Examinando por Materia "APRENDIZAJE AUTOMÁTICO (INTELIGENCIA ARTIFICIAL)"
Mostrando 1 - 20 de 162
Resultados por página
Opciones de ordenación
Publicación A Machine Learning driven Energy Management System for tropical household prosumer nodes(Universidad EAFIT, 2024) Giraldo Pérez, Juan Pablo; Mejía Gutiérrez, RicardoPublicación A Novel Ornstein Uhlenbeck Levy Model Conditioned on an Unknown Mean : Frecasting of the VIX(Universidad EAFIT, 2024) Aguirre Posada, Mario; Almonacid Hurtado, Paula María; Pérez Monsalve, Juan PabloPublicación A retail demand forecasting system of product groups characterized by time series based on “ensemble machine learning” techniques with feature enginnering(Universidad EAFIT, 2022) Mejía Chitiva, Santiago; Aguilar Castro, José LisandroPublicación Adopción de tecnologías de Inteligencia Artificial : un estudio para las empresas en Colombia(Universidad EAFIT, 2023) Vera Otálvaro, Lina María; Atis Ortega, Karen LisethThis document provides a broad description regarding the adoption of Artificial Intelligence (AI) technologies in Colombian companies. We managed to discriminate the analysis by economic sectors using a new module presented in the Encuesta Pulso Empresarial (EPE) of the National Administrative Department of Statistics (DANE) in 2022. The module collects data from more than 8,500 companies on the adoption, reasons for use and non-use of AI technologies. During the investigation, we discovered that organizing administrative processes is one of the main issues that the implementation of AI would solve. Nonetheless, high acquisition costs and lack of experienced personnel capable of using AI makes its implementation more difficult. Through a probability model, it is evident that, among other factors, using the Internet, digital platforms, investing in software equipment, carrying out research and development activities, increase the probability of adopting AI technologies. These research adds as an innovative element a set of variables considering perceptions and expectations that companies have related to present and future economic situations and its influence on the adoption of AI technologies.Publicación Ajuste fino de un modelo LLM para realizar reportes resumidos de expertos en trading, con integración de datos desde redes sociales(Universidad EAFIT, 2025) Restrepo Acevedo, Andrés Felipe; Martínez Vargas, Juan DavidThe contemporary financial market is characterized by its high complexity and the massive volume of structured and unstructured data generated daily, posing significant challenges for individual investors in terms of analysis and informed decision making. This project proposes the fine-tuning of a Small Language Model (SLM) integrated into a tool capable of generating financial analysis reports similar to those produced by experts. For the proof of concept (PoC), transcripts from financial analysis videos published by experts on their YouTube channels are utilized. The SLM is fine-tuned using instruction-based techniques and the incorporation of the LoRa(Low-Rank Adapters) method, with the aim of extracting and summarizing key information relevant to individual investors. The main objective of this tool is to assist individual investors by generating efficient and accessible reports, facilitating access to valuable information in natural language, and enhancing their ability to make data-driven decisions from unstructured data, all with minimal investment of time and resources. Experimental results demonstrate the viability of using fine-tuned Small Language Models (SLMs) for the generation of high-quality financial reports. Specifically, the selected model, finetune qlora unsloth llama 3.1 8B Instruct bnb 4bit v2 Q8 0, achieved an average score of 5.67 out of 10 in the evaluation conducted by a judge LLM, with an average cosine distance of 0.159 compared to the reference summaries generated by the foundational pretrained model GPT-4.1. This improvement represents a 97.5% increase in performance compared to the same base model, Llama 3.1 8B Instruct, without fine-tuning. Qualitatively, the model exhibits high fidelity and coherence in the extraction and synthesis of key information in moderately long contexts, although it faces challenges in thematic interpretation when dealing with considerably lengthy transcripts. Additionally, implementation of this tool is projected to save 560 hours annually for individual investors, along with an estimated annual reduction in API costs ranging from 7.52 to 25 for the channels analyzed in the proof of concept.Publicación Análisis comparativo de modelos predictivos para la estimación de PM2.5 : un enfoque basado en aprendizaje automático y predicción conformal(Universidad EAFIT, 2024) Camelo Valera, Matías; Martínez Vargas, Juan David; Sepúlveda Cano, Lina MariaFine particulate matter (𝑃𝑀2.5pollution poses a significant environmental and public health challenge, requiring accurate predictive models for its monitoring and control. This study compares different machine learning approaches, including Linear Regression, Random Forest, and XGBoost, with and without the inclusion of mobility variables, to estimate 𝑃𝑀2.5 levels. Additionally, inductive conformal prediction is implemented to quantify uncertainty in the estimates and provide confidence intervals with 𝛼=0.05. The results show that while XGBoost experiences performance deterioration during training when mobility variables are included, it achieves the best validation performance with the lowest mean absolute error and the highest coefficient of determination. Conformal prediction enabled the establishment of confidence intervals with 89.26% coverage, close to the expected 95%, ensuring model reliability across different spatial and temporal scenarios. In conclusion, the use of machine learning models combined with advanced validation and calibration techniques, such as conformal prediction, enhances the accuracy and reliability of 𝑃𝑀2.5 estimation. However, the quality of input variables, particularly mobility-related data, remains a challenge, highlighting the need to incorporate meteorological information and improve data resolution. These findings contribute to the development of more reliable predictive tools for environmental management and air quality policy decision-making.Publicación Análisis de discurso de los máximos responsables de las empresas participantes en el COLCAP(Universidad EAFIT, 2024) Cuervo Garcia, Dairo Alberto; Pantoja Robayo, Javier Orlando; Ceballos Cañón, Johan ArmandoPublicación Análisis de explicabilidad en modelos predictivos basados en técnicas de aprendizaje automático sobre el riesgo de re-ingresos hospitalarios(Universidad EAFIT, 2023) Lopera Bedoya, Juan Camilo; Aguilar Castro, José LisandroBig Data and medical care are essential to analyze the risk of re-hospitalization of patients with chronic diseases and can even help prevent their deterioration. By leveraging the information, healthcare institutions can deliver accurate preventive care, and thus, reduce hospital admissions. The level of risk calculation will allow planning the spending on in-patient care, in order to ensure that medical spaces and resources are available to those who need it most. This article presents several supervised models to predict when a patient can be hospitalized again, after its discharge. In addition, an explainability analysis will be carried out with the predictive models to extract information associated with the predictions they make, in order to determine, for example, the degree of importance of the predictors/descriptors. In this way, it seeks to make the results obtained more understandable for health personnel.Publicación Análisis de la responsabilidad civil en caso de una falla funcional en un vehículo operado con inteligencia artificial que resulte en daños a terceros(Universidad EAFIT, 2024) Tamayo Restrepo, Alejandro; Gil Berrio, Manuela; Villa García, Luis FelipeThis study will examine the implication of Artificial Intelligence (AI) with autonomous vehicles and its legal scopes, focusing on civil liability for functional failures. In addition, it will investigate whether the Colombian regulatory framework is prepared to address potential accidents caused by AI. For this purpose, two main areas will be explored: (i) the functioning of AI and (ii) the legal framework of tort liability in Colombia. Subsequently, the characteristics of AI and its dependence on human instructions will be analyzed to, finally, examine the theory of civil liability proposed by different authors and compare them with the legal perspective of foreign legal systems. This research seeks to understand and address the legal challenges of AI in autonomous vehicles, with the aim of contributing to the development of a fair and equitable legal framework in Colombia.Publicación Análisis de la tendencia de la solución de una interacción con un Chatbot de atención al cliente, basado en análisis de sentimiento y otras variables(Universidad EAFIT, 2023) Flórez Salazar, Luz Stella; Montoya Múnera, Edwin NelsonA chatbot is a program created with artificial intelligence. In the context of customer service, can establish conversations with customers and they are trained to resolve their queries, problems and complaints. A chatbot’s skill to identify when a customer is not meeting their request represents a challenge for companies that currently use this technology. One of the strategies to avoid quitting the conversation for this reason, is to escalate or transfer the conversation to a human agent. Therefore, it is essential to detect when it is time to carry out this escalation. This project evaluates different Natural Language Processing (NLP) techniques, rule-based labeling algorithms, classical supervised machine learning models and a simple neural network for classification, applied to interactions between a customer service chatbot and a user, in order to find a mechanism for automatic labeling of the data and to build a model that can be used to make the decision on whether the customer should continue interacting with the chatbot or if he should be transferred to a conversation with a human agent. The labeling mechanism could also be used to classify historical data, to later train a model. Different models and techniques are evaluated and those with the best performance in detecting the conversations that should escalate to a human agent are presented.Publicación Análisis de las características y atributos de una herramienta de inteligencia artificial aplicable al modelo de relacionamiento comercial del segmento de clientes preferenciales de Bancolombia(Universidad EAFIT, 2025) Saldarriaga Bonilla, Jonathan; Barcasnegras Ardila, Félix Ricardo; Uribe de Correa, Beatriz Amparo; Londoño Vélez, NataliaPublicación Análisis de los resultados de la aplicación del instrumento para la evaluación docente de la universidad EAFIT(Universidad EAFIT, 2024) Fernández Carmona, Laura Catalina; Guarín Zapata, Nicolás; Mola Ávila, José Antonio; Universidad EAFITPublicación Análisis de patrones de violencia armada en la frontera de Colombia con Venezuela usando algoritmos de aprendizaje automático(Universidad EAFIT, 2025) Lopera Pai, Daniela; Aguilar Castro, José LisandroPublicación Análisis de registros de mantenimiento de centrales de generación de energía con técnicas de procesamiento de lenguaje natural(Universidad EAFIT, 2024) Ocampo Davila, Andrés Alonso; Salazar Martínez, Carlos AndresPublicación Análisis predictivo de la deserción laboral en BPO : aplicaciones avanzadas de Machine Learning(Universidad EAFIT, 2023) Castelblanco Benítez, Julián; Almonacid Hurtado, Paula MariaPublicación Análisis predictivo del riesgo de default en microcrédito un enfoque de machine learning en sector financiero(Universidad EAFIT, 2024) Mendoza Trillos, Laura; Suárez Sierra, Biviana MarcelaPublicación Análisis y predicción de la deserción de empleados : un caso de estudio en la industria de software colombiana(Universidad Eafit, 2022) Sierra Buriticá, Eliana Marcela; Almonacid Hurtado, Paula MaríaThe objective of this study is to carry out the analysis and prediction of the desertion of employees of a software company in Medellín, based on a private database that contains 19 characteristics of 1497 workers, where 900 are active in the company and the rest have left their job. In the first place, a descriptive and exploratory analysis was carried out, where it was found that there was some variables that did not contribute information to the model, such as: Type of identification, start date of the contract, among others, also in this part the correlation of some variables and proceeded to eliminate them from the set of descriptive characteristics of the problem, since that leaving them would be leaving redundant information in the model. Second, they trained 4 machine learning models (Niave Bayes, Random Forest, Decision Tree, Logistic Regression) and the results obtained by each were compared, in order to find the one that best fits the problem of labor desertion, in this step it was found that the best classifier of machine learning is a decision tree (Decision Tree) with 14 layers, since metrics such as its curve of learning and ROC curve gave better results than the other two trained models.Publicación Análisis y predicción de recontactos en un contact center(Universidad EAFIT, 2021) Velásquez Gaviria, Diana Catalina; Quintero Montoya, Olga LucíaThe main purpose of this work is to solve in a methodical and formal way, making use of ma- chine learning models, a real problem of the productive sector that allows in addition to adding value for decision making, to provide a methodology and a compact, simple and reliable model that can be deployed and put into production in the technological platform that supports the call handling of a contact center so that benefits can be generated in the provision of the service for different sectors, generating efficiencies in the use of the channel and maximizing the customer experience in the attention of their requirements. To achieve this purpose, an airline dataset was taken containing the detail of all historical calls made by customers to a contact center during a period of 7 months (February to August 2019) and information associated with the performance of the agents handling those calls in order to predict whether users will generate at least one recontact to the contact center for the attention of their requirements before three days with respect to their initial contact. The methodology used was focused on making an ap- propriate selection of features and choosing a machine learning model that generates optimal results and enables an easy implementation allowing to identify in real time those customers with high probabilities of recontacting so that a strategy can be developed with them to impro- ve their experience. It was found that with 7 variables associated with the historical behavior of customers in the use of the channel such as frequency of calls (amount), average duration, number of agents who have served the customer (agents), time elapsed between first and last call (validity), number of days per month in which calls are made (AverageDaysInMonth), ave- rage number of calls per day (Average-Day) and time elapsed since the customer’s last call to the cut-off date of the analysis (Recency), it is possible to predict with a simple model and with very good results (AUC=88.9 %) whether a customer will call the contact center again.Publicación Análisis y predicción de ventas de motos haciendo uso de la metodología “Customer Value Map” y técnicas de Machine Learning(Universidad EAFIT, 2024) Díaz Cordero, Sandra Marcela; Martínez Vargas, Juan David; Vallejo Correa, Paola AndreaPublicación Analizando patrones de éxito en YouTube : un sistema de recomendación para creadores de contenidos educativos(Universidad EAFIT, 2024) Osorio Urrea, Vanessa; Ortiz Arias, Santiago; del Castillo Cortázar, Francisco Javier