Emerging Science Journal (ESJ)
Not a member yet
960 research outputs found
Sort by
Vision Transformer Embedded Feature Fusion Model with Pre-Trained Transformers for Keratoconus Disease Classification
Full text linkKeratoconus is a progressive eye disorder that, if undetected, can lead to severe visual impairment or blindness, necessitating early and accurate diagnosis. The primary objective of this research is to develop a feature fusion hybrid deep learning framework that integrates pretrained Convolutional Neural Networks (CNNs) with Vision Transformers (ViTs) for the automated classification of keratoconus into three distinct categories: Keratoconus, Normal, and Suspect. The dataset employed in this study is sourced from a widely recognized and publicly available online repository. Prior to model development, comprehensive preprocessing techniques were applied, including the removal of low-quality samples, image resizing, rescaling, and data augmentation. The dataset was subsequently partitioned into training, testing, and validation subsets to facilitate robust model training and performance evaluation. Eight state-of-the-art CNN architectures, including DenseNet121, EfficientNetB0, InceptionResNetV2, InceptionV3, MobileNetV2, ResNet50, VGG16, and VGG19, were utilized for feature extraction, while the ViT served as the classification head, leveraging its global attention mechanism for enhanced contextual learning, achieving near-perfect accuracy (e.g., DenseNet121+ViT: 99.28%). This study underscores the potential of hybrid CNN-ViT architectures to revolutionize keratoconus diagnosis, offering scalable, accurate, and efficient solutions to overcome limitations of traditional diagnostic methods while paving the way for broader applications in medical imaging. Doi: 10.28991/ESJ-2025-09-02-027 Full Text: PD
Driving Social Entrepreneurship Among Students: Investigating Through PLS-SEM and fsQCA Approaches in Emerging Economies
No full textThis study aims to identify the relationship between social self-efficacy, social innovation, resilience, and proactive personality concerning university students’ behavioral intention to engage in social entrepreneurship, particularly in emerging economies, like Bangladesh. A structured questionnaire was utilized to collect quantitative data from 540 students in various disciplines of study as part of the study's quantitative research methodology using partial least squares-Structural Equation Modelling (PLS-SEM) and fuzzy-set Qualitative Comparative Analysis (fsQCA). The analysis reveals that proactive personality traits are associated with the social entrepreneurship intention (SEI) and that leadership orientation is also significant to SEI. The study also demonstrates that social entrepreneurial activities tend toward higher social self-efficacy and resilience, making it crucial to focus on such characteristics while facing social risk and bearing innovations. This study's novelty lies in its focus on the unique combination of psychological traits—social self-efficacy, social innovation, resilience, and proactive personality—and their impact on university students' intention to engage in social entrepreneurship in emerging economies. Additionally, the research emphasizes the importance of integrating leadership skills and social innovation into academic curricula and policy development to foster social entrepreneurship. Practical implications indicate that leadership skills and social innovation should be included in the curricula of educational institutions, and supportive policies should be developed to create available resources for prospective social entrepreneurs
Empirical Analysis of Executive Capital, Innovation, and Risk-Taking in A-Share Tech Firms
Full text linkThis research aims to explore the impact of executive human capital (SMHC) on the performance of Chinese A-share technology listed companies, with a focus on the mediating roles of technological innovation and risk-taking. Using 13,733 data points from 2,796 A-share technology listed companies from 2014 to 2022 sourced from the CSMAR Database, the OLS regression method was employed for analysis. The research findings indicate that SMHC, including its stock, flow, and investment, significantly improves enterprise performance. Among them, investment has the most significant impact, enhancing both economic benefits and market value. Moreover, technological innovation and risk-taking play mediating roles, with positive and significant coefficients. This research enriches the understanding of the relationship between SMHC, technological innovation, risk-taking, and enterprise performance, providing new insights for enterprises to optimize their human capital management and enhance competitiveness
Process-Based Management in Public Administration: A Bibliometric Analysis
No full textThis bibliometric study aimed to analyze the trends and scope of scientific papers on process-based management in public administration worldwide between 2010 and 2024. A methodology based on the Scopus search engine was used to collect the data, which were selected according to specific criteria; thus, 29 documents were obtained. For the analysis, Microsoft Excel was used to produce graphs and tables, and VOSviewer software was used to create co-occurrence maps of countries, authors, and keywords. The results indicated that the United States and Brazil were the most productive countries in this area. Most papers were published as conference papers in areas such as computer science, business management, engineering, and social sciences, while scientific papers were the most cited. It was identified that process-based management has been applied in various sectors, such as healthcare institutions, educational institutions, food companies, and agricultural companies, among others. It was concluded that the reviewed papers highlight the obstacles derived from the organizational and operational complexity in public sector institutions, which limits the implementation of process-based management. This underlines the need for further research, especially for comparisons with private sector organizations
Odor Profiling of Blood Shells Using TGS Gas Sensor and PCA-SVM Analysis
Full text linkBlood cockles (Andara granulosa) are among the most popular animal protein sources due to their rich nutritional content and high economic value. The storage period and temperature are two critical factors that significantly influence the freshness of blood cockles. One key indicator of blood cockle quality is the odor they emit. An unpleasant or inappropriate odor can indicate contamination or a decline in quality, posing potential food safety risks. However, conventional methods of odor quality testing are often subjective, require specialized skills, and may not always be reliable. To address the limitations of human olfaction, advancements in gas sensor technology, specifically gas array sensors (also known as the electronic nose), have been developed. This research aims to profile the freshness of blood cockles by identifying their odor under different storage conditions using electronic nose technology. The study used fresh blood cockle meat, which was stored under varying temperature conditions: at room temperature, in a cooler, and in a freezer. The storage periods for the samples were 1, 2, 3, 4, and 5 days. The samples were placed in sealed bottles and tested using a gas array sensor. The data collected from this process were in the form of voltage readings, which were analyzed using machine learning techniques, specifically Principal Component Analysis (PCA). The data were then classified using a Support Vector Machine (SVM) model. The study results showed that the gas array sensor successfully classified the odor profiles, with PCA explaining 93.83% of the variance in the data. The SVM model achieved an accuracy of 89.66% for PCA-reduced data and 91.44% for non-PCA data
Music Education with Artificial Intelligence for Inclusive and Sustainable Early Childhood Learning
No full textThis study aims to evaluate the impact of a didactic strategy that incorporates generative artificial intelligence (AI) into music education, supporting oral language development in preschool children and promoting inclusive and sustainable early childhood learning. Using an action-research approach, a mixed-methods design was applied to assess the performance of 15 children aged 3 to 6 years, divided into experimental and control groups. The experimental group participated in AI-supported activities using tools such as Genially, Educaplay, and Wordwall, whereas the control group employed traditional methods. Quantitative data from pre-and post-tests, as well as qualitative observations, revealed that AI-enhanced sessions improved motivation, pronunciation, and engagement, particularly among children aged 5 and 6 years old. Although statistical tests showed no significant differences between groups, the intervention demonstrated pedagogical effectiveness by increasing interest and participation. The novelty of this work lies in applying generative AI in early music education to personalize learning and reduce inequality, aligning with several Sustainable Development Goals (SDGs 3, 4, 9, and 10). The findings offer valuable insights into designing inclusive educational experiences through the integration of ICT and AI, highlighting the need to enhance teacher training in emerging digital pedagogies and promote accessible music-based learning in diverse educational settings
Unlocking Potential Score Insights of Sentimental Analysis with a Deep Learning Revolutionizes
Full text linkOnline hate has emerged as a rapidly growing issue worldwide, often stemming from differences in opinion. It is crucial to use appropriate language and words on social media platforms, as inappropriate communication can negatively impact others. Consequently, detecting hate speech is of significant importance. While manual methods are commonly employed to identify hate and offensive content on social media, they are time-consuming, labor-intensive, and prone to errors. Therefore, AI-based approaches are increasingly being adopted for the effective classification of hate and offensive speech. The proposed model incorporates various text preprocessing techniques, such as removing extraneous elements like URLs, emojis, and blank spaces. Following preprocessing, tokenization is applied to break down the text into smaller components or tokens. The tokenization technique utilized in this study is TF-IDF (Term Frequency–Inverse Document Frequency). After tokenization, the model performs the classification of hate and offensive speech using the proposed BiLSTM-based SM-CJ (Scalable Multi-Channel Joint) framework. The BiLSTM-based SM-CJ model is particularly effective in detecting hate, offensive, and neutral tweets due to its ability to capture both forward and backward contexts within a given text. Detecting hate speech requires a comprehensive understanding of the text and the identification of patterns spanning across multiple words or phrases. To achieve this, the LSTM component of the BiLSTM model is designed to capture long-term dependencies by utilizing information from earlier parts of the text. The proposed SM-CJ framework further aligns the input sequence lengths fetched from the input layer, enabling the model to focus on specific segments of the input sequence that are most relevant for hate speech detection. This approach allows the model to accurately capture derogatory language, and subtle nuances present in hate speech. Finally, the performance of the proposed framework is evaluated using various metrics, including accuracy, recall, F1-score, and precision. The results are compared with state-of-the-art approaches, demonstrating the effectiveness of the proposed model. Doi: 10.28991/ESJ-2025-09-01-03 Full Text: PD
Crop Monitoring System Using IoT, Solar Energy and Decision Tree Algorithm
Full text linkPeru's diverse topographical regions offer optimal conditions for agriculture, but a lack of technology hinders efficiency, leading to food imports despite the country's potential. This paper aims to design an Internet of Things-based monitoring system where the specific objectives are focused on building a solar-powered power stage and integrating machine learning algorithms to help determine crop health. The development methodology includes the evaluation of the use of sensors to measure environmental and soil temperature and humidity, precipitation and hydrogen potential to help identify the health status of crops using machine learning algorithms (decision trees) and transmit the information to a Blynk real-time visualization server. The system components include a device based on an ESP32 module operating in low-power mode, a solar power stage, a data management stage with Blynk with Wi-Fi communication. The results show that the IoT device was adapted for outdoor environments protected by an IP65 housing and can operate for approximately 12 days with a 3000 mAh battery. The main result is that the Random Forest model stands out for having a 98% accuracy when inferring the state of crop conditions. Future improvements can include more efficient solar cells to improve the system's charging conditions. Doi: 10.28991/ESJ-2025-09-02-06 Full Text: PD
Impacting Information Technology and Telecommunications Infrastructure on the Digital Economy
Full text linkDigital technologies such as Artificial Intelligence, Big Data, and the Internet of Things are developing at breakneck speed, creating a solid wave of digital transformation globally. The digital economy is increasingly asserting its role as a new growth driver. Therefore, the study's objective is to explore the critical factors influencing the digital economy and propose policy recommendations for enhancing the digital economy. The study applies quantitative research methods, mainly through actual data surveying of economic experts, to evaluate factors affecting the digital economy. The authors used the structural equation model to measure the impact of factors on the digital economy in five big cities in Vietnam. The data collection strategy involves direct interviews via a structured questionnaire with a sample size of 800 economic experts and analysis using SPSS version 20.0 and Amos software. The finding identifies eight critical factors influencing the digital economy at a significant level of 0.01 and eight accepted hypotheses. The study's novelty contribution highlights the considerable influence of information technology and telecommunications infrastructure on Vietnam's digital economy. Finally, the authors proposed policy recommendations to enhance the digital economy; moreover, this model can be used in each locality or the whole country in researching and evaluating the impact of factors on the digital economy. Moreover, digital transformation has become a global trend, and digital recommendations will be applied to better manage environmental issues toward green growth and sustainable development. Doi: 10.28991/ESJ-2025-09-02-026 Full Text: PD
Impact of Social Networks on Entrepreneurial Innovation and Business Performance in SMEs
No full textThis study utilized the panel data of Small and Medium Enterprises Surveys (SMES: 2009-2015) and used the two-way fixed effects model to assess the relative association between social networks and entrepreneurial innovation and business performance in Vietnam. The results supported the hypotheses that social networks are positively associated with the likelihood of entrepreneurial innovation and business performance from the perspective of micro, small, and medium enterprises. A 1% increase in social networks was associated with an increase of 2.8 percentage points in incremental product innovation and 1.5 percentage points in process innovation. The more extensive social networks also helped enterprises perform their business better by increasing their real revenue, real profit, and rate of return on assets. In addition to contributing to the well documented literature of the role of social networks in entrepreneurial innovation and business performance, the novelty of the research was highlighting the relationship varied by ownership, organizational structure, and location. The local business association membership was also matter for entrepreneurial innovation and firm performance in the Vietnamese SMEs. The results were, therefore, informative for policymakers and SME entrepreneurs.
JEL Classification: O31, L25, L11