Sample Selection Method Research Articles

Mitigating data imbalance and noise: A divergence-based approach with enhanced sample selection

Label noise and distribution imbalance are common challenges in real-world datasets. To address noisy labels, sample selection methods are typically used. These methods identify small-loss data as correctly labeled and utilize only these clean samples for updating parameters during training. However, our thorough analysis shows that in scenarios with a long-tailed distribution, methods based on small-loss tend to favor samples from dominant classes, leading to a significant drop in performance. To tackle this issue, we introduce a divergence-based separation mechanism coupled with prototype distance for clean sample identification across each class, effectively mitigating the adverse effects of distribution imbalance on sample selection. Furthermore, drawing on recent theoretical insights that highlight the benefits of pseudo-labels in managing class imbalances, we enhance our approach by integrating semi-supervised and contrastive learning techniques. Comprehensive tests conducted on two synthetic noisy long-tailed datasets and one real-world dataset demonstrate the efficacy of our method in dealing with label noise and long-tailed distributions.

The Effect Of Market Segmentation And Product Distribution On Product Sales Volume PT. Garuda Mas Perkasa Medan

This research aims to determine and analyze the effect of market segmentation and product distribution on sales volume at PT Garuda Mas Perkasa Medan. The type of research in this study is quantitative with an associative approach. The population of this study were employees of PT Garuda Mas Perkasa Medan. The sample selection method uses saturated sampling, the number of samples used is 50 sample data. Before conducting research, first test the instrument with validity test and reliability test. Prerequisite test analysis includes normality test, multicollinearity test and heteroscedasticity test. The data analysis method uses multiple linear regression. The results of this study are in accordance with the first hypothesis, that market segmentation affects sales volume. This can be seen from the test results which show that the value of 3.915 is greater than 2.010 with a significance value of 0.000 less than 0.05. The results of this study are in accordance with the second hypothesis, that product distribution has an effect on sales volume. This can be seen from the test results which show that the value of 3.580 is greater than 2.010 with a significance value of 0.000 smaller than 0.05. The results of the study are in accordance with the third hypothesis, that market segmentation and product distribution simultaneously affect sales volume. This can be seen from the F test which shows a significance value smaller than 0.05, which is 0.000. Then Fcount> Ftable (26.845> 3.190).

The Effect of Compensation, Work Environment, and Career Development on Employee Turnover Intention of PT XYZ Final Assembly Department

This research aims to analyze the effect of compensation, work environment, and career development on turnover intention among employees in the Final Assembly Department at PT XYZ. Involving 87 respondents who are employees of the department, this research employs a saturated sampling method for sample selection. Data collection is conducted through the distribution of questionnaires using Google Form. Analysis is performed using Multiple Linear Regression Analysis, with t and f statistical tests to test hypotheses. The results indicate that compensation, work environment, and career development partially have a significant negative effect on employees' intention to leave their jobs. Additionally, simultaneously, these three factors also have a significant impact on turnover intention. This study provides a deeper understanding of the factors affecting turnover intention among employees in the Final Assembly Department PT XYZ, and it offers important implications for human resource management in efforts to retain competent and quality employees.

Dynamic selection for reconstructing instance-dependent noisy labels

As an inevitable issue in annotating large-scale datasets, instance-dependent label noise (IDN) can cause serious overfitting in neural networks. To combat IDN, label reconstruction methods have been developed with noise transition matrices or DNNs to simulate the transition from clean labels to noisy labels. Nevertheless, the absence of correct supervisions will lead to learning wrong noise transitions. This motivates us to select samples with clean labels to fetch the correct supervisions. However, the difficulty in obtaining prior knowledge of the noise rate prohibits the use of existing sample selection methods. To this end, we propose a dynamic sample selection method, namely Identity Mapping (IdMap), to overcome this limitation. Inspired by the feature-dependent characteristic of IDN, we first introduce the extracted instance features and pseudo-ground-truth labels to reconstruct noisy labels. A partial identity mapping between two labels is then established and samples with consistent identity mapping output are selected as clean data to update the classifier. Extensive experiments on both artificial and real-world noisy datasets demonstrate the superiority of IdMap compared with other state-of-the-art methods.

The influence of sales growth and capital intensity on tax aggressiveness with inventory intensity as a moderating variabel

This research aims to empirically prove the influence of sales growth and capital intensity on tax aggressiveness, with inventory intensity as a moderating variable. The independent variables used in this research are sales growth and capital intensity, while the dependent variable is tax aggressiveness, and inventory intensity serves as the moderating variable. The population of this research consists of Consumers Non-Cylicals sector companies listed on the Indonesia Stock Exchange from 2018 to 2022. The sample selection method is purposive sampling, yielding 12 companies or 60 observation samples. The data used are secondary data in the form of audited annual financial statements of each company. This research use Quantitative data analysis methods and associative problem formulation methods with causal relationships are used with the aid of E-views 12 software. The results indicate that sales growth and capital intensity simultaneously effect on tax aggressiveness. Partially, sales growth has a effect on tax aggressiveness, while capital intensity does not have a effect on tax aggressiviness. Additionally, the results indicate that inventory intensity does not moderate (weaken) the effect of sales growth on tax aggressiveness, nor does it moderate (weaken) the effect of capital intensity on tax aggressiveness.

Analysis of the influence of CAR, NPF, FDR AND BOPO on the financial performance (ROA) of banking in Indonesia

This article analyzes the influence of CAR, NPF, FDR and BOPO on banking financial performance (ROA) in Indonesia, which is the aim of this research. The sample selection method uses a purposive sampling technique at Sharia Commercial Banks registered with Bank Indonesia (BI) for the 2017-2022 period. The sample in this study consisted of 10 companies with a total sample data of 58. The data analysis technique used was multiple linear regression with the help of SPSS version 25. The results of the study showed that the NPF variable had an effect on financial performance (ROA), while CAR, FDR and BOPO did not. influence on financial performance (ROA). The selection of research objects at Sharia Commercial Banks resulted in limitations in the number of samples in this research. Apart from that, the independent variables in this study can only explain 19.8% of the independent variables.

Mask-Guided Vision Transformer for Few-Shot Learning.

Learning with little data is challenging but often inevitable in various application scenarios where the labeled data are limited and costly. Recently, few-shot learning (FSL) gained increasing attention because of its generalizability of prior knowledge to new tasks that contain only a few samples. However, for data-intensive models such as vision transformer (ViT), current fine-tuning-based FSL approaches are inefficient in knowledge generalization and, thus, degenerate the downstream task performances. In this article, we propose a novel mask-guided ViT (MG-ViT) to achieve an effective and efficient FSL on the ViT model. The key idea is to apply a mask on image patches to screen out the task-irrelevant ones and to guide the ViT focusing on task-relevant and discriminative patches during FSL. Particularly, MG-ViT only introduces an additional mask operation and a residual connection, enabling the inheritance of parameters from pretrained ViT without any other cost. To optimally select representative few-shot samples, we also include an active learning-based sample selection method to further improve the generalizability of MG-ViT-based FSL. We evaluate the proposed MG-ViT on classification, object detection, and segmentation tasks using gradient-weighted class activation mapping (Grad-CAM) to generate masks. The experimental results show that the MG-ViT model significantly improves the performance and efficiency compared with general fine-tuning-based ViT and ResNet models, providing novel insights and a concrete approach toward generalizing data-intensive and large-scale deep learning models for FSL.

Corporate Sustainability Performance: The Role of the Gender Diversity of the Board and Enterprise Risk Management

The research objective is to investigate the impact of board gender diversity on corporate sustainability performance, with enterprise risk management (ERM) as a moderating factor within health sector companies listed on the Indonesia Stock Exchange. The study adopts a secondary quantitative research approach and utilizes panel data linear regression analysis conducted through Econometric Views (E-Views). The study population consists of health sector firms listed on the Indonesia Stock Exchange from 2019 to 2022. The sample selection method used is purposive sampling, resulting in a sample size of 10 companies for the study. Secondary data is collected through the documentation method from the Indonesia Stock Exchange and the official websites of the selected companies. The findings reveal a negative relationship between board gender diversity and corporate sustainability performance, with a negative coefficient indicating that higher gender diversity among board members correlates with lower sustainability performance. Additionally, the study demonstrates that ERM strengthens the association between gender diversity and sustainability performance. Companies proficient in ERM implementation are better positioned to leverage the positive effects of gender diversity on sustainability performance. Understanding this interplay can assist organizations in making informed decisions regarding gender diversity policies and enterprise risk management strategies.

The Effect of ESG Risk Ratings, Board Size and Gender Diversity on Financial Performance: Econometric Case Study Indonesia 90 Companies 2020-2023

Aims: This study examines the influence of ESG Risk Ratings and the size of corporate boards on financial results. Furthermore, It investigates the impact of gender diversity on ESG Risk Ratings, Board Size, and the subsequent impact on Financial Performance. Study Design: In this case study, CFP is variable dependent, ESG and BS is variable independent, GD is the moderating variable, while Dar, LV, GR, and SZ are control variables. Place and Duration of Study: This study employs panel data from enterprises that meet the purposive sample requirements for the period of 2020-2023. Methodology: This study employs EViews 12 statistical software to conduct various tests, including the Descriptive Statistical Test, Model Selection Test, Multicollinearity and Heteroscedasticity Test, and Panel Data Regression Test. This study used purposive sampling method for sample selection. FP evaluated by the ratio of net income to total assets (ROA), ESG obtained from Morningstar Sustainalytics, BS determined by the total number of directors who serve on it, and GD is assessed by calculating the proportion of female board members. Result: The results of this inquiry indicate that ESG does not have an impact on FP, as demonstrated by the β = 0.0031 and Prob. = 0.3189 value. A different one independent variable, BS, also has no effect on FP, as evidenced by the value of β = 0.0096 and Prob. = 0.2130. As we transition to the moderation variable, it is evident that the GD variable does not moderate the relationship between ESG and FP (β = -0.011, Prob. = 0.3555). Meanwhile, the relationship between BS and FP is weakened by GD (β = -0.058, Prob. = 0.026). Conclusion: This study conclude that financial performance doesn’t influenced by ESG risk ratings and Board size. Beside, gender diversity weakens relationship between ESG and FP. The findings of this study have significant consequences for corporations, as they shed light on the elements that influence a company's financial performance, particularly with regard to sustainability standards. Further research can include variables such as comparison female directors and others.

The Effect of Return on Equity and Current Ratio on Price to Book Value (Case Study of PT Unilever Indonesia Tbk which is Listed on the Indonesia Stock Exchange for the 2015-2022 Period)

This study aims to find out how Price to Book Value is influenced by Return on Equity and Current Ratio. The focus of the research is PT Unilever Indonesia Tbk, which is listed on the Indonesia Stock Exchange. The sample for this research is the financial statements of PT Unilever Indonesia Tbk from 2015 to 2022. An impossible sample selection method was used. Instead, data analysis is carried out using several techniques. This includes classical assumption tests, correlation tests, multiple linear regression coefficient of determination tests, and partial and simultaneous hypothesis tests. The research results show that Return on Equity (X1) and Current Ratio (X2) do not partially influence Price to Book Value. Apart from that, Return on Equity does not partially affect Price to Book Value (Y) significantly at all

The Effect of Tax Planning on Company Value in Manufacturing Companies in the Food and Beverage Sector Listed on the IDX

This study aims to determine the influence of Tax Planning on Company Value. Tax Planning is used as an independent variable and Company Value as a dependent variable. The population in this study is food and beverage companies listed on the Indonesia Stock Exchange in 2018-2012. The sample consists of 6 companies with a period of 5 years. namely: Wilmar Cahaya Indonesia Tbk, Nippon Indosari Corporindo Tbk, Sekar Bumi Tbk, Sekar Laut Tbk, Siantar Top Tbk, and Ultrajaya Milk Industry and Trading Company, Tbk. The sample selection method is purposive sampling. The type of data used is quantitative data and data sources using secondary data. The analysis method used to test the hypothesis is to use simple linear regression analysis. The results show that tax planning does not have a significant effect on the Company's Value.

EFFECT OF SCHOOL FEEDING PROGRAMME PRACTICES ON STUDENTS ACADEMIC PERFORMANCE IN MATHEMATICS SUBJECT IN PUBLIC PRIMARY SCHOOLS IN RWANDA

The purpose of this study is to evaluate how school feeding programs impact academic performance in mathematics among students in public primary schools within Nyanza District. Its objectives were twofold: firstly, to assess the effectiveness of these feeding programs secondly, to determine the relationship between the feeding program and students academic achievements. The research employed a descriptive research design, utilizing both questionnaires and interviews to gather data. A sample size of 173 participants was determined using the Yamane formula, consisting of 155 teachers and 18 head teachers selected via simple random sampling and purposeful selection methods. Data collection involved administering teacher surveys and conducting interviews with head teachers. The gathered quantitative data were analyzed using IBM/SPSS, focusing on percentages and correlation coefficients. Key findings indicated mixed perceptions among respondents regarding the adequacy of food and cleanliness of water provided by the feeding programs. Approximately half of the respondents believed their students performed well in national mathematics exams and learned effectively after lunch, while a majority felt that learners were motivated during teaching activities. The study found a moderate positive correlation (coefficient of determination = 0.809) between aspects of the school feeding program and students academic performance in mathematics. Limitations of the study included potential biases in self-reported data and the narrow geographical focus on Nyanza District, limiting generalizability to other regions. Implications suggest that improving the quality and consistency of school feeding programs could positively influence academic outcomes in mathematics.

COVID-19: really weakening the effect of financial performance on corporate value of consumer cyclicals?

Covid-19 is a disease outbreak that affects company performance and value. This study aims to test whether Covid-19 can strengthen the effect of financial performance on firm value. Financial performance is proxied by financial ratios consisting of the Current Ratio (CR), Debt to Asset Ratio (DAR), and Return on Assets (ROA). Meanwhile, the Company Value is proxied by Tobin” Q. The Cyclical Consumer Sector is the population of this study. Purposive Sampling is the sample selection method used. Twenty-eight companies meet the criteria. The data processed is quarterly from 2017 to 2021, so the number of observations is four hundred and twenty. The study results show that only DAR and ROA affect firm value and that Covid-19 is a quasi-moderating and strengthening effect of ROA on firm value. The novelty of this study is that Covid-19 is treated not as an independent variable but as a moderating variable, and the data is quarterly. The results of this study have implications for companies to be able to anticipate the arrival of disease outbreaks because it dramatically affects the performance and value of the company. The influence of this plague can be debilitating or strengthening.

Challenges faced by learners who are survivors of school violence: implications for support structures

The main aim of this study is to explore challenges faced by learners who are survivors of school violence and the implications for support structures. Many schools are branded with violence, gangsterism, substance abuse and general crime. The scourge of violence in schools is cause for concern, citing daily reports that appear in the written and electronic media about high levels of violence, physical and sexual abuse, and gang-related activities in schools. Qualitative research design and phenomenology as a mode of inquiry were employed in this study. School principals' and teachers' experiences and perceptions were elicited through focus-group interviews, document reviews, and observations. The population comprised principals and teachers in five schools. Purposive sampling was used as a method of sample selection. The findings revealed that support is key, and lack of support can negatively impact learners’ access to schools and academic achievement in many ways. It was also found that most learners who leave school prematurely, through the fear of violence, do not report when they leave. The study recommended that support must be provided promptly, and parents, teachers, and caregivers must be capacitated to offer effective and efficient support.

Model of the Relationship of Religious, Image, Trust, and Satisfaction towards Member Loyalty in BMT in Jambi, Indonesia

This research aims to evaluate the influence of Religiosity, Image, Trust, Satisfaction, and Loyalty of Baitul Maal wat Tamwil (BMT) members in Jambi, Indonesia, both with and without mediation by the variables image, trust, and satisfaction. Considering that the development of BMT in Jambi is slower than in other provinces in Sumatra, this research was conducted in Jambi, Indonesia. The research design uses quantitative methods with a survey approach to collect primary data through questionnaires distributed to 100 BMT members in Jambi, Indonesia, using a purposive sampling method for sample selection. Quantitative data was then analyzed using the Structural Equation Model (SEM) with the help of SPSS 22 and SmartPLS software. The research results showed that Religiosity significantly influenced BMT's Image, Trust, and Member Satisfaction in Jambi, Indonesia. The image does not affect member loyalty to BMT, but Trust and Satisfaction significantly influence member loyalty to BMT in Jambi, Indonesia. This research is expected to provide theoretical contributions in the field of Sharia marketing management and consumer behavior at BMT, as well as practical contributions for BMT management and related parties as information for determining marketing strategies for BMT products in Jambi, Indonesia to increase satisfaction and loyalty of BMT members in Jambi, Indonesia

Peran Moderasi Ukuran Perusahaan pada Pengaruh Intellectual Capital terhadap Nilai Perusahaan

This research aims to determine the role of intellectual capital, especially when paying attention to the element of company size on company value. The analysis uses panel data regression. The sample selection method uses purposive sampling on non-financial companies listed on IDX-IC (Indonesian Stock Exchange) for the 2019-2022 period. The research results show that, of the three proxies used, SCE and CEE have a significant effect on company value, while the role of SIZE is only for SCE and CEE. HCE still does not play a role in determining market value, either directly or when moderated by SIZE. These findings prove that HCE is not always the main priority in determining company value, but other elements are considered more important by companies.

Exploring the Factors that Affect Creativity among Adolescents

Aims: Ideas are the key to innovation. In this competitive era the excellence of the mankind depends on their innovation. Innovative outcomes laid stress on how to think divergently. Creativity is varied in nature. Several factors are responsible to perform any creative task. Therefore, it is necessary to find out the most important factor which can help to expedite creativity among the person, more specifically to explore the scientific reason behind the specific factor and multivariate relationship between two or more factors. Study Design: In the present study the investigator used explanatory-case study method. Methodology: The present research included under non-probability sampling technique. For conducting the present research work systematically researcher used purposive sampling method for sample selection. It is basically selected for solving several purposes properly. Therefore the willing and accessible 31 subjects are chosen for this purpose. Out of 31 subjects, 20 students arefrom Bengali Medium School and remaining 11students are from English Medium School. These schools are affiliated to West Bengal Board of Secondary Education and CBSE Board. Results: A scatter analysis is used to compare two data sets for finding out the relationship between them. Simultaneously, curve fitting algorithm is used to examine the relationship between one or more predictors (independent variable) and a response variable (dependent variable) with the goal of defining a best fit model of the relationship. The major aim of this study is to find out the relationship between one or more predictors (independent variable) and a response variable (dependent variable) with the goal of defining a best fit model of the relationship. Conclusion: Exploring the effective factors of creativity is addressed on 14 years age of boys and girls.

THE DETERMINANTS OF CASH HOLDING

The aim of this research was to empirically prove the influence of profitability, leverage, sales growth, and capital expenditure on cash holding. The sample selection method used purposive sampling. The research sample was 63 primary consumer sector companies listed on the Indonesia Stock Exchange with a research period of three years and were processed by multiple regression analysis of panel data using the Eviews. The research results showed that there was only one variable, namely leverage, which had a negative impact on cash holding. Profitability did not have a positive impact on cash holding, while sales growth and capital expenditure did not have a negative impact on cash holding.

Pengaruh Komite Audir, Kompleksitas Perusahaan Dan Ukuran Perusahaan Terhadap Audit Delay

The aim of this research is to empirically prove the influence of the audit committee, company complexity and company size on audit delay. The population in the research is all food and beverage companies listed on the Indonesia Stock Exchange 2020-2022. The sample selection method chosen was the purposive sampling method so that 78 observations were obtained. The data analysis technique used is multiple linear regression analysis using the SPSS application. The research results show that the audit committee and company complexity have a significant positive effect on audit delay, while company size has a significant negative effect on audit delay.

Investigation of The Factors Affecting the Vaccine Preferences of Pre-Hospital Emergency Healthcare Professionals

Objective: The aim of this study was to examine the factors affecting the vaccine preferences of prehospital emergency health workers and the status of vaccine opposition. Materials and Methods: This study was planned as a descriptive study to examine the factors affecting the vaccine preferences of prehospital emergency health care workers working in 112 emergency health services stations and command and control centers affiliated to the Ministry of Health in Aydın province between April 2022 and January 2023, their anti-vaccine status and their attitudes towards Covid-19 vaccine. The population of the study consisted of physicians, emergency medical technicians/technicians (ATT/Paramedics) and ambulance drivers (drivers) working in 112 emergency health services stations and command control centers affiliated to the Ministry of Health in Aydın province (N: 577). The sample selection method was not used and it was aimed to reach the whole population. Data were collected by sending online questionnaire forms to social media groups. The Case Report Form consisting of 12 (twelve) questions including socio-demographic characteristics such as age, gender, educational status, Attitudes Towards Covid-19 Vaccine Scale and Anti-Vaccine Scale were used. The data obtained in the study were analyzed using the SPSS for Windows 25.0 program. Findings: The population of the study was 577 people. 89.7% (n:427) people participated in our study. 51.5% (n:220) of the participants were female. 48.9% (n:209) were found to have Covid-19 infection. 95.8% (n:427) of the participants were vaccinated. Among the vaccinated participants, 70.4% (n:288) received Biontech, 70.4% (n:288) received Sinovac and 5.1% (n:21) received Turkovac vaccine types. When the scores of the attitudes towards Covid-19 vaccine scale and its dimensions were compared according to the descriptive characteristics of the participants, it was found that the positive attitude dimension scores of male participants (3.81) were higher than those of female participants (3.51). It was determined that the positive attitude dimension scores of the participants with chronic diseases and the participants who were vaccinated, and the positive attitude dimension scores of the participants whose vaccine type was Biontech, Sinovac and Turkovak were higher than those of the participants who were not vaccinated. It was determined that the positive attitude dimension and attitudes towards Covid-19 vaccine scale scores of the participants whose vaccine dose was three doses and more were higher than the participants whose vaccine dose was one dose and two doses, and the positive attitude dimension and attitudes towards Covid-19 vaccine scale scores of the participants who had a death due to Covid-19 infection in the family or close environment were higher than those of the participants who did not. According to the findings of our study, the view that there is no vaccine protection in those with chronic diseases has a higher score. Participants who did not have Covid-19 infection, who were not vaccinated, and who had no death loss in their close environment had higher scores on the anti-vaccination scale. Conclusion: In this study, which examined the hesitancy of prehospital emergency healthcare workers about Covid-19 vaccination, it was found that Covid-19 vaccines were administered at a very high rate. However, it was also found that 4.2% of the employees had high hesitation about vaccines and were not vaccinated. Pre-hospital emergency healthcare workers have an important role and responsibility in the acceptance of Covid-19 vaccine by the public and other healthcare professionals, as in other disaster situations. In a pandemic, vaccination of all members of the society is necessary for the control of the pandemic. For this reason, scientific studies should be conducted for emergency healthcare workers who are not vaccinated and who are hesitant about vaccines, including basic concerns about vaccines and examining potential side effects of vaccines. Training programs should be organized to ensure that healthcare workers are vaccinated with existing Covid-19 vaccines.