OAOscar Mauricio Gelves Alarcón

Oscar Mauricio Gelves Alarcón

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Oscar Mauricio Gelves Alarcón has a total of 12 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    PREDICCIÓN DEL PRECIO DE LA UCHUVA EN COLOMBIA: comparación entre Métodos Clásicos y red neuronal LSTM.

    by Oscar Gelves, Elisa Navarro

    Los precios agrícolas en particular de frutas presentan alta volatilidad debido a factores productivos, de mercado, climáticos y logísticos, lo cual dificulta la toma de decisiones en los sistemas agroalimentarios. Por tanto, el diseño y utilización de modelos de pronósticos confiables se hacen necesarios con el fin de apoyar la planificación agrícola. El objetivo de este estudio es evaluar el desempeño de los modelos Brown, Holt Winters, SARIMA y Long Short – Term Memory (LSTM) en la predicción del precio mensual de la uchuva en Colombia durante el periodo comprendido 2015 -2025. Para la realización del diseño metodológico se utilizó un enfoque de tipo cuantitativo de análisis de serie de tiempo. Por tanto, se evaluó la estacionariedad de la serie por medio de pruebas estadísticas y se aplicó la transformación logarítmica con el objetivo de estabilizar la varianza de los datos. A continuación, la base de datos se dividió en conjuntos de entrenamiento y prueba con el fin de estimar y validar los modelos. El desempeño de los modelos en estudio fue evaluado por medio de los indicadores Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) y Root Mean Squared Error (RMSE). Los resultados obtenidos demuestran que los modelos Brown y Holt Winters representan de forma adecuada la tendencia general de la serie, sin embargo, tienden a suavizar de forma excesiva las variaciones abruptas. El modelo SARIMA presenta el menor nivel de ajuste ya que es afectada por los comportamientos no lineales y a la falta de patrones estacionales. En contraste, el modelo LSTM alcanza el mejor desempeño, mostrando niveles de error más bajos y con una mayor capacidad para capturar la dinámica de los precios de la uchuva.

  • Open Access

    Modelos predictivos para la estimación de oferta de fresas en Cundinamarca: un enfoque comparativo con aprendizaje automático y regresión

    by Oscar Mauricio Gelves Alarcón, Nataly Lorena Guarín Cortés, María Paula Peña Martínez, María Fernanda Rebolledo Marulanda

    Strawberry cultivation in Cundinamarca represents a key agricultural activity with a significant regional economic impact. This study aims to evaluate and compare different forecasting models to estimate the future supply of strawberries, considering variables such as cultivated area, yield, and production. Time series methods (Holt and Brown exponential smoothing), the AdaBoost algorithm, and a multiple linear regression model were applied, using historical data from the Ministry of Agriculture and Agronet. The results indicate that the multiple linear regression model exhibited the best performance, achieving a coefficient of determination of 0.9947 and outperforming the time series methods, which presented average errors exceeding 17%. Normality and homoscedasticity tests statistically validated the robustness of models. It is concluded that the proposed model provides an effective tool for agricultural planning in Cundinamarca allowing for projections of production increases associated with the expansion of cultivated areas and improvements in yield per hectare. Future research is recommended to integrate climatic variables and agricultural policy factors to enhance forecasting accuracy.

  • Open Access

    Modelo de Dinámica de Sistemas para la Gestión de un Programa Académico Universitario de Pregrado

    by Universidad Militar Nueva Granada, Oscar Mauricio Gelves Alarcón, David Alejandro Celemín Carrillo, Danna Alejandra Casteblanco Tenjica, Brayan Yesid Collazos Sarmiento

    El objetivo de este estudio es diseñar un modelo de dinámica de sistemas que aborde la evolución de la población de un programa de pregrado mediante un enfoque sistémico y causal. Se propone un enfoque que integre variables como la población estudiantil actual, los costos de matrícula y los resultados económicos del programa, con el fin de proporcionar una visión integral de la situación. La metodología de base se fundamenta en la utilización de un diagrama de Forrester para la construcción del modelo, el cual permite simular las variables hasta el año 2043. Además, se lleva a cabo una revisión literaria utilizando bases de datos como Scopus y Google Academics para contextualizar el modelo propuesto. Los resultados principales de esta investigación incluyen la capacidad del modelo para proyectar diferentes escenarios, como la falta de políticas gubernamentales económicas hacia las universidades. Asimismo, se destaca la identificación de temas y variables relevantes mediante la revisión literaria realizada. En conclusión, este estudio proporciona un modelo integral que permite proyectar la evolución de la población estudiantil en un programa de ingeniería, ofreciendo una herramienta para la toma de decisiones y la formulación de estrategias a largo plazo en el ámbito académico.


  • Open Access

    Propuesta de modelo de simulación de la Población Carcelaria en Colombia mediante Dinámica de Sistemas

    by Oscar Mauricio Gelves Alarcón, Nataly Guarin Cortes, Laura Gabriela Garnica Galvis, María Alejandra Alarcón Alarcón

    The main objective of this document is to present a system dynamics model that enables the simulation and prediction of the prison population in Colombia for the year 2050. This model is based on statistical data provided by INPEC and offers a detailed analysis of the prison population's behavior, considering factors such as detentions, inmate deaths, releases, births, and deaths of Colombian citizens. Additionally, the issue of overcrowding is comprehensively addressed by incorporating relevant auxiliary variables. Based on the model's construction, various scenarios are proposed to facilitate a deeper understanding of the underlying dynamics and trends. According to this analysis, the simulation estimates that the prison population in 2050 will be approximately 570,000 individuals.

  • Open Access

    Análisis de la aplicación del modelo CONWIP en los países latinoamericanos: Una revisión de literatura

    by Oscar Gelves

    El modelo Conwip fue diseñado en los años 90 del siglo XX por Spearman, publicándose en el libro Factory Physics con el objetivo de la disminución del inventario en proceso o WIP; para ello esta revisión de literatura tiene como objetivo reconocer el desarrollo de este modelo en el contexto latinoamericano, teniendo en cuenta las publicaciones durante los últimos años en diferentes bases de datos como Scopus, ScienceDirect, Scielo y Google Academics; en base a los resultados de la ecuación de búsqueda, se establece el tamizaje por medio de la metodología prisma y se definen las principales estadísticas por país, enfoque de investigación , herramientas complementarias y promedio por año de las publicaciones. Según el tamizaje realizado, se define que Brasil es el país de la región con mayor número de publicaciones en las cuales el enfoque de investigación, en su mayoría, son estudios de casos. El nivel de publicaciones en la región de Latinoamérica es bajo en comparación a otras regiones del mundo y es posible aplicar el modelo Conwip en diferentes contextos aparte de la manufactura, como, por ejemplo, la construcción.

  • Open Access

    Application of the experimental design for the analysis of nonattendance to scheduled medical appointments in the Colombian Health System

    by Elisa del Carmen Navarro Romero, Natalia García-Corrales, Óscar Mauricio Gelves-Alarcón, Wilfrido Javier Arteaga-Sarmiento

    This paper shows a study about nonattendance to medical appointments in the Colombian health system to determine which strategies have the greatest effect on the response variable (nonattendance) of scheduled medical appointments. Three factors were considered (calling the patient, reconfirming on the eve of the appointment, confirming understanding of the procedure), each with two or three levels randomly selected from a population. Two replicates were taken for each combination of treatments. The experimental runs were donerandomly to guarantee the absence of bias in data collection. With the above, a total of 180 experimental runs were carried out. This study aims to determine the factors that influence the population when they do not attend their medical appointments to propose a strategy that reduces this phenomenon. A design of experiments (DoE) was performed, specifically a completely randomized design (CRD) and a 3k factorial design

  • Open Access

    Estado del arte y la técnica de las prácticas lean en instituciones de salud en América Latina: Revisión de Literatura

    by Oscar Mauricio Gelves Alarcón, Elisa del Carmen Navarro Romero, Natalia García Corrales

    This literary review article aims to define the statistics related to scientific publications on the application of Lean health care tools in Latin American health institutions, for which a literary review of publications on the subject in the last ten years was defined Based on the database such as Scopus and gray literature, for the search development review articles and applications outside the Latin American region were excluded, in addition, a search equation is formulated using keywords such as Lean, Healthcare, Latin American country and Boolean. As a result of the equation and the corresponding screening, a total of 21 publications were found where it is established that Brazil is the country with the highest number of publications and the Lean tools most used in health institutions were VSM (Value Stream Mapping) and Lean Thinking.


  • Open Access

    Correlational analysis between the economics, socio-demographic indices and statistics of contagion due to Covid-19, applying the Clustering methodology in countries of America

    by Elisa del Carmen Navarro Romero, Óscar Mauricio Gelves Alarcón, Natalia García Corrales

    Introduction— This research is motivated, by the current world situation, caused by the pandemic declared by the WHO before the spread and severity of the coronavirus disease (COVID-19), notified for the first time in Wuhan (China) on December 31 of 2019. Through mathematical and statistical analysis, it seeks to show and explain in an expeditious manner, the causes for which there is a higher rate of contagion and lethality due to the virus, in different countries, taking into consideration patterns associated with social political behavior and economic, as a first approach to knowing statistics that allow generating forecasts for future periods, given the conditions. Objective— The main objective of this work is to define the correlation of the economic, social and demographic variables of the countries of America, with respect to the contagion of the virus, proposing a forecast model on the level of contagion in each cluster proposed by the different regions of the American continent. Methodology— The study performs clustering (grouping) of the countries of America with respect to their geographical position North America, Central America and the Caribbean islands and South America, followed by a search for statistical data on social, economic and demographic indicators of the countries of America in recent years and statistics of levels of contagion of COVID 19 in sources such as international organizations regulating health issues. Next, a characterization and correlation of the collected data was carried out, to finally, based on the results of the correlation, make a forecast of the level of contagion that would be reached by each of the regions. Results— The purpose of this document is to provide information on the countries of North America, Latin America and the Caribbean with respect to the analysis of mortality from COVID-19, through methods of analysis of mortality from all causes as one of the approaches proposed to contribute to the assessment of the true magnitude of the burden of the COVID-19 epidemic in these countries. Conclusions— The results show interesting information, since the Latin American curve turned out to be much less pronounced than that of the United States, in terms of contagion and deaths, despite the socio-demographic conditions, economic, technological and political opportunities. This analysis invites us to find out which are those correlations that directly impact the behavior of infections, taking into account variables such as age, gender, stratum, level of education, and other sociodemographic characteristics that may influence the spread or containment of the virus.