The development of mathematical thinking requires much more than the acquisition of content; it involves the activation of complex cognitive processes that allow us to analyze, plan, solve, and reflect on problematic situations. In this context, executive functions—such as working memory, cognitive inhibition, and mental flexibility— play a fundamental role in facilitating self-regulation, strategic decision-making, and adaptation to new problem solving conditions. This article analyzes the link between executive functions and mathematical thinking, highlighting how strengthening them contributes to deeper understanding, logical reasoning, and the ability to model real-life situations. A didactic integration of these functions in the classroom is proposed, based on institutional competencies, to promote meaningful learning and the development of higher-order cognitive skills. The proposed research methodology was qualitative, using a controlled trial method.
The classification of breast cancer using machine learning techniques has become a critical tool in modern medical diagnostics. This study analyzes the performance of hybrid models that combine traditional machine learning algorithms (TMLAs) with a convolutional neural network (CNN)-based VGG16 model for feature extraction to improve accuracy for classifying eight breast cancer subtypes (BCS). The methodology consists of three steps. First, image preprocessing is performed on the BreakHis dataset at 400× magnification, which contains 1820 histopathological images classified into eight BCS. Second, the CNN VGG16 is modified to function as a feature extractor that converts images into representative vectors. These vectors constitute the training set for TMLAs, such as Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Naive Bayes (NB), leveraging VGG16’s ability to capture relevant features. Third, k-fold cross-validation is applied to evaluate the model’s performance by averaging the metrics obtained across all folds. The results reveal that hybrid models leveraging a CNN-based VGG16 model for feature extraction, followed by TMLAs, achieve accuracy outstanding experimental accuracy. The KNN-based hybrid model stood out with a precision of 0.97, accuracy of 0.96, sensitivity of 0.96, specificity of 0.99, F1-score of 0.96, and ROC-AUC of 0.97. These findings suggest that, with an appropriate methodology, hybrid models based on TMLA have strong potential in classification tasks, offering a balance between performance and predictive capability.
30 December 2024 in Journal of Systems and Educational Management
The implementation of problem-based learning allows students to develop the ability to recognize problems, analyze them, establish solutions, and present results. The main interest of this work is to determine the variables that currently influence the implementation of problem-based learning as the guiding axis of a course and to establish a proposal to continue using it, given that the conditions of the student community have changed with everyday problems. The proposed research methodology was qualitative due to the controlled test method. applied to three groups of first semester, to identify existing problems on poor school performance, adaptation to collaborative work, reading comprehension, mathematical reasoning. With the results obtained in the diagnosis, improvement strategies were established and throughout the semester accompaniment was provided to the detected students, these strategies were modified according to the pre-results that were obtained.
This work implements a stable algorithm in Field-Programmable Gate Arrays (FPGAs) for which two architectures (unrolling the loop) were developed and analyzed. The algorithm recovers sources located at the boundary separating two homogeneous media that make up a two-dimensional non-homogeneous region from measurements on the boundary of such region. The problem of recovering these sources is an ill-posed inverse problem as small errors in the measurement can produce important changes in the source location. Inverse source problems have many applications in different areas, such as engineering and medicine, making the proposed implementation important. The first architecture (mode one) allows considering different operating speeds, which is an advantage depending on whether we work with fast or slow signals. The second one (mode two) reduces resource consumption by exploiting the characteristics of the source identification algorithm, which is an advantage for multichannel problems such as inverse electrocardiography or electroencephalography. The architectures were tested on four FPGAs of the 7 Series of Xilinx: Spartan-7 xc7s100fgga484, Virtex-7 xc7v585tffg1157, Kintex-7 xc7k70tfbg484, and Artix-7 xc7a35tcpg236. The two FPGA implementations of the algorithm were validated using synthetic examples implemented in MATLAB. The results presented here can be extended to concentric spheres and complex geometries.
Objective: This work presents an implementation of a stable algorithm that recovers sources located at the boundary separating two homogeneous media in field-programmable gate arrays. Two loop unrolling architectures were developed and analyzed for this purpose. This inverse source problem is ill-posed due to numerical instability, i.e., small errors in the measurement can produce significant changes in the source location. Methodology: To handle the numerical instability when recovering these sources, the Tikhonov regularization method in combination with the Fourier series truncation method are applied in the stable algorithm. This stable algorithm is implemented in two different architectures developed in this work: The first architecture (Mode 1) allows for different operating speeds, which is an advantage depending on whether we work with fast or slow signals. The second one (Mode 2) reduces resource consumption by exploiting the characteristics of the source identification algorithm, which is an advantage for multichannel problems such as inverse electrocardiography or electroencephalography. Results: The architectures were tested on four devices of the 7 Series of Xilinx: Spartan-7 xc7s100fgga484, Virtex-7 xc7v585tffg1157, Kintex-7 xc7k70tfbg484, and Artix-7 xc7a35tcpg236. The two hardware implementations of the stable algorithm were validated using synthetic examples implemented in MATLAB, which shows the advantages of each architecture. Contributions: We developed two efficient architectures based on a loop unrolling design for source identification problems. These are effective strategies to divide and assign tasks to the configurable hardware, and they appear as an appropriate technique for implementing the algorithm. The first one is simple and allows for different operating speeds. The second one uses a control system based on multiplexors that reduce resource consumption and complexity of the design and can be used for multichannel problems. From the numerical test, we found the regularization parameters. The synthetic examples developed here can be considered for similar problems and can be extended to concentric spheres.