Balanced Sampling Research Articles

Use of a Large Language Model to Assess Clinical Acuity of Adults in the Emergency Department

The introduction of large language models (LLMs), such as Generative Pre-trained Transformer 4 (GPT-4; OpenAI), has generated significant interest in health care, yet studies evaluating their performance in a clinical setting are lacking. Determination of clinical acuity, a measure of a patient's illness severity and level of required medical attention, is one of the foundational elements of medical reasoning in emergency medicine. To determine whether an LLM can accurately assess clinical acuity in the emergency department (ED). This cross-sectional study identified all adult ED visits from January 1, 2012, to January 17, 2023, at the University of California, San Francisco, with a documented Emergency Severity Index (ESI) acuity level (immediate, emergent, urgent, less urgent, or nonurgent) and with a corresponding ED physician note. A sample of 10 000 pairs of ED visits with nonequivalent ESI scores, balanced for each of the 10 possible pairs of 5 ESI scores, was selected at random. The potential of the LLM to classify acuity levels of patients in the ED based on the ESI across 10 000 patient pairs. Using deidentified clinical text, the LLM was queried to identify the patient with a higher-acuity presentation within each pair based on the patients' clinical history. An earlier LLM was queried to allow comparison with this model. Accuracy score was calculated to evaluate the performance of both LLMs across the 10 000-pair sample. A 500-pair subsample was manually classified by a physician reviewer to compare performance between the LLMs and human classification. From a total of 251 401 adult ED visits, a balanced sample of 10 000 patient pairs was created wherein each pair comprised patients with disparate ESI acuity scores. Across this sample, the LLM correctly inferred the patient with higher acuity for 8940 of 10 000 pairs (accuracy, 0.89 [95% CI, 0.89-0.90]). Performance of the comparator LLM (accuracy, 0.84 [95% CI, 0.83-0.84]) was below that of its successor. Among the 500-pair subsample that was also manually classified, LLM performance (accuracy, 0.88 [95% CI, 0.86-0.91]) was comparable with that of the physician reviewer (accuracy, 0.86 [95% CI, 0.83-0.89]). In this cross-sectional study of 10 000 pairs of ED visits, the LLM accurately identified the patient with higher acuity when given pairs of presenting histories extracted from patients' first ED documentation. These findings suggest that the integration of an LLM into ED workflows could enhance triage processes while maintaining triage quality and warrants further investigation.

Postoperative recreational sports and subjective shoulder function: a comprehensive analysis following intramedullary stabilization of displaced midshaft clavicular fractures.

Introduction Sports-related outcomes and the role of recreational sports activities for shoulder function after intramedullary stabilization (IMS) of displaced midshaft clavicular fractures (DMCFs) in the general population are not well known. In this study, we aimed to determine the sport-related outcomes (return-to-sports [RTS] rate, type of sports, time until RTS, and intensity) and to explore the role of sports after IMS of DMCFs. Materials and Methods This single-center, retrospective, cohort study included patients who underwent IMS of DMCFs between 2009 and 2022 at a Level II trauma center in Germany, experienced no major complications, and had completed at least 1 year of follow-up. Propensity score matching was conducted to obtain a balanced sample of patients who did not engage (cases) and engaged (controls) in postoperative sports activities by adjusting for age and fracture complexity. Groups were compared to assess the impact of regular sports activities on subjective shoulder functioning at follow-up, as evaluated using the Disabilities of the Arm, Shoulder, and Hand and Oxford Shoulder Score (OSS) questionnaires, after controlling for the patient- (i.e., sex and smoking) and treatment- (i.e., surgery duration and physical therapy) factors. Results Among the 199 patients included, the RTS rate was 97.5%, and 160 patients practiced regular postoperative activity, mostly in the same sport and intensity. In the matched cohort (39 cases and 39 controls), practicing regular sports activities postoperatively was the only independent factor associated with a higher OSS in the multiple regression analysis (unstandardized regression coefficient = 2.40; Bias-corrected and accelerated 95% confidence interval [0.28, 4.69]). Conclusions The sport-related outcomes after IMS of DMCFs in our cohort were comparable to those achieved after plate osteosynthesis, and IMS reported in the literature. Recreational sports activities benefitted subjective shoulder function, thereby encouraging further research and potentially influencing management policies. Level of evidence Grade IV-a retrospective observational cohort study.

Do product offering and service quality affect customer satisfaction in Islamic and conventional banks? Evidence from an oil-based economy

Purpose This paper aims to investigate the effect of product offering and other service quality (SQ) dimensions on the satisfaction of the customers of both Islamic and conventional banks, using evidence from an oil-based economy that is based on a prolonged SERVQUAL model with 11 dimensions and other statistical analysis methods. Design/methodology/approach The data were collected from 461 Islamic and conventional bank customers in Qatar via a survey and several tests were used to test certain hypotheses. Component analysis, factor analysis and gap and ascendency analysis were used in this study. Afterward, a correlation analysis and regression model were used to examine the hypotheses and validate the instruments used. Findings The results show that regardless of the type of bank, customers always have greater expectations of the services than they had perceived. A customer’s expectation of the product on offer is the only dimension that is significantly different in relation to the two types of banks. However, reliability, competence, responsiveness, credibility and empathy dimensions are significantly different of the two types of banks in customers’ perception of quality. In addition, the results suggest that both types of banks need to concentrate their efforts on the product offering, competence and courtesy dimensions. Research limitations/implications The size of our sample of Islamic and conventional banks is unequally balanced. Future studies might therefore choose an equally balanced sample. Practical implications Bank managers in both types of banks need to continue improving the quality of their service including product offering and to adopt advanced methods to enhance customer satisfaction (CS) and reduce the gaps in quality in the dimensions used. Furthermore, managers in both types of banks need to put more emphasis on product offering, competence, courtesy and communication if they wish to improve SQ. Moreover, Islamic banks must guarantee that they possess competent, highly trained personnel who are familiar with Islamic finance products, so as to enhance the quality of service and attract customers. Originality/value To the best of the author’s knowledge, this is the first study to investigate the effect of product offering and other dimensions of SQ on CS in both Islamic and conventional banks by using 11 dimensions of SQ. In addition, it provides evidence of gaps in SQ, at the dimensions level, for both types of banks in an oil-based economy. The results of this study are valuable in helping decision-makers and bank managers who wish to raise the level of SQ and improve CS and in validating the results from other countries with a dual financial system.

Domain Feature Decomposition for Efficient Object Detection in Aerial Images

Object detection in UAV aerial images faces domain-adaptive challenges, such as changes in shooting height, viewing angle, and weather. These changes constitute a large number of fine-grained domains that place greater demands on the network’s generalizability. To tackle these challenges, we initially decompose image features into domain-invariant and domain-specific features using practical imaging condition parameters. The composite feature can improve domain generalization and single-domain accuracy compared to the conventional fine-grained domain-detection method. Then, to solve the problem of the overfitting of high-frequency imaging condition parameters, we mixed images from different imaging conditions in a balanced sampling manner as input for the training of the detection network. The data-augmentation method improves the robustness of training and reduces the overfitting of high-frequency imaging parameters. The proposed algorithm is compared with state-of-the-art fine-grained domain detectors on the UAVDT and VisDrone datasets. The results show that it achieves an average detection precision improvement of 5.7 and 2.4, respectively. The airborne experiments validate that the algorithm achieves a 20 Hz processing performance for 720P images on an onboard computer with Nvidia Jetson Xavier NX.

Choir singing is associated with enhanced structural connectivity across the adult lifespan.

The global ageing of populations calls for effective, ecologically valid methods to support brain health across adult life. Previous evidence suggests that music can promote white matter (WM) microstructure and grey matter (GM) volume while supporting auditory and cognitive functioning and emotional well-being as well as counteracting age-related cognitive decline. Adding a social component to music training, choir singing is a popular leisure activity among older adults, but a systematic account of its potential to support healthy brain structure, especially with regard to ageing, is currently missing. The present study used quantitative anisotropy (QA)-based diffusion MRI connectometry and voxel-based morphometry to explore the relationship of lifetime choir singing experience and brain structure at the whole-brain level. Cross-sectional multiple regression analyses were carried out in a large, balanced sample (N = 95; age range 21-88) of healthy adults with varying levels of choir singing experience across the whole age range and within subgroups defined by age (young, middle-aged, and older adults). Independent of age, choir singing experience was associated with extensive increases in WM QA in commissural, association, and projection tracts across the brain. Corroborating previous work, these overlapped with language and limbic networks. Enhanced corpus callosum microstructure was associated with choir singing experience across all subgroups. In addition, choir singing experience was selectively associated with enhanced QA in the fornix in older participants. No associations between GM volume and choir singing were found. The present study offers the first systematic account of amateur-level choir singing on brain structure. While no evidence for counteracting GM atrophy was found, the present evidence of enhanced structural connectivity coheres well with age-typical structural changes. Corroborating previous behavioural studies, the present results suggest that regular choir singing holds great promise for supporting brain health across the adult life span.

Does corporate governance influence environmental, social and governance disclosure practices of state‐owned enterprises? An international study

AbstractThis paper provides fresh insights into the impact of corporate governance mechanisms on environmental, social and governance (ESG) disclosure practices of state‐owned enterprises (SOEs). To accomplish this study's research objective, we collected ESG and corporate governance data from Refinitiv Eikon's database on a balanced sample of 253 SOEs from 37 worldwide countries over 5 years (2018–2022), obtaining a total of 1265 observations. A battery of fixed and random effects panel regression models has been estimated to test the impact of board characteristics, like board size, board independence, board gender diversity, number of board meetings, Chief Executive Officer (CEO) duality and the presence of a corporate social responsibility (CSR) committee, on overall and individual ESG disclosure scores of sampled SOEs. Results show that while board size and CEO duality negatively affect SOE ESG disclosure, board independence and gender diversity, as well as the number of board meetings and the existence of a CSR committee, exert a positive influence.

Coping with the Inequity and Inefficiency of the H-Index: A Cross-Disciplinary Empirical Analysis

This paper measures two main inefficiency features (many publications other than articles; many co-authors’ reciprocal citations) and two main inequity features (more co-authors in some disciplines; more citations for authors with more experience). It constructs a representative dataset based on a cross-disciplinary balanced sample (10,000 authors with at least one publication indexed in Scopus from 2006 to 2015). It estimates to what extent four additional improvements of the H-index as top-down regulations (∆Hh = Hh − Hh+1 from H1 = based on publications to H5 = net per-capita per-year based on articles) account for inefficiency and inequity across twenty-five disciplines and four subjects. Linear regressions and ANOVA results show that the single improvements of the H-index considerably and decreasingly explain the inefficiency and inequity features but make these vaguely comparable across disciplines and subjects, while the overall improvement of the H-index (H1–H5) marginally explains these features but make disciplines and subjects clearly comparable, to a greater extent across subjects than disciplines. Fitting a Gamma distribution to H5 for each discipline and subject by maximum likelihood shows that the estimated probability densities and the percentages of authors characterised by H5 ≥ 1 to H5 ≥ 3 are different across disciplines but similar across subjects.

School closures and their implications for student outcomes: evidence from Lithuania

ABSTRACT This paper studies the effect of school closures on student outcomes in the Lithuanian context. Using administrative student-level data from 2013–2017 and propensity score matching, we create a balanced sample of control and treatment groups. In contrast to other studies, we focus on students in the final years of high school, possibly eliciting the upper bar of the disruption effect. We also follow students after high school graduation, providing evidence on labour market outcomes. We find that school closure has a small, negative effect on only some exam outcomes and the probability of enrolling at a university, suggesting that the disruption effect is small even for students in the final years of high school.

Rebalancing network with knowledge stability for class incremental learning

Class incremental learning (CIL) has been proposed to solve the problem of learning to classify new classes while maintaining the performance on old classes. A typical strategy is to update the old classification model with entire new class data and a few old exemplars, which face serious performance degradation on old classes. The main reasons come down to class imbalance between old and new classes along with catastrophic forgetting towards old classes. Most existing CIL methods have proposed solving the above two issues in classification space, ignoring their adverse effects of overlapping between old and new classes and confusion among old classes in feature space. In this paper, we propose a rebalancing network with knowledge stability (RNKS), aiming to adequately retain the model performance on old classes in CIL by solving class imbalance and catastrophic forgetting in feature and classification space simultaneously. In detail, the proposed RNKS mainly consists of multi-proxies rebalancing (MPR) and hybrid knowledge distillation (HKD). MPR, focusing on class imbalance, employs multi-proxies metric learning to decrease the feature overlapping between old and new classes, together with balanced data sampling to correct the skewed decision boundary. HKD, coping with catastrophic forgetting, encourages the updated model to reproduce identical feature topologies and predictions of old classes as the old model via feature relation-based and response-based distillations. Experiments on CIFAR-100 and ILSVRC datasets demonstrate the effectiveness of this work against the state-of-the-art approaches.

Relationship of Attentional Control and Intolerance of Uncertainty with Hoarding Behaviour among University Boarding Students

Background: Hoarding disorder is increasingly recognized as a significant mental health condition characterized by an excessive acquisition and difficulty in discarding possessions. While previous research has identified genetic, environmental, and neurobiological factors contributing to hoarding behaviors, studies in the Pakistani context, particularly among university students, are scarce. This study investigates the role of attentional control and intolerance of uncertainty in relation to hoarding behaviors among university boarding students in Pakistan. Objective: The primary objective of this study was to examine the relationships between hoarding behaviors, attentional control, and intolerance of uncertainty, and to explore the demographic influences on these variables, with a focus on gender differences. Methods: Utilizing a correlational and cross-sectional design, this study recruited a balanced sample of 200 university boarding students (100 males and 100 females) via simple random sampling. The Intolerance of Uncertainty Scale (IUS), Attention Control Scale (ATC), and Saving Inventory Scale (SI-R) were employed to measure the respective constructs. Data were collected through paper-based questionnaires, with analysis conducted using SPSS version 25. Statistical methods included descriptive statistics, Pearson's correlation, and t-tests for gender comparison. Results: The results indicated a moderate positive correlation between hoarding behaviors and intolerance of uncertainty (r = .33, p < .01), and a negative correlation between attentional control and intolerance of uncertainty (r = -.19, p < .01). Gender differences were significant, with males exhibiting higher mean scores in hoarding behaviors (M = 45.13, SD = 10.67) compared to females (M = 39.12, SD = 10.38), and similar patterns were observed for other studied variables. Education and socioeconomic status showed no significant effects on the relationships studied. Conclusion: The study underscores the significant impact of intolerance of uncertainty and attentional control on hoarding behaviors among university students in Pakistan, with distinct gender differences suggesting the need for targeted interventions. This research contributes to the understanding of hoarding in a novel cultural context and supports the development of specific therapeutic strategies focusing on cognitive and emotional regulation.

Orthorexia Nervosa and Healthy Orthorexia in a Physically Active North American Population.

The Teruel Orthorexia Scale (TOS) defines two related but distinct constructs: Orthorexia Nervosa (OrNe), a pathological fixation on a healthy diet, and Healthy Orthorexia (HeOr), an interest in a healthy diet independent of psychopathology. Here, we (a) assessed both types of Orthorexia in a large North American sample using the TOS and (b) explored if engaging in regular physical activity was associated with a greater risk of Orthorexia. A cohort of physically active adults (n = 927; 41% men) completed the TOS, as well as the Rapid Assessment of Physical Activity (RAPA), to broadly assess aerobic physical activity level and participation in strength and/or flexibility training. As expected, scores for HeOr and OrNe differed between participants, with lower scores for Orthorexia Nervosa in our physically active non-clinical sample. Higher HeOr scores were associated with lower BMI, and this was true for both men and women. We also found that measures of Orthorexia were associated with self-reported physical activity: active adults reporting more aerobic physical activity had higher HeOr scores, with the most active men having the highest scores. Notably, adults who reported regular strength training had higher scores for both HeOr and OrNe, with men who strength trained showing higher OrNe scores than women. Here, those who participate in regular strength training are more likely to exhibit orthorexic behaviors, and this effect was more pronounced for men than women. Prior work has validated the TOS in young, primarily female samples of non-English speakers outside the United States: present data from an age-diverse, physically active, gender balanced sample support the use of TOS for measurement of Orthorexia Nervosa and Healthy Orthorexia in English speakers and suggest that more work is needed to assess potential gender differences in these constructs.

Sequential and Concurrent Mixed-Mode Designs: A Tailored Approach

Abstract Due to rising costs and declining response rates, surveys are increasingly moving from face-to-face interviewing to a self-administered mixed-mode design. Mixed-mode surveys can be conducted using a concurrent or a sequential design. A sequential design in which the web mode is offered first is a common strategy for mixed-mode surveys as it reduces survey costs. However, when deciding which mode choice sequence to use, sample balance should also be considered. One approach to achieving a balanced sample might be to tailor the sequence of the choice of modes, or the mode choice sequence. For this purpose, we use an indicator that assigns the sampled persons to the different mode choice sequences to minimize the variability of response probabilities. In this study, we compare the sample composition achieved with a concurrent and a sequential design. Additionally, we investigate whether indicator-based tailoring of the two mode choice sequences can improve sample composition. We implemented a randomized experiment in the 2021 German General Social Survey (ALLBUS), which surveyed the general population aged 18 and older in private households (N = 5,342) using a mixed-mode design (web and mail). In a first step, respondents were randomly assigned to a concurrent or a sequential design. We find that the two mode choice sequences lead to a similar sample composition. Next, we identify age as the best available single indicator of the variables known before the survey to tailor the mode choice sequence. Our analyses show that a tailored approach based on age improves the sample composition slightly.

Understanding Physician's Perspectives on AI in Health Care: Protocol for a Sequential Multiple Assignment Randomized Vignette Study.

As the availability and performance of artificial intelligence (AI)-based clinical decision support (CDS) systems improve, physicians and other care providers poised to be on the front lines will be increasingly tasked with using these tools in patient care and incorporating their outputs into clinical decision-making processes. Vignette studies provide a means to explore emerging hypotheses regarding how context-specific factors, such as clinical risk, the amount of information provided about the AI, and the AI result, may impact physician acceptance and use of AI-based CDS tools. To best anticipate how such factors influence the decision-making of frontline physicians in clinical scenarios involving AI decision-support tools, hypothesis-driven research is needed that enables scenario testing before the implementation and deployment of these tools. This study's objectives are to (1) design an original, web-based vignette-based survey that features hypothetical scenarios based on emerging or real-world applications of AI-based CDS systems that will vary systematically by features related to clinical risk, the amount of information provided about the AI, and the AI result; and (2) test and determine causal effects of specific factors on the judgments and perceptions salient to physicians' clinical decision-making. US-based physicians with specialties in family or internal medicine will be recruited through email and mail (target n=420). Through a web-based survey, participants will be randomized to a 3-part "sequential multiple assignment randomization trial (SMART) vignette" detailing a hypothetical clinical scenario involving an AI decision support tool. The SMART vignette design is similar to the SMART design but adapted to a survey design. Each respondent will be randomly assigned to 1 of the possible vignette variations of the factors we are testing at each stage, which include the level of clinical risk, the amount of information provided about the AI, and the certainty of the AI output. Respondents will be given questions regarding their hypothetical decision-making in response to the hypothetical scenarios. The study is currently in progress and data collection is anticipated to be completed in 2024. The web-based vignette study will provide information on how contextual factors such as clinical risk, the amount of information provided about an AI tool, and the AI result influence physicians' reactions to hypothetical scenarios that are based on emerging applications of AI in frontline health care settings. Our newly proposed "SMART vignette" design offers several benefits not afforded by the extensively used traditional vignette design, due to the 2 aforementioned features. These advantages are (1) increased validity of analyses targeted at understanding the impact of a factor on the decision outcome, given previous outcomes and other contextual factors; and (2) balanced sample sizes across groups. This study will generate a better understanding of physician decision-making within this context. DERR1-10.2196/54787.

How health risk of COVID-19 resurgence may trigger individual tendencies regarding COVID-19-related cyberchondria: Investigating the I-PACE model

Cyberchondria during the COVID-19 pandemic has posed significant challenges to public well-being. Previous lstudies have implicated perceived health risk in COVID-19-related cyberchondria. However, the effect of perceived COVID-19 health risk on cyberchondria and potential underlying mechanisms remains unclear. In this study, we developed a model based on the general addictive behavior framework of the I-PACE (Interaction-Person-Affect-Cognition-Execution) model to describe how induction of health risk of COVID-19 resurgence may influence individual tendencies regarding COVID-19-related cyberchondria. Through a scenario-based experiment, we tested this model with a balanced sample of 984 participants from China. The results revealed that the effect of induction of health risk of COVID-19 resurgence on tendencies regarding COVID-19-related cyberchondria was fully mediated by individual's risk perception and health information craving. Additionally, moderation analysis showed that an individual's general inhibitory control did not moderate the development of cyberchondria. Theoretical and practical implications are provided.

Kreĭn twin support vector machines for imbalanced data classification

Conventional classification assumes a balanced sample distribution among classes. However, such a premise leads to biased performance over the majority class (with the highest number of instances). The Twin Support Vector Machines (TWSVM) obtained great prominence due to their low computational burden compared to the standard SVM. Besides, traditional machine learning seeks methods whose solution depends on a convex problem or semi-positive definite similarity matrices. Yet, this kind of matrix cannot adequately represent many real-world applications. The above defines the need to use non-negative measures as an indefinite function in a Reproducing Kernel Kreĭn Space (RKKS). This paper proposes a novel approach called Kreĭn Twin Support Vector Machines (KTSVM), which appropriately incorporates indefinite kernels within a TWSVM-based gradient optimization. To code pertinent input patterns within an imbalanced data discrimination, our KTSVM employs an implicit mapping to a RKKS. Also, our approach takes advantage of the TWSVM scheme’s benefits by creating two parallel hyperplanes. This promotes the KTSVM optimization in a gradient-descent framework. Results obtained on synthetic and real-world datasets demonstrate that our solution performs better in terms of imbalanced data classification than state-of-the-art techniques.

Estimating the timber value of a forest property using geographically balanced samples and unoccupied aerial vehicle data

Abstract A common task in forestry is to determine the value of a forest property, and timber is the most valuable component of that property. Remotely sensed data collected by an unoccupied aerial vehicle (UAV) are suited for this purpose as most forest properties are of a size that permits the efficient collection of UAV data. These UAV data, when linked to a probability sample of field plots, enable the model-assisted (MA) estimation of the timber value and its associated uncertainty. Our objective was to estimate the value of timber (€/ha) in a 40-ha forest property in Finland. We used a systematic sample of field plots (n = 160) and 3D image point cloud data collected by an UAV. First, we studied the effects of spatial autocorrelation on the variance estimates associated with the timber value estimates produced using a field data-based simple expansion (EXP) estimator. The variance estimators compared were simple random sampling, Matérn, and a variant of the Grafström–Schelin estimator. Second, we compared the efficiencies of the EXP and MA estimators under different sampling intensities. The sampling intensity was varied by subsampling the systematic sample of 160 field plots. In the case of the EXP estimator, the simple random sampling variance estimator produced the largest variance estimates, whereas the Matérn estimator produced smaller variance estimates than the Grafström–Schelin estimator. The MA estimator was more efficient than the EXP estimator, which suggested that the reduction of sampling intensity from 160 to 60 plots is possible without deterioration in precision. The results suggest that the use of UAV data improves the precision of timber value estimates compared to the use of field data only. In practice, the proposed application improves the cost-efficiency of the design-based appraisal of a forest property because expensive field workload can be reduced by means of UAV data.

State-owned enterprises as countercyclical instruments: Quasi-experimental evidence from the infrastructure sector

This paper examines the effects of a negative macroeconomic shock on the financial performance of state-owned enterprises (SOEs) in infrastructure. The main aim of the paper is to explore whether SOEs serve as countercyclical instruments or whether they end up generating fiscal costs during a downturn, perhaps amplifying negative shocks. The paper introduces a new measure of fiscal injections that includes subsidies, equityrecapitalizations, and loans from the government, other SOEs, or state-owned financial institutions. The empirical setup uses a quasi-experimental setting that exploits the differential effects of a drastic fall in oil prices (in 2014–15) for SOEs in countries that experience a negative aggregate demand drop as a consequence of this shock versus similar firms in countries that do not suffer a downturn from the shock. The results—based on a balanced sample using coarsened exact matching and a differences-in-differences estimation—show that SOEs in treatment countries received increases in fiscal injections as a percent of average assets of 3.5 percent the year after the shock, equivalent to a significant recapitalization. These fiscal injectionstake the form of equity injections and loans from the government and state-owned financial enterprises. The results also show that capital expenditure as a percent of average assets in fully owned infrastructure SOEs decreased by 3.5 percentage points the year after the negative shock, equivalent to a 40 percent decline relative to its average level. In sum, this paper presents evidence that shows that SOEs do not serve as countercyclical tools and, in fact, may amplify negative shocks by requiring fiscal injections for SOEs while also lowering their capital expenditures.

Ratio Type Estimator for Balanced Sampling Plan excluding Adjacent Units

The study variables are usually correlated with the auxiliary variables and therefore their correlation could easily be computed. In this paper, the correlation coefficient is used for the estimation procedure, and therefore, we proposed a Horvitz-Thompson ratio-type estimator using correlation coefficient for Balanced Sampling plan excluding Adjacent units (BSA plan). It has been illustrated theoretically and empirically that the proposed Horvitz-Thompson ratio-type estimator is more precise than the Horvitz-Thompson estimator based on BSA plan. The proposed estimator provides an opportunity to utilise the auxiliary information for the estimation of population mean for BSA plan and useful for many real-life experiments. International Journal of Statistical Sciences, Vol.24(1), March, 2024, pp 103-114

Comparative Analysis of the Determinants of Entrepreneurial Activities in the Middle East and Latin America

This study aims to contribute to advancing the understanding of entrepreneurial ecosystems, Latin American development, and the evolution and future perspectives of the Middle East. It used a cross-sectional research design and quantitative data with 750 observations from the Global Entrepreneurship Monitor, specifically the National Experts Survey and the Adult Population Survey. The results were analyzed by the Partial Least Squares Technique (PLS-SEM) by grouping countries into two balanced samples of underexplored regions: the Middle East and Latin America. The two regions, Latin America and the Middle East, have diverse entrepreneurial ecosystems; each condition impacts entrepreneurial activities in different ways. In Latin America, the most significant variable is “Physical Infrastructure”, while in the Middle East, the most significant determinants are “Commercial and Professional Infrastructure” and “Entrepreneurship Culture”. This study shows that to support entrepreneurial activities, each region requires different settings for their entrepreneurial ecosystems. It contributes to the literature on regional entrepreneurial ecosystems and to less explored regions to advance our understanding, which will drive better policies and actions.

Fine-Grained Prototypes Distillation for Few-Shot Object Detection

Few-shot object detection (FSOD) aims at extending a generic detector for novel object detection with only a few training examples. It attracts great concerns recently due to the practical meanings. Meta-learning has been demonstrated to be an effective paradigm for this task. In general, methods based on meta-learning employ an additional support branch to encode novel examples (a.k.a. support images) into class prototypes, which are then fused with query branch to facilitate the model prediction. However, the class-level prototypes are difficult to precisely generate, and they also lack detailed information, leading to instability in performance. New methods are required to capture the distinctive local context for more robust novel object detection. To this end, we propose to distill the most representative support features into fine-grained prototypes. These prototypes are then assigned into query feature maps based on the matching results, modeling the detailed feature relations between two branches. This process is realized by our Fine-Grained Feature Aggregation (FFA) module. Moreover, in terms of high-level feature fusion, we propose Balanced Class-Agnostic Sampling (B-CAS) strategy and Non-Linear Fusion (NLF) module from differenct perspectives. They are complementary to each other and depict the high-level feature relations more effectively. Extensive experiments on PASCAL VOC and MS COCO benchmarks show that our method sets a new state-of-the-art performance in most settings. Our code is available at https://github.com/wangchen1801/FPD.