Outlier detection in multivariate data is a critical challenge with applications in fields such as finance, medicine, and industrial process monitoring. This study investigates the Data-Driven Cluster Analysis Method, designed to improve the identification of atypical observations through adaptive clustering strategies. Specifically, the research examines the role of the stopping criterion k(max) -- the maximum number of clusters considered -- in determining the method's efficiency and accuracy. Using Monte Carlo simulations with contaminated normal, exponential, and point mass distributions, the study evaluates whether excessively large k(max) values contribute meaningfully to model performance or merely increase computational cost. Results demonstrate that the optimal number of clusters, selected via the Bayesian Information Criterion (BIC), consistently falls well below the imposed k(max )threshold, regardless of dimensionality, or contamination level. Furthermore, as sample size increases, the gap between the selected k and the k(max) limit widens, while runtime grows proportionally. These findings suggest that overly conservative settings for k(max) are unnecessary and can be replaced by more parsimonious values without compromising detection accuracy. The study reinforces DDCAM's robustness and stability while highlighting opportunities for computational optimization.
This study investigates the relationship between participation in Scientific Initiation (SI) and admission and success rates in graduate programs at the Federal University of Vic & cedil;osa (UFV). The research was motivated by the need to understand the factors that influence access and academic continuity, especially in the Brazilian context, where graduate school dropout rates pose a challenge for both students and institutions. Using descriptive statistics and logistic regression techniques, data were obtained through the Federal University of Vic & cedil;osa's open data portal, and specific data were analyzed using R software. Variables such as age, gender, state of origin, and academic standing were examined, enabling the identification of patterns and trends in student trajectories. It was possible to draw an average profile of UFV students, both at the undergraduate and graduate levels, and to assess the relationship between scientific initiation and graduate studies. The results indicate that participation in scientific initiation has a positive association with admission to graduate studies. Students who participated in scientific initiation during their undergraduate years had a significantly higher rate of academic continuation compared to those who did not participate in this program. The results of this research underscore the importance of institutional policies that encourage scientific initiation, promoting greater academic engagement and contributing to a reduction in dropout rates in graduate programs. The study also contributes to the knowledge base on the relationship between scientific initiation and retention in graduate programs.
This study optimizes resource allocation in queueing network systems to improve operational efficiency using appropriate algorithms. Specifically, Particle Swarm Optimization (PSO) was applied to the buffer and server allocation problem (BSAP) in queueing networks with Markovian arrivals and service times, multiple servers, and finite buffers. In this context, the buffer and server allocation problem (BSAP) stands out, whose solution methodology can be applied to various real-world situations modeled as queues or queueing networks, such as manufacturing line systems, healthcare services, traffic models, among others. BSAP is computationally challenging as a nonlinear programming problem without a closed-form analytical solution, necessitating derivative-free methods like PSO. The PSO algorithm is considered a promising tool for finding efficient solutions, thus enabling improved resource management. The study examines the algorithm's ability to deliver cost-effective and suitable solutions that accommodate variations in relative costs between servers and buffers, as well as the specific characteristics of each network topology.
This article explores the application of Structural Equation Modeling (SEM) in the analysis of empirical education data, detailing its operationalization, result interpretation, and model quality assessment. SEM is a statistical technique that combines factor analysis and multiple regression, enabling the examination of dependency relationships between observed variables and latent constructs. The study uses data from the "Teaching Work in Basic Education in Brazil" research to analyze six constructs: preparation for career entry, activity control level, frequency of collaborative activities, classroom conditions, educational unit conditions, and professional satisfaction. The results show that preparation for career entry has the highest total effect on professional satisfaction, while the greatest direct effect is exerted by the activity control level construct. The article emphasizes the importance of theoretical support for defining items and associations in the model, as well as rigorous validation measures such as Cronbach's Alpha and fit indices (RMSEA, SRMR, CFI, TLI). The application to educational data demonstrates the potential of SEM to deepen the analysis of teaching work, highlighting its contributions to educational research and encouraging methodological advances and new research agendas in the field of Education.
This research study aims to statistically analyze the impact of social media usage on the mental well-being of employees. Social media plays a significant role in everyday life, impacting social, professional, and personal aspects. Despite many advantages, excessive use of social media may harm users' mental health. The study's objective is to investigate how employees' mental health, job satisfaction, stress levels, anxiety, and overall productivity are influenced by their use of social media in both public and private sector organizations. In this regard, the data of 410 respondents showed an age range of 21 to 31 years old (M = 26.25). For this research study, employees / skilled personnel/schoolteachers (Public and Private), banking sector personnel, hotel staff, and courier services organizations in Pakistan were nominated to collect the samples. The age group of employees was the focal point of the study, as most young workers are interested in using social media. Correlation, regression, NOVA, and coefficient analysis reveal a strong relationship between social media and the mental well-being of employees, which leads to testing of Hypotheses H1 to H3. Resultantly, the Beta value beta = 0.220 (p < .001), F-statistic = 20.672, and t-value of 4.547 were found statistically significant. The positive angles incorporate enhanced social connections, self-expression, and access to support organizations. Comparatively, the adverse results, for example, cyberbullying, social correlation, rest unsettling influences, and feelings of dread toward missing an opportunity, are likewise investigated.
