Examinando por Materia "CIENCIA DE LOS DATOS"
Mostrando 1 - 8 de 8
Resultados por página
Opciones de ordenación
Publicación Agente de inteligencia artificial para el apoyo a la primera impresión diagnóstica a partir de descripciones sintomáticas expresadas en lenguaje natural(Universidad EAFIT, 2025-11-24) Bertel Morales, Juan Pablo; Jaramillo Múnera, Yomin EstivenThis thesis proposes the development of an artificial intelligence (AI) agent capable of supporting the generation of an initial diagnostic impression based on symptoms expressed in natural language. The project is grounded in the recognition that medical diagnosis is a complex task prone to errors, particularly when it relies on subjective and unstructured descriptions. To support clinical decision-making, natural language processing and machine learning techniques were applied following the CRISP-DM methodology. The model was trained using the synthetic DDxPlus dataset, which enabled the simulation of clinical scenarios without compromising real patient information. In the process, symptoms were transformed into synthetic anamneses through semantic normalization and subsequently vectorized using various biomedical embedding models. These representations were then used to train a supervised model tasked with associating each narrative with the confirmed diagnosis. As an additional evaluation, a “stress test” was conducted in a simulated environment, in which a healthcare professional interacted directly with the system to assess its ability to interpret real symptomatic descriptions and generate preliminary diagnostic suggestions in a coherent, consistent, and safe.Publicación Dengue Forecasting in Medellín Using Climate-Driven Time Series and Deep Learning Model(Universidad EAFIT, 2025) Niño Barbosa, Daniela Ximena; Almonacid Hurtado, Paula MaríaPublicación Estudio exploratorio de técnicas de visión por computadora para la predicción de demanda : caso de estudio Cooperativa de Caficultores del Alto Occidente de Caldas(Universidad EAFIT, 2026-02-02) Olarte Hernández, Jaber Andrés; Arbeláez Estrada, Juan CarlosPublicación Modelos de aprendizaje automático para el pronóstico de la demanda en eventos promocionales en el comercio electrónico : un estudio aplicado a ventas minoristas(Universidad EAFIT, 2026-02-05) Duque Giraldo, Tomás; Fonseca Valero, Diego FernandoThe growing importance of e-commerce has intensified the need for accurate demand forecasts, especially during promotional events that generate abnormal consumption peaks such as Black Friday or Christmas. These events lead to sudden increases in sales that traditional statistical methods fail to anticipate adequately. As a result, stock shortages or overstocking may occur. To address this problem, this paper applies advanced machine learning techniques capable of incorporating nonlinear patterns and contextual variables, with the aim of improving the accuracy of demand estimates in such scenarios. In particular, a tree-based machine learning predictive model known as LightGBM is evaluated against traditional forecasting methods, using a public dataset of US retail sales. This identifies the most effective approach under conditions of high volatility and examines its relevance during holiday and promotional periods when consumer behavior changes significantly.Publicación Patrones espacio–acústicos en sonidos respiratorios multicanal : integración de procrustes, clustering y modelos supervisados en pacientes con EPOC(Universidad EAFIT, 2025-03-03) Escobar Pajoy, Sebastián; Fonseca Valero, Diego FernandoAuscultation is a fundamental tool for assessing respiratory conditions; however, its interpretation is limited by examiner subjectivity and by the strong influence that recording location exerts on the acoustic properties of lung sounds. This work proposes a multichannel analysis strategy aimed at characterizing the spatial organization of respiratory sounds in patients with COPD and evaluating the relative contribution of different thoracic regions to the detection of adventitious events. Using simultaneous recordings from seven chest locations in the ICBHI 2017 dataset, a collection of respiratory segments was constructed and described through spectral and cepstral features. Multichannel configurations were projected into low dimensional spaces and aligned using Procrustes analysis, enabling comparable geometric representations across subjects and breathing cycles. An unsupervised clustering scheme applied to these representations revealed recurrent spatial patterns associated with different distributions of wheezes and crackles. In addition, supervised Random Forest models were trained for adventitious sound detection, incorporating feature-importance analyses and channel-ablation experiments to examine the contribution of each thoracic region. The results indicate that multichannel spatial information contains structured patterns that can be leveraged both to group thoracic configurations and to enhance the interpretability of classification models, contributing to more robust and explainable representations of pulmonary acousticsPublicación Perfilamiento vocacional de jóvenes en condiciones vulnerables del Área Metropolitana del Valle de Aburrá, bajo un marco de analítica de datos(Universidad EAFIT, 2026) Acosta Gómez, Ángela Camila; Pérez Rave, Jorge IvánPublicación Pronóstico de aportes hídricos para operación energética a partir de modelos directos (ML-TS)(Universidad EAFIT, 2026) Díaz Giraldo, Harold Nolberto; Saldarriaga Aristizábal, Pablo Andrés; No aplicaPublicación Significant gender differences in sociocognitive constructs of STEM students : an analysis supported by machine learning techniques(Universidad EAFIT, 2025-06-04) Gutiérrez Rivera, Juan Sebastián; González Palacio, Liliana; Montoya-Noguera, Silvana; González Palacio, Liliana; Montoya Noguera, Silvana; Universidad EAFIT Alianza 4U