Revista Ingenierías Universidad de Medellín

DETECTION OF OPERATIONAL FAILURES WITH ARTIFICIAL NEURAL NETWORKS: APPLICATION TO THE TENNESSEE EASTMAN PROCESS

Giovanni Morales — 2024-04-30

The purpose of this article is to compare results of fault detection for the Tennessee Eastman (TE) process with the application of artificial neural networks (ANN). The Neuralnet library of the open-source program R, as well as the Keras library of the open-source program Python were used for the training of ANN. The TE process simulation data were down loaded from Harvard University’s server, and subsequently analyzed, defining the trends in the operational variables during the appearance of failures. With the database, the training and validation of different ANN structures were developed, considering the parameters number of hidden neurons, activation function, and number of hidden layers. According to the results, the training and validation of the ANNs with the Neuralnet library yielded a lower performance in fault detection than that obtained with the Keras library. The ANN with the best performance in detecting failures in the TE process was obtained by the application of the Keras library. This ANN considered 52 input variables, 11 neurons in the hidden layer, and one neuron in the output layer, using a logistic function (ANN represented as 52:11:1 logistic) and reporting a prediction efficiency of 92% for the detection of faults with an external test set, which is convenient for future implementation in industrial processes.

INGENIERÍA INVERSA PARA EL MANTENIMIENTO DEL SOFTWARE: UNA REVISIÓN SISTEMÁTICA DE LITERATURA

Raúl Antonio Aguilar Vera — 2024-05-02

Reverse engineering covers a wide range of software life cycle activities starting with the existing implementation, recovering, or recreating the design and deciphering the requirements implemented in the system. The objective of this article is to present a state of the art on Reverse Engineering techniques used in the context of software maintenance. For this purpose, a Systematic Literature Review was carried out, in which 47 primary studies from 4 databases, published in this century, were analyzed. Research questions were raised regarding the distribution of publications throughout this century and, about the type of maintenance on which they focus; Likewise, the techniques and metrics reported for reverse engineering processes in maintenance activities were analyzed. A significant increase (just over 200 %) could be observed in the second decade of this century compared to the first. On the other hand, it was observed that perfective maintenance is the one in which the most techni ques were identified, although the majority are static. From the analysis of the selected studies, a total of 22 types of reengineering techniques used for software maintenance/evolution tasks were identified, reporting a set of eight types of metrics used to evaluate the quality of reengineering processes in the evolution of the software.

RECONOCIMIENTO DE TÉCNICAS OFENSIVAS EN ARTES MARCIALES: UN MAPEO SISTEMÁTICO

Jairo Josué Cristobal Franco — 2024-05-10

Motivation: The precise identification of punches and kicks in martial arts sporting competitions is a critical issue, often complex, and at times subject to controversies due to the subjective judgment of referees. Problem: The subjectivity in assessing punches and kicks during martial arts sporting competitions poses a significant challenge regarding impartiality and accuracy in refereeing. Solution Approach: This study analyzes the most recent contributions to punch and kick recognition in martial arts competitions. It reviews classification techniques, commonly used sensors, and the performance achieved in identifying these movements. Results: The analysis provides a general overview of implemented punch and kick classification techniques. This contributes to understanding recent advancements in this field and how they can enhance objectivity and precision in refereeing martial arts competitions. Conclusions: This study underscores the growing interest in machine learning techniques for classifying punches and kicks in martial arts, encompassing a wide range of classifiers, from traditional methods to deep learning models. The combination of inertial sensors and depth cameras emerges as a promising avenue. Future research is expected to thoroughly compare and characterize these approaches, paving the way for implementing artificial intelligence systems in martial arts competitions and potentially revolutionizing the objectivity in assessing movements in this sport.

PRÁCTICAS DE ACCESIBILIDAD EN EL DESARROLLO DE SOFTWARE: MAPEO SISTEMÁTICO DE LA LITERATURA

María Elena Fernández Castillo — 2024-05-10

Incorporar la accesibilidad en el desarrollo de software conlleva beneficios significativos como la inclusión de personas con discapacidad. Reconocer la diversidad de usuarios con diferentes requisitos y necesidades dentro de la creación de software facilita su utilización por la mayor cantidad de usuarios sin importar sus condiciones individuales. El objetivo de esta investigación es obtener una base de conocimientos que permita identificar aquellas prácticas de las que se tenga evidencia de su uso en el desarrollo de software accesible. Se llevó a cabo un Mapeo Sistemático de la Literatura (MSL) enfocado en identificar prácticas que incorporan actividades de accesibilidad en el ciclo de vida del software, resultando en 29 estudios que se ajustaron a los criterios de selección. Los resultados muestran que la fase con más evidencia de incorporación de actividades de accesibilidad en el desarrollo de software es la fase de pruebas, donde se utilizan evaluadores para verificar la accesibilidad dentro de páginas web, seguido de la fase de construcción, donde se utilizan metodologías que incluyen la accesibilidad dentro del proceso de desarrollo de software, en la fase de diseño es común que se incorporen artefactos que guíen el desarrollo de accesible y, por último, en la fase de requisitos es usual emplear la descripción de requisitos para obtener y documentar requisitos de accesibilidad.

IDENTIFICATION OF A UNIVERSITY COMMUNITY ON FACEBOOK FOR THE DISSEMINATION OF SCIENCE AND CULTURE

Paola Velazquez-Solis — 2024-05-10

Social network analysis is a method that allows the identification and exa mination of structures of various types of data, objects, or user groups, as well as the interactions created by a community and the relationships that exist between them. To characterize the virtual community of a higher education institution, the Fruchterman-Reingold, Yifan Hu, and Noverlap algorithms were applied using the Gephi tool for the analysis and modeling of 30 Facebook fan pages dedicated to the promotion of culture and the dissemination of science. The results show a relationship with the direction and number of edges for each of the 30 nodes for each algorithm applied. The focus of the analysis provided information about the dynamics of the virtual university community, allowing a visual understanding of how members connect and communicate within the Facebook social network.

TOOL FOR ESTIMATING USER STORIES ASSOCIATED WITH NON-FUNCTIONAL REQUIREMENTS

Liliana González-Palacio — 2024-05-10

One of the initial steps in developing any software product is defining its functional and non-functional requirements, the latter also known as quality requirements. Several stakeholders are involved in this process. However, estimating the effort needed to implement these requi rements, particularly the quality ones, is a complex task. To overcome this challenge, we present Story Points Predictor (SPP), a tool for predicting the effort required to implement quality requirements using artificial intelligence techniques. Based on historical data, SPP estimates the size of the requirements and classifies them into three groups: small, medium, and large. Development teams can use SPP to complement their effort estimation process for requirements. Experimental results show that SPP has a 72% accuracy.

PROPOSAL OF A TIME SERIES-BASED MODEL FOR THE CHARACTERIZATION AND PREDICTION OF DROPOUT RATES AT THE NATIONAL OPEN AND DISTANCE UNIVERSITY

Gabriel Elías Chanchí G. — 2024-05-15

Dropout rates are a key indicator of educational quality, making it imperative for educational institutions to design strategies to reduce them, thereby contributing to improved student retention and the achievement of academic objectives. While dropout research has primarily focused on machine learning methods applied to in-person education datasets, this article introduces a novel approach based on time series models for dro pout rates analysis at the National Open and Distance University (UNAD). Methodologically, an adaptation of the CRISP-DM methodology was undertaken in four phases, namely: F1. Business and data understanding, F2. Data preparation, F3. Model building and evaluation, and F4. Model deployment. In terms of results, an open dataset on UNAD dropout, ob tained from the SPADIES system between 1999 and 2021, was employed. Using Python libraries statsmodels and pandas, an ARIMA model was implemented, displaying optimal error metrics. This ARIMA model was utilized to forecast future dropout rates at UNAD, projecting a future dropout rate fluctuating around 23%. In conclusion, the ARIMA model developed for UNAD stands as an innovative and essential tool in the educational realm, capable of accurately anticipating dropout rates for upcoming semesters. This provides UNAD with a unique advantage in strategic decision-making.