76%. However, there is a slight decline in performance metrics as the percentage of missing data increases. In scenarios where block-wise missing data affects multiple omics, the model performance actually surpasses that of scenarios where missing data is present in only one omics. One possible explanation for this might be that the other scenarios introduce a greater diversity of observation profiles, leading to a more robust model. Depending on the specific omics being studied, there is greater consistency in feature selection when comparing block-wise missing data scenarios.

Group feature screening for ultrahigh-dimensional data missing at random

<abstract> <p>Statistical inference for missing data is common in data analysis, and there are still widespread cases of missing data in big data. The literature has discussed the practicability of two-stage feature screening with categorical covariates missing at random (IMCSIS). Therefore, we propose group feature screening for ultrahigh-dimensional data with categorical covariates missing at random (GIMCSIS), which can be used to effectively select important features. The proposed method expands the scope of IMCSIS and further improves the performance of classification learning when covariates are missing. Based on the adjusted Pearson chi-square statistics, a two-stage group feature screening method is modeled, and theoretical analysis proves that the proposed method conforms to the sure screening property. In a numerical simulation, GIMCSIS can achieve better finite sample performance under binary and multivariate response variables and multi-classification covariates. The empirical analysis through multiple classification results shows that GIMCSIS is superior to IMCSIS in imbalanced data classification.</p> </abstract>

Decoding a Cognitive Performance State from Behavioral Data in the Presence of Auditory Stimuli.

Cognitive performance state is an unobserved state that refers to the overall performance of cognitive functions. Deriving an informative observation vector as well as the adaptive model and decoder would be essential in decoding the hidden performance. We decode the performance from behavioral observation data using the Bayesian state-space approach. Forming an observation from the paired binary response with the associated continuous reaction time may lead to an overestimation of the performance, especially when an incorrect response is accompanied by a fast reaction time. We apply the marked point process (MPP) framework such that the performance decoder takes the correct/incorrect responses and the reaction time associated with correct responses as an observation. We compare the MPP-based performance with two other decoders in which the pairs of binary and continuous signals are taken as the observation; one decoder considers an autoregressive (AR) model for the performance state, and the other one employs an autoregressive-autoregressive conditional heteroskedasticity (AR-ARCH) model which incorporates the time-varying and adaptive innovation term within the model. To evaluate decoders, we use the simulated data and the n-back experimental data in the presence of multiple music sessions. The Bayesian state-space approach is a promising way to decode the performance state. With respect to individual perspective, the estimated MPP-based and ARCH-based performance states outperform the AR-based estimation. Based on the aggregated data analysis, the ARCH-based performance decoder outperforms the other decoders. Performance decoders can be employed in educational settings and smart workplaces to monitor one's performance and contribute to developing a feedback controller in closed-loop architecture to improve cognitive performance.

Frequentist and Bayesian Sample Size Determination for Single-Arm Clinical Trials Based on a Binary Response Variable: A Shiny App to Implement Exact Methods

Frequentist and Bayesian Sample Size Determination for Single-Arm Clinical Trials Based on a Binary Response Variable: A Shiny App to Implement Exact Methods

Phase change plasmonic metasurface for dynamic thermal emission modulation

Plasmonic metasurfaces with adjustable optical responses can be achieved through phase change materials (PCMs) with high optical contrast. However, the on–off behavior of the phase change process results in the binary response of photonic devices, limiting the applications to the two-stage modulation. In this work, we propose a reconfigurable metasurface emitter based on a gold nanorod array on a VO2 thin film for achieving continuously tunable narrowband thermal emission. The electrode line connecting the center of each nanorod not only enables emission excitation electrically but also activates the phase transition of VO2 beneath the array layer due to Joule heating. The change in the dielectric environment due to the VO2 phase transition results in the modulation of emissivity from the plasmonic metasurfaces. The device performances regarding critical geometrical parameters are analyzed based on a fully coupled electro-thermo-optical finite element model. This new metasurface structure extends the binary nature of PCM based modulations to continuous reconfigurability and provides new possibilities toward smart metasurface emitters, reflectors, and other nanophotonic devices.

Skill formation and income: Insights from the Indonesian migrant workers

Indonesia is among the top five countries with the highest number of seafarers globally and this seafaring expertise extends to Indonesian migrant workers on cruise ships. The enhancement of workers' skills on cruise ships is vital as it can bolster their work proficiency, paving the way for future career advancement and increased earnings. This study aimed to examine the relationship between training and the income increase of Indonesian migrant workers on cruise ships. Using a quantitative approach, the analysis involved 320 Indonesian migrant workers. The dependent and independent variables were income increase and training, respectively, with control variables including education, age, and the length of work for Indonesian migrant workers on cruise ships. Binary logistic regression was used to examine the relationship between the binary response variables. The results showed that there was a positive relationship between on-the-job training and the income increase of Indonesian migrant workers. Statistically, Indonesian migrant workers who engaged in on-the-job training had a 1.9 times higher income increase compared to those with off-the-job training. Conversely, the length of work showed a negative relationship with income increase. The analysis indicated that as the tenure of Indonesian migrant workers on cruise ships extended, the likelihood of income decreased by .8 times. The significance of on-the-job training for Indonesian migrant workers lies in its role in enhancing skills, competitiveness, and total performance, ultimately resulting in higher income.

Random State Initialized Logistic Regression for Improved Heart Attack Prediction

One of the primary causes of death in Indonesia is heart attacks. Therefore, an effective method of pre-diction is required to determine whether a patient is experiencing a heart attack. One efficient approach is to use machine learning models. However, it is still rare to find machine learning models that have good performance in predicting heart attacks. This study aims to develop a machine learning model on Logistic Regression algorithm in predicting heart attack. Logistic Regression is one of the machine learning meth-ods that can be used to study the relationship between a binary response variable [0,1] and a set of pre-dictor variables, and can be used directly to calculate probabilities. In this study, a random state is ini-tialized in the Logistic Regression model in order to stabilize the training of the machine learning model and increase the precision of the proposed method. The results of this study show that the proposed model can be a method that has good performance in predicting heart attack disease.

A Hybrid Paillier Encryption and Randomized Response Technique Model for Surveying Sensitive Questions with Binary Response

A Hybrid Paillier Encryption and Randomized Response Technique Model for Surveying Sensitive Questions with Binary Response

Modelling Factors that Predict Differences in Childhood Mortality in Lagos Communities Using Prognostic Logistic and Poisson Regression Models

Lagos State is a city with one of the largest concentration of people in the world with heterogenous behaviour and cultural beliefs. There are different prognostic models in the medical sciences, yet their real life application, especially to childhood mortality is limited. There are variations in childhood mortality rate across different communities in Lagos State. Childhood mortality is a subject of interest to World Health Organization (WHO) and one of the major Millennium Development Goals. In 2014, the Special Adviser to the Lagos State Governor on Public Health, Dr. Yewande Adeshina said that under-5 and infant mortality rates in Lagos state have reduced due to various health interventions being implemented in the State. However, the truth of the matter is that childhood mortality is still high and this is an indication that we still have lots of work to do in this regard. In this paper, prognostic models were used in modelling factors that predict the differences in childhood mortality in Lagos communities. Six models, two each from logistic regression, linear regression and Poisson regression models were used. Primary data were collected from mothers that fall in the age bracket (15-49), who reside in any of the 5 divisions of Lagos State, namely Ikorodu, Badagry, Lagos Mainland/Ikeja, Lagos Island and Epe. Five variables were identified as covariates. The prognostic multi-variable models were employed. The binary logistic regression model with 5 covariates was selected as the best model for the binary response variable, while the Poisson regression model with 4 covariates was selected as the best model for the count response variable. At the end of the research, Ikorodu, Badagry and Epe communities have higher than expected childhood mortality rates. Also, we estimated childhood mortality rate in Lagos State and measured the variations in childhood mortality across Lagos communities. The factors that predict these variations were detected and control measures were recommended to reduce the difference in childhood mortality rate in Lagos State.

Binary Classification with Imbalanced Data.

When the binary response variable contains an excess of zero counts, the data are imbalanced. Imbalanced data cause trouble for binary classification. To simplify the numerical computation to obtain the maximum likelihood estimators of the zero-inflated Bernoulli (ZIBer) model parameters with imbalanced data, an expectation-maximization (EM) algorithm is proposed to derive the maximum likelihood estimates of the model parameters. The logistic regression model links the Bernoulli probabilities with the covariates in the ZIBer model, and the prediction performance among the ZIBer model, LightGBM, and artificial neural network (ANN) procedures is compared by Monte Carlo simulation. The results show that no method can dominate the other methods regarding predictive performance under the imbalanced data. The LightGBM and ZIBer models are more competitive than the ANN model for zero-inflated-imbalanced data sets.

Determinants of Repayment Performance of Agricultural Loans Clients of RBDI, Abuyog Branch

The study aimed to present the profile of the agricultural loan clients of RBDI, describe the loan assessment process, describe the loan monitoring process in RBDI, identify the determinants of loan repayment performance of RBDI agricultural loan clients and to recommend courses of action to improve loan application assessment, loan monitoring, and loan collection activities of RBDI. The study used the Logit Regression Analysis and the model Y = β0 + β1X1 + β2X2 + β3X3 + β4X4 + . . . . . . . βnXn + e was developed. The binary response is non-payment (default) and payment (non-default). The determinants were the socio-demographic and socio-economic characteristics of the agricultural loan clients, as well as the loan application and monitoring activities of the bank. The profile of the agricultural loan clients of RBDI showed commonalities of the clients in terms of age (63% are above 45 years old), most of them were not able to finish college (53% secondary level) and most are married (77%) with more than 3 dependents (73%). Age was determined as a statistically significant factor affecting the non-payment or loan default. The increase in family size and schooling dependents also raises the probability of default. The result that higher educational level and increase in dependents reduces the probability of default has to be further verified. As the size of land increases and the farming experience increases, the probability of loan default decreases.

Restricted Latent Class Models for Nominal Response Data: Identifiability and Estimation.

Restricted latent class models (RLCMs) provide an important framework for diagnosing and classifying respondents on a collection of multivariate binary responses. Recent research made significant advances in theory for establishing identifiability conditions for RLCMs with binary and polytomous response data. Multiclass data, which are unordered nominal response data, are also widely collected in the social sciences and psychometrics via forced-choice inventories and multiple choice tests. We establish new identifiability conditions for parameters of RLCMs for multiclass data and discuss the implications for substantive applications. The new identifiability conditions are applicable to a wealth of RLCMs for polytomous and nominal response data. We propose a Bayesian framework for inferring model parameters, assess parameter recovery in a Monte Carlo simulation study, and present an application of the model to a real dataset.

Debiased inference for heterogeneous subpopulations in a high-dimensional logistic regression model

Due to the prevalence of complex data, data heterogeneity is often observed in contemporary scientific studies and various applications. Motivated by studies on cancer cell lines, we consider the analysis of heterogeneous subpopulations with binary responses and high-dimensional covariates. In many practical scenarios, it is common to use a single regression model for the entire data set. To do this effectively, it is critical to quantify the heterogeneity of the effect of covariates across subpopulations through appropriate statistical inference. However, the high dimensionality and discrete nature of the data can lead to challenges in inference. Therefore, we propose a novel statistical inference method for a high-dimensional logistic regression model that accounts for heterogeneous subpopulations. Our primary goal is to investigate heterogeneity across subpopulations by testing the equivalence of the effect of a covariate and the significance of the overall effects of a covariate. To achieve overall sparsity of the coefficients and their fusions across subpopulations, we employ a fused group Lasso penalization method. In addition, we develop a statistical inference method that incorporates bias correction of the proposed penalized method. To address computational issues due to the nonlinear log-likelihood and the fused Lasso penalty, we propose a computationally efficient and fast algorithm by adapting the ideas of the proximal gradient method and the alternating direction method of multipliers (ADMM) to our settings. Furthermore, we develop non-asymptotic analyses for the proposed fused group Lasso and prove that the debiased test statistics admit chi-squared approximations even in the presence of high-dimensional variables. In simulations, the proposed test outperforms existing methods. The practical effectiveness of the proposed method is demonstrated by analyzing data from the Cancer Cell Line Encyclopedia (CCLE).

A joint normal-binary (probit) model for high-dimensional longitudinal data

In many biomedical studies multiple responses are collected over time, which results in highdimensional longitudinal data. It is often of interest to model the continuous and binary responses jointly, which can be done with joint generalized mixed models in which the association is modelled through random effects. Investigating the association between the responses is often limited to scrutinizing the correlations between the latent random effects. In this article, this approach is extended by deriving closed-form formulas for the manifest correlations (and corresponding standard errors), which reflects the correlation between the observed responses as observed. In addition, the marginal joint model is constructed, from which predictions of subvectors of one response conditional on subvectors of other response(s) and potentially a subvector of the history of the response can be derived. Corresponding prediction and confidence intervals are constructed. Two case studies are discussed, in which further pseudo-likelihood methodology is applied to reduce the computational complexity.

Dynamic fit index cutoffs for categorical factor analysis with Likert-type, ordinal, or binary responses.

Scale validation is vital to psychological research because it ensures that scores from measurement scales represent the intended construct. Fit indices are commonly used to provide quantitative evidence that a proposed factor structure is plausible. However, there is a mismatch between guidelines for evaluating fit of the factor models and the data that most researchers have. Namely, fit guidelines are based on the simulations that assume item responses are collected on a continuous scale whereas most researchers collect discrete responses such as with a Likert-type scale. In this article, we show that common guidelines derived from assuming continuous responses (e.g., root-mean-square error of approximation < 0.06, comparative fit index > 0.95) do not generalize to factor models applied to discrete responses. Specifically, discrete responses provide less information than continuous responses, so less information about misfit is passed to fit indices. Traditional guidelines, therefore, end up being too lenient and lose their ability to identify that a model may have a poor fit. We provide one possible solution by extending the recently developed dynamic fit index framework to accommodate discrete responses common in psychology. We conduct a simulation study to provide evidence that the proposed method consistently distinguishes between well-fitting and poorly fitting models. Results showed that our proposed cutoffs maintained at least 90% sensitivity to misspecification across studied conditions, whereas traditional cutoffs were highly inconsistent and frequently exhibited sensitivity below 50%. The proposed method is included in the dynamic R package and as a web-based Shiny application to make it easily accessible to psychologists. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Exact likelihoods for N‐mixture models with time‐to‐detection data

SummaryThis paper is concerned with the formulation of ‐mixture models for estimating the abundance and probability of detection of a species from binary response, count and time‐to‐detection data. A modelling framework, which encompasses time‐to‐first‐detection within the context of detection/non‐detection and time‐to‐each‐detection and time‐to‐first‐detection within the context of count data, is introduced. Two observation processes which depend on whether or not double counting is assumed to occur are also considered. The main focus of the paper is on the derivation of explicit forms for the likelihoods associated with each of the proposed models. Closed‐form expressions for the likelihoods associated with time‐to‐detection data are new and are developed from the theory of order statistics. A key finding of the study is that, based on the assumption of no double counting, the likelihoods associated with times‐to‐detection together with count data are the product of the likelihood for the counts alone and a term which depends on the detection probability parameter. This result demonstrates that, in this case, recording times‐to‐detection could well improve precision in estimation over recording counts alone. In contrast, for the double counting protocol with exponential arrival times, no information was found to be gained by recording times‐to‐detection in addition to the count data. An R package and an accompanying vignette are also introduced in order to complement the algebraic results and to demonstrate the use of the models in practice.

Medical student specialty decision-making and perceptions of neurosurgery. Part 1: Role of gender.

Although women account for 50% of medical school graduates, less than 30% of neurosurgery residency applicants and less than 10% of neurosurgeons are female. In order to diversify the field of neurosurgery and recruit more women, it is necessary to understand why there is a disproportionately low entry rate into neurosurgery by female medical students. Factors contributing to specialty decision-making and perceptions of neurosurgery among medical students and residents, specifically differences by gender, have not been studied. The authors aimed to investigate these differences using quantitative and qualitative methods. A Qualtrics survey was administered at the authors' institution to all medical students and resident physicians to assess factors influencing medical specialty decision-making and perceptions of neurosurgery. Likert scale responses converted to numerical values on a 5-point scale were analyzed with the Mann-Whitney U-test. The chi-square test was performed on binary responses. Semistructured interviews were conducted in a subset of survey respondents and were analyzed by using the grounded theory method. Of the 272 survey respondents, 48.2% were medical students and 61.0% were female. When making specialty decisions, female medical students considered maternity/paternity leave more (p = 0.028) than their male counterparts. Female medical students were more hesitant toward neurosurgery due to maternity/paternity needs (p = 0.031) and the technical skill required (p = 0.020) than male medical students. Across both genders, the majority of medical students were hesitant toward neurosurgery due to opportunities for work/life integration (93%), length of training (88%), malignancy of the field (76%), and perceived happiness of the people in the field (76%). Female residents indicated they were more likely than male residents to consider the perceived happiness of the people in the field (p = 0.003), shadowing experiences (p = 0.019), and elective rotations (p = 0.004) when making specialty decisions. Two major themes emerged from the semistructured interviews: 1) maternity needs were more of a concern for women and 2) length of training was a concern for many individuals. Compared with their male counterparts, female students and residents consider different factors and experiences when choosing a medical specialty and have different perceptions of neurosurgery. Exposure to and education within neurosurgery, specifically maternity needs, may help address hesitancy in pursuing a neurosurgical career among female medical students. However, cultural and structural factors may need to be addressed within neurosurgery in order to ultimately increase representation of women.

Views and perceptions of amyloid imaging in a preclinical Alzheimer’s disease trial

AbstractBackgroundStudies have shown that many cognitively unimpaired older individuals are interested in learning their Alzheimer’s disease (AD) biomarker results. Yet, little is known about what motivates participation in preclinical AD trials that require amyloid biomarker testing and disclosure, and whether these motivations differ by biomarker results (elevated or not‐elevated amyloid).MethodUsing data from the EARLY trial, a preclinical AD trial testing the safety and efficacy of a BACE1 inhibitor, we analyzed participant responses on the Views and Perceptions of Amyloid Imaging (VPAI) scale (Table 1) that was collected during the screening process; that is, before undergoing biomarker testing and disclosure. Participants were asked to rate how important 9 potential reasons for seeking amyloid imaging were to them. Response options included “not at all” (1), “a little” (2), “somewhat” (3), “very” (4), and “extremely” (5). We categorized raw scores into binary responses; those with high endorsement (scores ≥4) and those with low‐to‐moderate endorsement (scores &lt;4). We stratified participant responses by amyloid group to examine whether their biomarker status impacted participant item‐level VPAI scores.ResultPrior to biomarker testing and disclosure, participants most often endorsed the desire to contribute to research as a motivation to participate (Table 1). In general, the not‐elevated amyloid group had higher endorsement for the VPAI items than the elevated amyloid group. The proportions of participants endorsing “confirm feeling that I might be developing dementia” and “preparing family members” were observed to be higher among participants with elevated compared to not‐elevated amyloid; whereas “curiosity” was endorsed by a higher proportion of the not‐elevated amyloid group.ConclusionWhile there are individual differences in reasons for seeking an amyloid scan, preclinical AD trial participants may be largely motivated by altruistic factors. Differences in factors that motivate participants to learn biomarker results may be apparent between amyloid groups even before biomarker disclosure.

Unraveling the relationship between heart rate and agitation in people with dementia

AbstractBackgroundAgitation is a critical behavioral and psychological symptom of dementia, leading alone to increased caregiver burden, more frequent hospitalization, and earlier institutionalization. The results presented here provide preliminary insights into one facet of multimodal agitation detection, specifically analyzing the relationship between heart rate and agitation.MethodTo date, 26 people with dementia have been included in our study to detect agitation using a multimodal approach1. In this analysis, heart rate is derived from the wristwatch wearable measuring photoplethysmography (PPG). Data points were included in subsequent analysis if the signal quality index of the original PPG signal was &gt; = 0.8 (on a scale from 0‐1). For each complete survey, the mean heart rate (HR) was extracted from 1‐minute intervals in a 6‐minute window centered on the survey timestamp (± 3 minutes). For each mean HR per minute, the presence or absence of agitation was derived from the Pittsburgh Agitation Scale integrated into the survey. The data was centered within‐patient. Due to the nested structure of the data, hierarchical mixed models were used to predict presence of agitation as a binary response variable.ResultThe resulting dataset of high‐quality HR data consists of 1286 data points, from 21 patients (Age mean: 81 (8.3), nfemale: 8) The 3‐parameter model built included mean heart rate as an explanatory fixed effect, and two random intercepts: day and patient. In the model, increased HR compared to a participant’s average HR over a week, was significantly associated with an increase in odds of agitation (p&lt; 0.001). Every 1 beat increase in mean HR over a minute, from the participants’ mean heart rate, increased the odds of that 1‐minute interval indicating agitation by 7.68% (CI: 5.43‐10.1).ConclusionThese results provide insight into the relationship between heart rate and agitation, possibly indicating activation of the autonomic nervous system, due to the association of increased HR with increased odds of agitation. Analyses like these provide the basis for future agitation detection algorithms that account for a multi‐faceted patient response.1. Davidoff, et al, 2022, Innovation in Aging, Volume 6, Issue 7, 2022, igac064, https://doi.org/10.1093/geroni/igac064

Medical student specialty decision-making and perceptions of neurosurgery. Part 2: Role of race/ethnicity.

Although individuals underrepresented in medicine (URM) make up 33% of the United States population, only 12.6% of medical school graduates identify as URM; the same percentage of URM students comprises neurosurgery residency applicants. More information is needed to understand how URM students are making specialty decisions and their perceptions of neurosurgery. The authors aimed to evaluate the differences between URM and non-URM medical students and residents in terms of the factors that contribute to specialty decision-making and perceptions of neurosurgery. A survey was administered to all medical students and resident physicians at a single Midwestern institution to assess factors influencing medical student specialty decision-making and perceptions of neurosurgery. Likert scale responses converted to numerical values on a 5-point scale (strongly agree was the high score of 5) were analyzed with the Mann-Whitney U-test. The chi-square test was performed on the binary responses to examine associations between categorical variables. Semistructured interviews were conducted and analyzed using the grounded theory method. Of 272 respondents, 49.2% were medical students, 51.8% were residents, and 11.0% identified as URM. URM medical students considered research opportunities more than non-URM medical students in specialty decision-making (p = 0.023). When specialty decision-making factors were assessed, URM residents less strongly considered the technical skill required (p = 0.023), their perceived fit in the field (p < 0.001), and seeing people like them in the field (p = 0.010) than their non-URM counterparts when making specialty decisions. Within both medical student and resident respondent cohorts, the authors found no significant differences between URM and non-URM respondents in terms of their specialty decision-making being affected by medical school experiences such as shadowing, elective rotations, family exposure, or having a mentor in the field. URM residents were more concerned about the opportunity to work on health equity issues in neurosurgery than non-URM residents (p = 0.005). The predominant theme that emerged from interviews was the need for more intentional efforts to recruit and retain URM individuals in medicine and specifically neurosurgery. URM students may make specialty decisions differently than non-URM students. URM students were more hesitant toward neurosurgery due to their perceived lack of opportunity for health equity work in neurosurgery. These findings further inform optimization of both new and existing initiatives to improve URM student recruitment and retention in neurosurgery.