Programación Matemática y Software

User Interfaces Flow: Modeling practical cases using User Interface Transition Diagrams

Jorge Cervantes-Ojeda — 2025-10-02

Here, we explore the application of the User Interface Transition Diagram (UITD) for modeling the flow between user interfaces in interactive software systems. While UITDs are known tools in requirements elicitation, our work addresses the gap in showing their effectiveness in modeling three common user interaction scenarios that are frequently encountered in practice: (i) actions repeated across multiple interfaces, (ii) actions available only to users with extended privileges, and (iii) dynamically enabled or disabled buttons based on conditional logic. These solutions not only make it easier for customers to understand the model with the user actions and system responses but also facilitate a smooth transition from requirements specification to design phases by precisely defining user interactions. Despite the uniqueness of each software system, many share common traits that can be effectively modeled using UITDs. By offering generic solutions for these scenarios, this work aims to enhance the modeling capabilities of UITDs, promoting their broader adoption in the software industry.

Tools for Addressing the Fundamentals and Counting Techniques in Probability

María Cristina Medel López — 2025-10-02

Probability and Statistics are two areas of research and application of applied mathematics. They are useful in different fields of science, such as formal and experimental science, technology (design, development and monitoring of technological projects) and even in the industrial and productive sector (quality and maintenance of components or systems), which explains the relevance of their teaching, as well as their learning. However, during high school and higher education, it is tedious and even complicated to apply theories to solve practical problems in each of the areas mentioned. The first obstacle that students face is that they do not know how to count, which means that they do not know how to use counting techniques to solve probability problems. In this work, different problems are discussed which will allow students to better understand the importance of this area in their probability courses. In addition, the use of free software-type mathematical tools such as WolframAlpha, GeoGebra and Excel is proposed, to support the understanding of the use of counting techniques in probability theory.

Numerical Analysis with the FDTD Method to Study the Laser Effect in a 3D System with Disorder and Experiment with ZrO2:Yb Nanocrystals

Jesús Jonathan Martínez-Ocampo — 2025-10-02

In this work, the optical field and the mirrorless laser effect generated in a 3D disordered system containing nanoemitters are studied. The studied system is non-linear, complex and has no analytical solution. The objective is to study the laser effect numerically and experimentally with percolation, which highlights the importance of approaching the study numerically using the Finite Differences with Time Domain (FDTD) Method. To complement the results of the numerical analysis, an experiment is carried out in which Yb nanoemitters are incorporated into ZrO2. These nanoemitters are distributed in a disorderly manner in the system and are non-coherently excited by an external source. The absorption and emission spectrum of ZrO2:Yb is obtained for different concentrations, and the experimental results confirm the theoretical predictions made.

Applying Machine Learning methodologies to enhance trading decisions in cryptocurrency assets

Víctor Leonardo Teja Juárez — 2025-10-02

Cryptocurrency trading involves the buying and selling of digital assets, such as Bitcoin (BTC) and Ethereum, with the aim of obtaining financial gains through specialized platforms known as exchanges. The relevance of this practice lies in its ability to capitalize on the notable market volatility, allowing for significant returns. This study focuses on the application of machine learning algorithms for strategic decision-making in the cryptocurrency realm, with a particular emphasis on sentiment analysis derived from Reddit.com posts to evaluate market perception. The inherent volatility of the cryptocurrency market, along with psychological influences and information asymmetries, underscores the importance of sentiment analysis for predicting price movements and optimizing trading strategies. This analysis classifies sentiment into positive, negative, or neutral categories, thereby guiding trading decisions. Additionally, a recurrent neural network is employed to predict BTC prices using historical data, complementing the sentiment analysis. The evaluation of technical indicators allows for identifying the optimal time to operate in the market, and backtesting reveals notable returns, especially in BTC with 49.88%, Ethereum (38.74%), Binance Coin (32.89%), Cardano (29.74%), and Solana (27.64%). The study demonstrates that machine learning models offer accurate predictions and reduce biases compared to traditional trading platforms. Nonetheless, the need for continuous adaptation and diversification is highlighted due to market volatility and regulatory uncertainties. Future research is suggested to focus on testing strategies across various cryptocurrencies and consulting with financial experts to mitigate risks and enhance investment outcomes.

Application of micro-genetic algorithms to the hyperparameter optimization in classification methods

Antonio Guerrero Juárez — 2025-10-04

This study proposes the use of micro-genetic algorithms as a hyperparameter optimization technique to improve the accuracy and efficiency of certain classification methods. Four models were initially evaluated without hyperparameter optimization, and subsequently, a specifically designed micro-genetic algorithm was applied to fine-tune their hyperparameters. The objective was to analyze the impact of this technique on improving classification accuracy. The results demonstrated that the implementation of micro-genetic algorithms not only significantly increased the accuracy of the classification methods but also reduced training time, indicating improved efficiency. These findings suggest that micro-genetic algorithms can be an effective tool for optimizing the performance of classification methods and solving classification problems with greater accuracy and speed.

Block-based fragile watermarking for tamper detection and recovery content of digital images

Elizabeth Campos Ponce — 2025-10-05

Nowadays the internet as well as advances in information and communication technologies, has facilitated the transmission and storage of multimedia data, such as digital images. This has brought with it some disadvantages, such as copyright infringements, as well as edits and/or alterations to the content that compromise the authenticity and integrity of multimedia files. To address these issues, various solutions have been proposed in scientific literature, including steganography, data encryption, and digital watermarking techniques. This article proposes a method based on fragile watermarking that enables the detection and recovery of content in grayscale images when it has been altered and/or modified. Initially, the image is segmented into four quadrants, which correspond to each other diagonally. Using an additional internal segmentation in the form of blocks within each quadrant and applying average calculations in conjunction with the extraction of the most significant bits, watermarks are generated and embedded into the pixels of each block using the least significant bit substitution technique. To increase the security of the proposed method, a secret key is used to permute the insertion index in each quadrant. Experimental results demonstrate that the proposed method offers high imperceptibility in terms of the structural similarity index and peak signal-to-noise ratio. Regarding the detection capacity for modifications, alterations such as copy-move, copy-paste, and random cropping, with efficiency measured in terms of false positive rates, false negative rates, and alteration detection rates. Lastly, content restoration results are presented, with fidelity evaluated using the same imperceptibility metrics.