Latin American Journal of Computing
Not a member yet
191 research outputs found
Sort by
Hybrid CNN-Transformer Model for Severity Classification of Multi-organ Damage in Long COVID Patients
Get PDFGlobal COVID-19 spread has necessitated the use of rapid and accurate diagnostic procedures to support clinical decision-making, particularly in resource-limited environments. In this work, a hybrid deep model combining Convolutional Neural Networks (CNN) and Transformer architecture is proposed to diagnose COVIDx CXR-3 dataset chest X-ray images into three classes of severity levels: Mild, Moderate, and Severe. The methodology incorporates data preprocessing techniques such as resizing, normalization, augmentation, and SimpleITK organ segmentation. A DenseNet121-based CNN extracts local features, while global dependencies are extracted by a Vision Transformer. The features from both are fused and fed to a classification head to generate the predictions. The training was done in PyTorch with learning rate 0.0001, batch size 32 and optimized with Adam optimizer for 50 epochs. Performance measures like Accuracy, Precision, Recall, F1-Score, and Confusion Matrix were computed to measure performance. Results show that the CNN-transformer model which outperforms the CNN-only model that achieved 88%. This integration has demonstrated a better capability in severity classification and great potential in helping clinicians prioritize care, optimize treatment plans, and allocate resources, thereby improving outcomes in COVID-19 management. Global COVID-19 spread has necessitated the use of rapid and accurate diagnostic procedures to support clinical decision-making, particularly in resource-limited environments. In this work, a hybrid deep model combining Convolutional Neural Networks (CNN) and Transformer architecture is proposed to diagnose COVIDx CXR-3 dataset chest X-ray images into three classes of severity levels: Mild, Moderate, and Severe. Themethodology incorporates data preprocessing techniques such as resizing, normalization, augmentation, and SimpleITK organ segmentation. A DenseNet121-based CNN extracts local features, while global dependencies are extracted by a Vision Transformer. The features from both are fused and fed to a classification head togenerate the predictions. The training was done in PyTorch with learning rate 0.0001, batch size 32 and optimized with Adam optimizer for 50 epochs. Performance measures like Accuracy, Precision, Recall, F1-Score, and Confusion Matrix were computed to measure performance. Results show that the CNN-transformer model which outperforms the CNN-only model that achieved 88%.This integration has demonstrated a better capability in severity classification and great potential in helping clinicians prioritize care, optimize treatment plans, and allocate resources, thereby improving outcomes in COVID-19 management
Cloud Computing in Ecuadorian Higher Education: A Case Study on Use, Benefits, and Challenges at UTEQ
Get PDFDigital transformation continues to reshape higher education, with cloud computing emerging as a key enabler of enhanced accessibility, collaboration, and academic management. This study investigates the use of cloud computing in Ecuadorian universities by identifying its benefits, barriers, and opportunities through a survey of key stakeholders in the education system. A quantitative approach was employed using a structured questionnaire to collect data on participants’ knowledge levels, tools used, perceived advantages, challenges, and expectations. The main benefit identified was accessibility from any location (92%), followed by enhanced collaboration (73%) and the modernization of educational practices (43%). The primary challenges included lack of training (67%), limited connectivity (58%), associated costs (46%), and concerns about data security and privacy (34%). These findings underscore the need to strengthen technological infrastructure and provide targeted training to optimize the effective use of cloud computing. Regarding future perspectives, 71% of respondents advocated for greater integration into teaching and learning, while 64% suggested expanding its use across academic and administrative domains. Cloud computing represents a strategic asset for Ecuadorian higher education. However, its full adoption requires addressing infrastructure and capacity-building challenges through policies that promote collaboration, innovation, and the efficient management of institutional resources.Digital transformation continues to reshape higher education, with cloud computing emerging as a key enabler of enhanced accessibility, collaboration, and academic management. This study investigates the use of cloud computing in Ecuadorian universities by identifying its benefits, barriers, and opportunities through a survey of key stakeholders in the education system. A quantitative approach was employed using a structured questionnaire to collect data on participants’ knowledge levels, tools used, perceived advantages, challenges, and expectations. The main benefit identified was accessibility from any location (92%), followed by enhanced collaboration (73%) and the modernization of educational practices (43%). The primary challenges included lack of training (67%), limited connectivity (58%), associated costs (46%), and concerns about data security and privacy (34%). These findings underscore the need to strengthen technological infrastructure and provide targeted training to optimize the effective use of cloud computing. Regarding future perspectives, 71% of respondents advocated for greater integration into teaching and learning, while 64% suggested expanding its use across academic and administrative domains. Cloud computing represents a strategic asset for Ecuadorian higher education. However, its full adoption requires addressing infrastructure and capacity-building challenges through policies that promote collaboration, innovation, and the efficient management of institutional resources
Desarrollo de un prototipo de aplicación de escritorio para bancos de sangre
Get PDFIn this research, a prototype desktop application was developed to optimize the management of blood banks, with emphasis on inventory control and scheduling appointments for donations. The objective of the study was to develop an accessible software solution that automates administrative processes and contributes to a more efficient and secure management of blood resources. The methodology used was the agile Scrum framework, with the development organized in biweekly iterations. SQL Server was used as the database manager due to its robustness and security. Validation was carried out in a simulated environment representative of a blood bank, using scenarios designed to emulate real tasks in the management of donations. In addition, a usability evaluation was conducted with real users through structured interviews, who identified strengths and opportunities for improvement in the interface. The results showed a significant reduction in record-keeping errors, greater operational efficiency and improved interaction with donors. The usability evaluation revealed needs for refinements to the visual experience, proposals such as the incorporation of interactive tutorials and improvements to the data entry forms. It is concluded that the prototype satisfactorily fulfills the proposed objectives, representing an effective technological solution to strengthen transfusion safety and support clinical decision making in contexts of limited infrastructure.En esta investigación, se desarrolló un prototipo de aplicación de escritorio orientado a optimizar la gestión de bancos de sangre, con énfasis en el control de inventarios y la programación de citas para donaciones. El objetivo del estudio fue desarrollar una solución informática accesible que automatice procesos administrativos y contribuya a una gestión más eficiente y segura de los recursos sanguíneos. La metodología empleada fue el marco ágil Scrum, con el desarrollo organizado en iteraciones quincenales. La herramienta utiliza SQL Server como sistema gestor de base de datos por su fiabilidad y seguridad. La validación se llevó a cabo en un entorno simulado representativo de un banco de sangre, utilizando escenarios diseñados para emular tareas reales en la gestión de donaciones. Además, se realizó una evaluación de usabilidad con usuarios reales mediante entrevistas estructuradas, quienes identificaron fortalezas y oportunidades de mejora en la interfaz. Los resultados evidenciaron una reducción significativa de errores en los registros, mayor eficiencia operativa y una mejora en la interacción con los donantes. La evaluación de usabilidad reveló necesidades de refinamiento en la experiencia visual, propuestas como la incorporación de tutoriales interactivos y mejoras en los formularios de entrada de datos. Se concluye que el prototipo cumple satisfactoriamente los objetivos planteados, representando una solución tecnológica eficaz para fortalecer la seguridad transfusional y apoyar la toma de decisiones clínicas en contextos de infraestructura limitada
Traffic Congestion in Ecuador: A Comprehensive Review, Key Factors, Impact, and Solutions of Smart Cities
Get PDFThis study explores the main sources of traffic congestion in Ecuadorian cities and propose solutions to address this issue. The findings reveal that the main causes are natural disasters disrupting the transportation infrastructure and leading to chaotic traffic flow, lack of infrastructure maintenance, inadequate education, cultural issues, improper traffic signal timing, or the absence of exclusive lanes for public transportation. Fast transit projects have also faced obstacles, including a lack of political leadership, underestimation of implementation complexities, rushed planning processes, resistance from stakeholders like bus operators, and inaccurate cost estimations. Vehicle pollution is another consequence of lower-quality fuel and the topography of highland cities, which demand more engine power. The proposed solutions are categorized into three types: smart city technologies, implementing regulations, and enhancing public transportation systems. To address traffic accidents, it is recommended to identify high-risk areas, monitor fleet variables of buses, educate the population on responsible driving practices, and implement designated driver applications. By considering and implementing these solutions, Ecuadorian cities can alleviate traffic congestion, enhance transportation efficiency, reduce pollution, and improve road safety.Este estudio explora las principales fuentes de congestión de tráfico en ciudades ecuatorianas y las soluciones propuestas. Los hallazgos revelan que las principales causas son desastres naturales que interrumpen la infraestructura de transporte y provocan un flujo de tráfico caótico, falta de mantenimiento de la infraestructura, educación inadecuada y problemas culturales, sincronización inadecuada de los semáforos o ausencia de carriles exclusivos para el transporte público. Los proyectos de tránsito rápido también han enfrentado obstáculos, incluyendo falta de liderazgo político, subestimación de complejidades de implementación, procesos de planificación apresurados, resistencia de partes interesadas como operadores de autobuses y estimaciones de costos inexactas. La contaminación vehicular es consecuencia de la calidad inferior del combustible y la topografía de las ciudades de la sierra, que demanda más potencia del motor. Las soluciones se dividen en: tecnologías de ciudades inteligentes, regulaciones de implementación y mejoras en los sistemas de transporte público. Para abordar los accidentes de tráfico, se recomienda identificar áreas de alto riesgo, monitorear flotas de autobuses, educar a la población sobre prácticas de conducción responsables e implementar aplicaciones de conductor designado. Estas soluciones pueden aliviar la congestión de tráfico, mejorar la eficiencia del transporte, reducir la contaminación y aumentar la seguridad vial
A Visual Review and Bibliometric Analysis of Cloud Computing Traffic Flow Forecasting for a Digital Africa
Get PDFThe use of cloud computing has grown globally in recent years. When allocating resources, cloud resources and traffic flow management need to be closely watched and controlled. This paper conducted a bibliographic study based on the Scopus database and the Institute of Electrical and Electronics Engineers (IEEE) to evaluate the adoption rate of resource management and traffic flow forecasting in cloud computing in Africa. There is still uncertainty about cloud computing adoption on the African continent. It is difficult to say that 54 African countries are fully prepared to adopt cloud computing. The growth of internet connectivity and potential economic development in Africa are contributing to cloud computing\u27s rapid growth. Cloud computing enables individuals, private companies, and the government to access computing resources and services remotely. It has the potential to significantly impact the education, healthcare, and economic sectors; however, there are challenges such as limited infrastructure and data security concerns in Africa. This study demonstrates the limitations of the implementation of cloud computing in African countries such as South Africa, Nigeria, Namibia, Botswana, Zimbabwe, Uganda, Kenya, Cameron, Egypt, and Ghana. Only 18.5% of the continent is researching the implementation of cloud computing, and the lack of cloud implementation is a persistent issue because of scarce resources. Africa\u27s adoption of cloud computing can be increased through alternative solutions suggested in the study.The use of cloud computing has grown globally in recent years. When allocating resources, cloud resources and traffic flow management need to be closely watched and controlled. This paper conducted a bibliographic study based on the Scopus database and the Institute of Electrical and Electronics Engineers (IEEE) to evaluate the adoption rate of resource management and traffic flow forecasting in cloud computing in Africa. There is still uncertainty about cloud computing adoption on the African continent. It is difficult to say that 54 African countries are fully prepared to adopt cloud computing. The growth of internet connectivity and potential economic development in Africa are contributing to the rapid growth of cloud computing. This technology enables individuals, private companies and the government to access computing resources and services remotely. It has the potential to significantly impact the education, healthcare and economic sectors; however, there are challenges such as limited infrastructure and data security concerns in Africa. This study demonstrates the limitations of the implementation of cloud computing in African countries such as South Africa, Nigeria, Namibia, Botswana, Zimbabwe, Uganda, Kenya, Cameron, Egypt and Ghana. Only 18.5% of the continent is researching the implementation of cloud computing, and the lack of cloud implementation remains a persistent issue because of scarce resources. However, Africa\u27s adoption of cloud computing can be increased through alternative solutions suggested in the study
A Systematic Literature Review on Defense Techniques Against Routing Attacks in Internet of Things
Get PDFThe proliferation of the Internet of Things (IoT) has attracted different sectors such as agriculture, manufacturing, smart cities, transportation, etc. to adopt these technologies. Most IoT networks utilize Routing Protocol for Low Power and Lossy Networks (RPL) to exchange control and data packets across the network. However, RPL is susceptible to routing attacks such as rank attacks, DIS-flooding, etc. In recent years different defense techniques have been proposed to act against these attacks i.e., Secure-Protocol, conventional Intrusion Detection Systems (IDS), and Machine Learning (ML)-based. This systematic literature review explores 39 published papers in the domain of defense techniques against routing attacks in RPL-based IoT. We review. The findings of this study suggest that most Secure-Protocol can detect and mitigate routing attacks utilizing distributed placement, ML-based can detect most attacks but lack mitigation mechanisms, and conventional IDS technique utilizes a hybrid approach in detection and placement strategies. Additionally, this study reveals that India publishes more research papers in ML-based and Secure-Protocol. Furthermore, flooding attacks are the most discussed attacks in the selected studies. Finally, Cooja Contiki is the most used simulation tool.The proliferation of the Internet of Things (IoT) has attracted different sectors such as agriculture, manufacturing, smart cities, transportation, etc. to adopt these technologies. Most IoT networks utilize Routing Protocol for Low Power and Lossy Networks (RPL) to exchange control and data packets across the network. However, RPL is susceptible to routing attacks such as rank attacks, DIS-flooding, etc. In recent years different defense techniques have been proposed to act against these attacks i.e., Secure-Protocol, conventional Intrusion Detection Systems (IDS), and Machine Learning (ML)-based. This systematic literature review explores 39 published papers in the domain of defense techniques against routing attacks in RPL-based IoT. We review. The findings of this study suggest that most Secure-Protocol can detect and mitigate routing attacks utilizing distributed placement, ML-based can detect most attacks but lack mitigation mechanisms, and conventional IDS technique utilizes a hybrid approach in detection and placement strategies. Additionally, this study reveals that India publishes more research papers in ML-based and Secure-Protocol. Furthermore, flooding attacks are the most discussed attacks in the selected studies. Finally, Cooja Contiki is the most used simulation tool
SIGMA: Wireless System with Geolocation for Environmental Monitoring
Get PDFThe increase in the number of automotive parks, the emissions generated by industries, and the forest fires, among others, deteriorate the air quality of the Metropolitan District of Quito. Low-cost devices (sensors) distributed throughout the city to collect and deliver information on concentrations of gaseous pollutants in real time are essential for preserving the health of the citizens. This kind of technology can contribute to improving air quality by controlling the emissions of harmful substances into the atmosphere. This paper shows a prototype system for environmental monitoring using open hardware and software technologies. The system comprises two subsystems: a transmitter (mobile) and a receiver (fixed). The transmitter unit has been installed in a public transport vehicle (a taxi or any public transportation), which allows the acquisition of environmental parameters such as carbon monoxide, ozone, nitrogen, temperature, humidity, geographic location, time, and date. The obtained measurements are sent in real-time to a receiver subsystem, mainly consisting of a server, where the received data is processed and published in a pollution map. This data informs citizens by geographical areas, about the different levels or concentration ranges of a particular gas, and general air pollution in the city.The increase in the number of automotive parks, the emissions generated by industries, and the forest fires, among others, deteriorate the air quality of the Metropolitan District of Quito. Low-cost devices (sensors) distributed throughout the city to collect and deliver information on concentrations of gaseous pollutants in real time are essential for preserving the health of the citizens. This kind of technology can contribute to improving air quality by controlling the emissions of harmful substances into the atmosphere. This paper shows a prototype system for environmental monitoring using open hardware and software technologies. The system comprises two subsystems: a transmitter (mobile) and a receiver (fixed). The transmitter unit has been installed in a public transport vehicle (a taxi or any public transportation), which allows the acquisition of environmental parameters such as carbon monoxide, ozone, nitrogen, temperature, humidity, geographic location, time, and date. The obtained measurements are sent in real-time to a receiver subsystem, mainly consisting of a server, where the received data is processed and published in a pollution map. This data informs citizens by geographical areas, about the different levels or concentration ranges of a particular gas, and general air pollution in the city
Editorial
Get PDFWelcome to Volume 12, Issue 1 of the Latin-American Journal of Computing (LAJC).
It is a privilege to introduce this latest edition, which brings together eight high-quality research articles addressing some of the most pressing challenges in the fields of computing and technology. These studies showcase the innovation and dedication of researchers committed to advancing knowledge and tackling real-world problems.
This issue covers a diverse range of relevant and timely topics, including environmental monitoring, data visualization, systematic reviews, renewable energy systems, microservices architecture, and smart city solutions. Each article offers valuable insights and practical applications, reflecting the journal\u27s mission to foster impactful and rigorous research.
Among the contributions of this issue, readers will find:
An innovative wireless geolocation system designed to monitor and reduce pollution in urban areas.
A systematic data visualization framework tailored to support strategic decision-making in global trade.
Insights into routing attack defense mechanisms for Internet of Things (IoT) networks.
A user-friendly web-based tool for sizing photovoltaic systems aligned with Ecuador’s renewable energy targets.
Comprehensive reviews and methodologies that enhance cloud computing adoption, protect mobile users, and streamline urban traffic management.
These contributions emphasize the significance of interdisciplinary collaboration in addressing complex societal issues, both in Latin America and on a global scale.
On behalf of the editorial committee, I want to extend my gratitude to all authors, reviewers, and the editorial team whose dedication makes this issue possible. We remain committed to supporting our contributors and readers by providing a platform for sharing cutting-edge research and fostering academic collaboration.
We hope this issue inspires further innovations and meaningful conversations within the computing community.
Gabriela Suntaxi
Editor-in-Chief
Bienvenidos al Volumen 12, Número 1 de la revista Latin-American Journal of Computing (LAJC).
Me complace presentarles esta última edición, que reúne ocho artículos de investigación de alta calidad que abordan algunos de los desafíos más apremiantes en los campos de la computación y la tecnología. Estos estudios destacan la innovación y dedicación de investigadores comprometidos con el avance del conocimiento y la resolución de problemas del mundo real.
Este número cubre una amplia gama de temas relevantes y oportunos, que incluyen monitoreo ambiental, visualización de datos, revisiones sistemáticas, sistemas de energía renovable, arquitecturas de microservicios y soluciones para ciudades inteligentes. Cada artículo ofrece ideas valiosas y aplicaciones prácticas, reflejando la misión de la revista de fomentar investigaciones rigurosas y con impacto.
Entre las contribuciones de este número, los lectores encontrarán:
Un innovador sistema inalámbrico de geolocalización diseñado para monitorear y reducir la contaminación en áreas urbanas.
Un marco sistemático de visualización de datos, orientado a apoyar la toma de decisiones estratégicas en el comercio global.
Perspectivas sobre los mecanismos de defensa contra ataques de enrutamiento en redes del Internet de las Cosas (IoT).
Una herramienta web fácil de usar para el dimensionamiento de sistemas fotovoltaicos, alineada con los objetivos de energía renovable de Ecuador.
Revisiones exhaustivas y metodologías que fomentan la adopción de la computación en la nube, protección a los usuarios móviles y optimización de la gestión del tráfico urbano.
Estas contribuciones subrayan la importancia de la colaboración interdisciplinaria para abordar problemas sociales complejos, tanto en América Latina como a nivel global.
En nombre del comité editorial, deseo expresar mi gratitud a todos los autores, revisores y al equipo editorial, cuya dedicación hace posible esta edición. Seguimos comprometidos en apoyar a nuestros colaboradores y lectores, brindando una plataforma para compartir investigaciones de vanguardia y fomentar la colaboración académica.
Esperamos que este número inspire nuevas innovaciones y conversaciones significativas dentro de la comunidad de la computación.
Gabriela Suntaxi
Editora en Jef
Análisis de Sentimientos en la Red Social “X”, Percepción Pública sobre el Presidente del Ecuador, Daniel Noboa (noviembre 2023 - abril 2024).
Get PDFThis study examines public opinions and social evaluations regarding the President of Ecuador, Daniel Noboa, on the social network “X” through sentiment analysis techniques. To this end, web scraping was implemented, allowing for the efficient and cost-effective collection of 3177 relevant tweets. Subsequently, a manual labeling process of the emotions reflected in the messages was carried out, aiming to ensure a more rigorous and representative classification of users\u27 sentiments. The results revealed that 79.7% of the tweets were neutral, indicating a lack of a defined stance in most mentions. However, 16.6% of the messages expressed a negative orientation, showing a significant presence of criticism and disapproval toward the president. In contrast, only 3.7% of the tweets reflected a positive attitude, indicating a relatively low level of explicit support. These findings suggest that President Noboa’s image on “X” is predominantly neutral, with a significant tendency toward criticism and limited positive endorsement. The study demonstrates the usefulness of web scraping and sentiment analysis as key tools for evaluating public opinion on social media, providing valuable insights into the sociopolitical dynamics in digital environments.El presente estudio examina las opiniones y valoraciones sociales en torno al presidente del Ecuador, Daniel Noboa en la red social “X” mediante técnicas de análisis de sentimientos. Para ello, se implementó web scraping, lo que permitió recopilar 3177 tweets relevantes de manera eficiente y económica. Posteriormente, se efectuó un proceso de etiquetado manual de las emociones reflejadas en los mensajes, con el objetivo de asegurar una clasificación más rigurosa y representativa del sentir de los usuarios. Los resultados revelaron que el 79.7% de los tweets eran neutrales, lo que indica una falta de postura definida en la mayoría de las menciones. No obstante, se detectó que el 16.6% de los mensajes manifestaban una orientación negativa, evidenciando una presencia significativa de críticas y manifestaciones de desaprobación hacia el presidente. Por otro lado, únicamente el 3.7% de los tweets reflejaban una actitud positiva, lo que indica un nivel relativamente bajo de respaldo explícito. Estos hallazgos sugieren que la imagen del presidente Noboa en “X” es mayormente neutral, con una tendencia significativa hacia la crítica y un bajo respaldo positivo. El estudio demuestra la utilidad del web scraping y el análisis de sentimientos como herramientas clave para evaluar la opinión pública en redes sociales, proporcionando información valiosa para la comprensión de la dinámica sociopolítica en entornos digitales
Data Domain Servitization for Microservices Architecture
Get PDFMicroservices have emerged as a software design paradigm where small, autonomous services interact to meet business requirements. However, transitioning from monolithic systems to microservices presents challenges, especially when multiple subdomains share transactional tables to maintain referential integrity across separate databases. Ensuring each microservice handles business data while adhering to ACID properties (atomicity, consistency, isolation, durability) is crucial.This requires unique, consistent, and low-dependency data from a business domain perspective. A Systematic Literature Review (SLR) is a secondary research method used to evaluate the current body of scientific literature.It helps identify existing work, highlight research gaps, and propose new research directions. In software engineering, SLRs offer a comprehensive overview of studied research areas. This article reports an empirical study based on a systematic literature review aimed at identifying modeling techniques for segmenting data structures during microservice design. The review found limited methods to address the appropriate level of data granularity per microservice. These findings highlight a need for further research into processes and methodologies that can effectively handle data segmentation and consistency within microservice architectures.Los microservicios es un estilo de construcción de software basada en servicios pequeños que se diseñan para responder en entornos empresariales. Sin embargo, modelar dominios de negocios o descomponer sistemas monolitos para obtener microservicios acarrea problemas si se generan varios subdominios que comparten tablas transaccionales que mantienen integridad referencial y que se almacenan en base de datos diferentes. Adicionalmente, la forma de descomponer para cada microservicio debe asegurar la atomicidad, consistencia, aislamiento y durabilidad (ACID) de los datos del negocio, por lo tanto, se deben mantener datos únicos, consistentes y con baja dependencia desde la perspectiva del dominio de negocio. La revisión sistemática de Literatura es un método secundario de investigación que es utilizado para analizar el estado actual de la literatura científica. Es útil para encontrar publicaciones existentes en la temática planteada, identificar espacios donde no se ha investigado y plantear nuevos temas de investigación a la comunidad científica. Este método ha sido utilizado en el área de la Ingeniería de Software para dar una vista general y estado actual sobre las áreas de investigación estudiadas. En este artículo se hace reporte de un Estudio Empírico basado en la revisión Sistemática de Literatura conducido con la finalidad de determinar la existencia de técnicas de modelamiento para la segmentación de las estructuras de datos durante el diseño y construcción de Microservicios. Se da explicación del procedimiento de investigación realizado. Se pudo encontrar artículos relacionados con la temática planteada. Se encuentra que existen limitados métodos y procedimientos para lograr la comprensión y el nivel adecuado de granularidad de datos requerida en cada microservicio. A vista del resultado obtenido, se abre el escenario para el planteamiento de nuevos procesos de investigación en la temática