La deserción universitaria afecta negativamente a muchos estudiantes, este suceso puede estar relacionado con problemas personales, cuestiones económicas, entre otros. Ante tal situación surge la importancia de desarrollar una forma de predecir estos casos, para esto se propuso el uso de técnicas de Machine Learning, las utilizadas fueron Regresión Logística, Naive Bayes, Red Neuronal Perceptrón Multicapa, Árbol de Decisión, Support Vector Machine y Random Forest; se seleccionó un Dataset, que pasó por una limpieza de datos, se corrigieron los datos faltantes y los valores atípicos; luego se eliminaron los registros cuya variable de salida era Matriculado, centrándose en los tipos Abandono y Graduado. Cada modelo fue entrenado y probado mediante validación cruzada con pliegues, finalmente, se compararon en función de métricas de precisión, exactitud y exhaustividad, donde se concluyó que la Regresión Logística es la técnica que mejores resultados proporciona para predecir la deserción universitaria en el dataset considerado.

In recent years, the need for analytics on large volumes of data has become increasingly important. It turns out to be extremely useful in making strategic decisions about different applications. In this way, appropriate mechanisms must be designed to carry out data processing and integration with different platforms to take advantage of their best features. In this work, an architecture that works on cloud services is shown to migrate data stored in Big Query to an analytics engine such as Elasticsearch and take advantage of its potential in query, insert and display operations. This is accomplished through the use of Cloud Functions and Pub / Sub. The integration of these platforms through the proposed architecture showed 100% effectiveness when transferring data to another, maintaining an insertion rate of 4,138.30 documents per second, demonstrating its robustness, efficiency, and versatility when performing a data migration. This pretends to establish an architecture solution when it comes about handling a large amount of data as in the real world.

Este artículo presenta la posibilidad de trabajar en el desarrollo de software utilizando no solo metodologías ágiles para los procesos, sino también el uso de Design Thinking para la planificación del Plan de Pruebas del Software. Ya que durante el proceso de gestión del software existe la dificultad de comprenderlo, para una de las etapas de la gestión de software, en este caso la planificación de las pruebas del software, el que ejecuta las pruebas necesita saber lo que el usuario requiere, debido a que el producto de software es intangible y difícil de medir, el que realiza las pruebas debe escuchar al usuario final y realizar las pruebas de acuerdo a los requerimientos que se solicita. Para mejorar el proceso de comprensión de lo que el usuario realmente requiere, se utiliza sus pilares como: empatía, colaboración y creación de prototipos para una mayor interpretación, la cual permitirá obtener una actitud asertiva al equipo de pruebas. Por otro lado, tenemos que realizar el plan de pruebas del software, en base a esto se realizó un estudio sobre la posibilidad de utilizar Design Thinking para ayudar al proceso de planificación de pruebas de software.

Stakeholder classification is carried out manually using methods such as brainstorming, interviews with experts, and checklists. These methods present a subjective character as they depend on the appreciation of the interviewees. This characteristic affects the accuracy of this classification, making that the project managers do not make the correct decisions. The research aims to suggest a fuzzy inference system for the classification of stakeholders, which will improve the quality of such classification in the projects. The proposal carries out the machine learning and the adjustment of the fuzzy inference system to classify the stakeholders by executing four algorithms based on artificial neural networks: ANFIS, HYFIS, FS.HGD, and FIR.DM. It analyzes the results of applying them in 10 iterations by calculating the measures: percentage of correct classifications, false-positive cases, false-negative cases, and mean square error. The ANFIS system show the best results. The fuzzy inference system for stakeholder classification generated improves the quality of this classification using machine learning, allowing to make better decisions in a project.

Software requirements are the most critical phase focused on documenting, eliciting, and maintaining the stakeholders’ requirements. Risk identification and analysis are preemptive actions designed to anticipate and prepare for potential issues. Usually, this classification of risks is done manually, a practice that the personal judgment of the risk analyst or the project manager might influence. Machine learning (ML) techniques were proposed to predict the risk level in software requirements. The techniques used were logistic regression (LR), multilayer perceptron (MLP) neural network, support vector machine (SVM), decision tree (DT), naive bayes, and random forest (RF). Each model was trained and tested using cross-validation with k-folds, each with its respective parameters, to provide optimal results. Finally, they were compared based on precision, accuracy, and recall metrics. Statistical tests were performed to determine if there were significant differences between the different ML techniques used to classify risks. The results concluded that the DT and RF are the techniques that best predict the risk level in software requirements.