Ciencias Exactas y Ciencias de la Salud
Permanent URI for this collectionhttps://hdl.handle.net/11285/551039
Pertenecen a esta colección Tesis y Trabajos de grado de las Maestrías correspondientes a las Escuelas de Ingeniería y Ciencias así como a Medicina y Ciencias de la Salud.
Browse
Search Results
- Domain-adapted pretraining and topic modeling for identifying skills categories in job postings(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2025-12-05) Madera Espíndola, Diana Patricia; Ceballos Cancino, Héctor Gibrán; Vázquez Lepe, Elisa Virginia; mtyahinojosa, emipsanchez; González Gómez, Luis José; Fahim Siddiqui, Muhammad Hammad; Cantú Ortiz, Francisco Javier; Escuela de Ingeniería y Ciencias; Campus Estado de México; Butt, SaburThe need to identify and cluster related skills in job postings has become increasingly essential as the labor market becomes more complex, driven by the rapid growth in job market data and continuous shifts in economic conditions, technology, and skill requirements. This task is especially challenging for postings in low-resources languages such as Spanish, as there is a lack of models specifically trained to handle these language variations. Previous work in this regard involves taxonomies created by experts such as ESCO, intended to be used as reference points via measured skills. However, some issues associated with these systems stem from their reliance on region-specific taxonomies as well as their rigidity to adapt to the changing environment of the market. Thus, we proposed a method to improve skill identification performance within the Mexican automotive industry by grouping equivalent skills present in Spanish job postings through the integration of text normalization, a Domain-Adaptive Pre-training (DAPT) Spanish BERT model, the use of BERTopic for pseudo-labels extraction, the improvement of vocabulary representation via VGCN embeddings, and similarity metrics such as keyword overlap and cosine similarity for final refined clustering. The scope of this research is to evaluate our approach by using an Adjusted Rand Index (ARI) score in skill classification on a dataset exhibiting a long-tail distribution across both the head and tail data, comparing the results to those of an initial Non-DAPT model, since, to the best of our knowledge, no direct approach exists that is comparable to either our ensemble model or the distribution of our dataset.
- Implementación de un plan de ventas y operaciones (S&OP)(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2025-11-25) Banderas Márquez, Adriana Valeria; Coronado Mondragón, Christian Etianne; mtyahinojosa, emipsanchez; Vázquez Hernández, Jesús; Escuela de Ingeniería y Ciencias; Campus MonterreyEste proyecto implementa un proceso estructurado de Sales and Operations Planning (S&OP) en Flexometal, empresa mexicana dedicada a la fabricación de tubería de acero, con enfoque en la línea galvanizada RSD, que representa el 32 % de sus ventas totales. Mediante un diagnóstico operativo y una revisión sistemática de literatura, se identificaron deficiencias en la planeación de la demanda, el manejo de inventarios y la coordinación entre ventas, operaciones y finanzas, lo que generaba altos costos, variabilidad en la producción y un nivel de servicio inferior al objetivo corporativo. A partir de estos hallazgos, se establecieron siete objetivos operativos, entre ellos: reducir costos en 10 - 14 %, disminuir inventarios en al menos 15 %, mejorar la precisión del pronóstico en 15–20 %, elevar el nivel de servicio entre 15–18 % y fortalecer la colaboración interdepartamental. El proceso implementado adaptó las cinco fases del S&OP al contexto de Flexometal, incorporando herramientas como pronósticos estadísticos, políticas de inventario, análisis financiero y una estructura mensual de reuniones colaborativas. Tras diez meses de ejecución, el modelo generó mejoras tangibles: reducción del MAPE de 31 % a 22 %, optimización del inventario en 19 %, incremento del nivel de servicio a 93.8 % y una disminución total de costos del 12 %, equivalentes a ahorros estimados de 1.6 millones de pesos anuales. El S&OP se consolidó como un sistema sostenible, replicable y orientado a la toma de decisiones basada en datos.
- Deep learning framework to predict and generate new fluorescent molecules from experimental data(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2025-07) Azizi, Mina; Aguirre Soto, Héctor Alán; emipsanchez; Ray, Mallar; Bernal Neira, David Esteban; Mendoza Cortés, José Luis; School of Engineering and Sciences; Campus Monterrey; Flores Tlacuahuac, AntonioFluorescent molecules play important roles in biological imaging, diagnostics, and materials science. However, identifying efficient and effective fluorophores remains challenging, as traditional trial-and-error experimentation and in silico computations are both costly and time-consuming. To address this, this thesis presents a deep learn- ing approach to streamline the discovery process by predicting optical properties and generating novel fluorescent molecules directly from experimental data. The study is based on FluoDB, a publicly available dataset collected from the literature, containing over 55,000 fluorophore–solvent pairs with experimentally measured optical prop- erties. Graph Convolutional Network (GCN) models were trained to predict four key optical properties and effec- tively captured complex structure–property relationships, achieving R² values ranging from 0.49 to 0.87 across the different targets. A Conditional Variational Autoencoder (CVAE) was also implemented to generate novel fluores- cent molecules based on solvent identity and target absorption range. In total, 2573 valid and structurally diverse molecules were generated, with a variety of predicted optical behaviors. Together, the predictive model and genera- tive models provide a useful and data-driven approach to accelerate exploration and design of functional fluorescent materials.
- Estrategia de transformación para comercio digital(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2025-03-28) Pastén Calderón, Francisco Javier; Navarro Tuch, Sergio Alberto; emipsanchez; Mora Vargas, Jaime; Bustamante Bello, Martín Rogelio; Escuela de Ingeniería y Ciencias; Campus Ciudad de MéxicoEsta tesis tiene como objetivo analizar y diseñar una estrategia integral de transformación digital para garantizar que la institución mantenga su competitividad y relevancia en el mercado actual, el cual es altamente competitivo. La transformación digital representa una oportunidad crucial para romper con el statu quo y reinventarse, lo que implica no solo la adopción de nuevas tecnologías, sino una reestructuración fundamental de las operaciones, los procesos y la organización del negocio para la generación de valor. La transformación digital ayuda a las empresas e instituciones a superar las limitaciones actuales mejorando la experiencia del cliente, agilizando los procesos, analizando los datos para hacer más eficiente la toma de decisiones, reduciendo los costos, y facilitando así las estrategias que pueden asegurar el éxito a largo plazo. Según la consultora McKinsey (McKinsey, 2024), lograr la transformación requiere diseñar una estrategia centrada en el valor, construir una sólida base de talento interno, desarrollar modelos operativos escalables, habilitar la tecnología distribuida, garantizar la accesibilidad y la gobernanza de los datos, y fomentar procesos sólidos de gestión y adopción del cambio. La tesis emplea varios marcos y metodologías aprendidas durante al maestría en materias como comportamiento organizacional, administración de la innovación tecnológica e introducción al comercio electrónico así como literatura adicional para complementar o actualizar la información, incluyendo “Build a Digital Commerce Tech Stack for Sales Success” (Luke Tipping, 2024), “Teams Topologies" (Pais)" para la evolución de las estructuras de equipo, "Swift de Tanzu Labs" (Smullen, 2024) para el desarrollo ágil y "Construcción de microservicios" para la transformación arquitectónica, para guiar la revisión digital. En el contexto de la industria del empeño en México, se destaca la necesidad de transformación debido a los cambios en el comportamiento de los clientes, el crecimiento de las fintechs como competencia, la reducción del tráfico peatonal en las sucursales y los cambios en los tipos de activos. Las recomendaciones estratégicas incluyen diversificar los productos de financiamiento, mejorar la experiencia del cliente a través de la digitalización, modernizar el proceso comercial, revitalizar las experiencias en las sucursales y liderar con tecnologías emergentes. Con la adopción de estas estrategias, la empresa pretende adaptarse a las tendencias del mercado y mantener su relevancia para ayudar a quienes más lo necesitan.
- An explainable autoencoder integrating regression and classification trees for anomaly detection(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2025) Caballero Dominguez, Zoe; Raúl Monroy Borja, Raúl; mtyahinojosa, emipsanchez; Graff Guerrero, Mario; García Ceja, Enrique Alejandro; González Mendoza, Miguel; Escuela de Ingeniería y Ciencias; Campus Estado de México; Medina Pérez, Miguel AngelAnomaly detection, or outlier detection, is a critical field since anomalies are data points that deviate from normal patterns and are used to represent critical information, such as fraud, diseases, or cyber-attacks. These applications are considered high-risk scenarios which involve high-stakes decision-making. Therefore, understanding the reasoning behind machine learning models used in this area has become an essential requirement. Despite its growing importance, explainable outlier detection remains a challenge since improving model accuracy while maintaining explainability creates a significant trade-off. Furthermore, anomaly detection models are mostly designed for one type of data, either numerical or categorical. This represents a disadvantage when both data types are present in the dataset's attributes, as real-world applications often contain, since transforming categorical values to numerical ones, or vice-versa, can produce information loss and reduced performance. In this thesis, we seek to address both challenges by proposing a novel explainable semi-supervised anomaly detection model that integrates classification and regression trees into an autoencoder architecture. We named our proposal: Explainable Outlier Tree-based Encoder (EOTE). EOTE is able to detect anomalies by creating a reconstruction of the input instance based on the relationships between attributes learned from normal samples. The harder it is for EOTE to reconstruct the instance correctly, the higher the probability of being an outlier is given to the instance. We evaluate EOTE against 12 anomaly detection and one-class classifiers across 110 datasets containing attributes of one data type (numerical or nominal) and a mix of both. Our experiments reveal that EOTE is one of the top-performing algorithms at detecting outliers in datasets with only numerical and nominal attributes, as well as datasets with mixed data attributes. Therefore, without sacrificing performance, EOTE is capable of producing interpretable outputs for its classification. This combination makes EOTE a suitable classifier for anomaly detection in high-risk applications.
- Priority-aware collision avoidance via optimal velocity in multi-robot systems(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2025) Sánchez Vaca, Luis Humberto; Sánchez Ante, Gildardo; mtyahinojosa, emipsanchez; Castañeda Cuevas, Herman; Hinojosa Cervantes, Salvador Miguel; Mercado Ravell, Diego Alberto; Escuela de Ingeniería y Ciencias; Campus Monterrey; Abaunza González, HernánThis thesis presents a decentralized control framework for prioritized multi-robot navigation that integrates Reciprocal Velocity Obstacles (RVO) with Bare-Bones Particle Swarm Optimization (BB-PSO). While velocity-based methods provide real-time geometric collision-avoidance guarantees, they often lead to oscillatory or conservative behaviors in dense environments and do not account for heterogeneous task priorities. On the other hand, optimization-based planners can shape agent behavior but lack inherent safety guarantees unless they are explicitly constrained. To address these limitations, the proposed framework combines two paradigms. First, RVO constructs a set of safe and admissible velocities. Then, BB-PSO selects the optimal velocity within this set based on a cost function that integrates priority-aware behaviors. This mechanism enables robots to navigate smoothly while respecting different task urgencies. Each robot independently computes its control command using local information about other agents, making this a fully decentralized operation. A simulation framework was developed to evaluate the proposed method across scenarios with different robot densities, priority distributions, and motion constraints. Experiments compare the hybrid controller against a greedy baseline using three metrics: arrival time, distance traveled, and collision occurrences. Results show that the hybrid approach improves navigation efficiency and significantly benefits high-priority agents by reducing their travel time and path deviation while maintaining safe interactions for the entire team. Overall, this thesis contributes a novel prioritized navigation strategy that combines geometric safety, real-time feasibility, and adaptive optimization. The approach represents a promising step toward scalable, priority-aware multi-robot systems that operate in complex and dynamic environments, with potential applications in automated warehouses, hospital logistics, and service robotics.
- A real-time FLL-based observer to enhance rotor-shaft motion sensing(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2025) Morazán Luz, José Raúl; Galluzzi Aguilera, Renato; mtyahinojosa, emipsanchez; Escobar Valderrama, Gerardo; Escuela de Ingeniería y Ciencias; Campus Ciudad de México; Ibarra Moyers, Luis MiguelReliable measurements of rotary machines' motion generally require the use of speed and position sensors. Conventional solutions, which are mechanically coupled to the rotor-shaft, are often considered expensive and are subject to degradation; despite their high accuracy. Low-cost alternatives provide an estimate of angular motion by sensing the magnetic field of a target fixed to the machine's shaft. Although these magnetic-based solutions have gained popularity due to their contactless feature and ease of installation under restricted dimensions, the use of linear magnetic sensors has been limited as their output signals are subject to distortion. This is relevant for alternate current machines' field-oriented control, as position error propagates towards the output torque, due to geometric transformation of coordinates over the stator phase currents. The published literature reveals that the most accurate and computationally efficient solution consists of a frequency-locked loop observer based on a fourth-order harmonic oscillator, typically used for grid synchronization, to enhance the signals of linear magnetic sensors for rotor-shaft motion sensing. However, aspects of its implementation over a real-time system have not been discussed; representing a research gap. Hence, in this thesis, the mentioned estimation method is implemented as a real-time algorithm, where it is evaluated in terms of its computational resource utilization through processor-in-the-loop tests, and both response time and accuracy through variable speed tests with a field-oriented controlled synchronous machine, using the motion estimates as feedback. As a reference, these characteristics are equally assessed by reproducing the mentioned experiments, for a quadrature encoder and the generic approach to process the output signals of linear Hall-effect sensors for motion sensing. Experimental results demonstrate that the proposed method, while requiring more computational resources, it is able to be executed over a microcontroller unit along with a field-oriented control strategy. After fine-tuning, the proposed observer is able to achieve a response time similar to the quadrature encoder, and improve the accuracy throughout the use of linear Hall-effect sensors; decreasing the angular speed estimate's stationary error under 19 % with respect to the quadrature encoder reference. Additionally, an attenuation over the dq-currents is achieved; reducing the mean dispersion in steady-state from 31.65 mA to 24.87 mA in the d-axis, and 26.52 mA to 11.4 mA in the q-axis. Thus, improving the performance of the driving machine. This statement is further supported through a harmonic analysis over the measured signals in steady state, where the proposed method demonstrated harmonic cancellation over components derived from the number of poles from the sensing array and the controlled motor.
- General aging model for schematic eyes(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2025) Fernandez Pozas, Pablo Antonio; Jaimes Nájera, Alfonso; emimmayorquin, emipsanchez; Gutierrez Vega, Julio Cesar; Peréz García, Benjamín de Jesús; Escuela de Ingeniería y Ciencias; Campus MonterreyA general methodology for modeling the aging process of the human eye is proposed. The two main changes that occur with age are: an increase in the dimensions of the crystalline lens an a flattening of its internal GRIN. These two processes are modeled respectively with a set of scaling factors and the introduction of a composite function called the plateau function. The methodology is put to test with three different schematic eye models: The Poisson Gauss, the Composite Modified Luneburg and the AVOCADO model. In each model the aging methodology was implemented and yielded satisfactory results in terms of: physiological fidelity, position of the focal plane and evolution of Zernike’s primary spherical aberration.
- Voice fraud mitigation: developing a deep learning system for detecting cloned voices in telephonic communications(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12-03) Villicaña Ibargüengoyti, José Rubén; Montesinos Silva, Luis Arturo; emimmayorquin; Santos Díaz, Alejandro; Mantilla Caeiros, Alfredo Víctor; School of Engineering and Sciences; Campus Ciudad de MéxicoThis study addresses the increasing threat in recent years of voice fraud by cloned voices in phone calls. This problem can compromise personal security in many aspects. The primary goal of this work is to develop a deep learning-based detection system for distinguishing between real and cloned voices in Spanish, focusing on calls made over telephone lines. To achieve this, a dataset was generated from real and cloned audio samples in Spanish. The audios captured were simulated under various telephone codecs and noise levels. Two deep learning models, a convolutional neural network (which in this project is named Vanilla CNN) and a transfer learning (MobileNetV2) approach, were trained using spectrograms derived from the audio data. The results indicate a high accuracy in identifying real and cloned voices, reaching up to 99.97% accuracy. Also, many validations were performed under different types of noise and codecs included in the dataset. These findings highlight the effectiveness of the proposed architectures. Additionally, an ESP32 audio kit was integrated with Amazon Web Services to implement voice detection during phone calls. This study contributes to voice fraud detection research focused on the Spanish language.
- Enhancing BGP security with MAD anomaly detection system and machine learning techniques(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12) Romo Chavero, María Andrea; Cantoral Ceballos, José Antonio; emipsanchez; Botero Vega, Juan Felipe; Navarro Barrón, Francisco Javier; School of Engineering and Sciences; Campus Monterrey; Pérez Díaz, Jesús ArturoAnomalies in the Border Gateway Protocol (BGP) represent a signicant vulnerability in the Internet’s infrastructure, as they can cause widespread disruptions, trafc misdirection, and even security breaches. Proactive detection of these anomalies is vital to preserving network stability and preventing potential cyberattacks. In response to this challenge, we present the Median Absolute Deviation (MAD) anomaly detection system, which combines traditional statistical methods with advanced machine learning (ML) techniques for more precise and dynamic detection. Our approach introduces a novel adaptive threshold mechanism, allowing the system to adjust based on the changing conditions of network trafc. This dynamic thresholding signif- icantly improves the accuracy, precision, and F1-score of anomaly detection compared to the previous xed-threshold version. Additionally, we integrate the MAD system with a diverse ML classiers, including Random Forest, XGBoost, LightGBM, CatBoost, and ExtraTrees to enhance the system’s ability to identify complex patterns that indicate unusual BGP behavior.We evaluate our detection system on well-documented BGP anomaly events, such as the Slammer worm, Nimda, Code Red 1 v2, the Moscow blackout, and the Telekom Malaysia misconguration. The results show that our system when combined with ML models achieves an overall accuracy and F1-score of 0.99, demonstrating its effectiveness across various anomaly types. By using both statistical and ML models, the system is able to capture irregularities that could signal security threats, offering a more comprehensive detection solution.This research highlights the importance of combining statistical anomaly detection with ML to obtain a balance between accuracy and computational efciency. The system’s low resource requirements and minimal pre-processing make it highly scalable, allowing it to be potentially deployed in real-time on large-scale networks.