A computational approach for early breast cancer detection using Dynamic Infrared Thermography (DIT) was developed. Thermograms are represented by multivariate time series extracted from thermal hotspots in the breast, capturing five features: maximum and mean temperature, spatial heterogeneity, heat flux, and tumor depth, over 20 thermograms. Features are estimated based on the inverse solution of the Pennes bio-heat equation. Classification is performed using a Time Series Forest (TSF) and a Long Short-Term Memory (LSTM) network. The TSF achieved an accuracy of 86%, while the LSTM reached 94% accuracy. These results indicate that dynamic thermal responses under cold-stress conditions reflect tumor angiogenesis and metabolic activity, demonstrating the potential of combining multivariate thermographic sequences, biophysical modeling, and machine learning for non-invasive breast cancer screening.

Didactic control systems are tools that facilitate the understanding of topics related to control and stabilisation. However, due to the high cost of these devices, their implementation in schools is limited. This work presents the design, construction, and testing of a low-cost didactic prototype of a Furuta-type pendulum for the analysis of topics related to this system. The construction included parts designed in SolidWorks and produced with 3D printing, threaded rods, a precision linear potentiometer, a DC motor, a BTS7960 H-bridge, and an Arduino Nano. The functionality of the project was verified through the implementation of two digital tools: a Kalman filter to improve the measured signal and a PID controller to regulate the angular speed of the motor. It was demonstrated that the prototype can be stabilised under the operational considerations of PID control without the use of an angular position sensor. Moreover, the Kalman filter must be kept small in order to avoid information loss.

Las noticias falsas se crean con el objetivo de manipular, danar o desinformar. En los últimos años, este tipo de noticias ha impactado negativamente en diferentes sectores de la sociedad, como son: política, salud y movimientos sociales. El problema aumenta debido a su amplia y veloz propagacion, que supera con creces la rapidez en la que un humano puede controlar o contener este fenomeno. La búsqueda de soluciones para detener la difusión de información falsa ha motivado el desarrollo de métodos computacionales para su deteccion automática. Comúnmente, tales enfoques son diseñados desde una perspectiva del procesamiento de lenguaje natural. Particularmente, en este trabajo se estudia el impacto del uso de diversas representaciones de características del contenido de la noticia para la deteccion de noticias falsas en Español con técnicas de aprendizaje automático, incluyendo arquitecturas profundas.

In this study, a numerical analysis of the mathematical model of a single-phase transformer is presented, incorporating a comprehensive expression for the mutual magnetic flux to assess its effect on energy transfer efficiency. Drawing on Faraday’s and Ampère’s laws, the fundamental relationships between magnetic flux, currents, and voltages in the windings are established. Through numerical simulations, the influence of mutual leakage flux is investigated, alongside structural parameters such as resistance, reactance, and the number of turns. The results obtained confirm the model’s effectiveness in predicting transformer behaviour under varying load and network conditions, offering an analytical tool for the design and enhancement of these electrical machines. This approach not only deepens the theoretical understanding of transformers but also supports the development of more efficient and reliable electrical systems.

This project details the implementation of an artificial neural network (ANN) as the principal element of a system engineered to identify and predict future control states of a thrust-propelled seesaw in an open-loop configuration. The primary objective was to maintain the seesaw in a balanced 90° position. The system’s dynamic behaviour was analysed under minimal external disturbances, facilitating development and evaluation in a controlled environment. Experimental data were captured via a programmable Arduino UNO board transmitting over the serial port and recorded in an Excel file for subsequent processing. A Kalman filter was applied to refine the data, from which a random subset was selected to train the neural network. A comprehensive analysis of the results is presented herein, demonstrating the ANN’s satisfactory performance in the control task.