In 2020, there were more than 1.2 million new skin cancer diagnoses, and melanoma was the most recurrent type of cancer. On the other hand, melanoma is the least common but most serious form of skin cancer affecting both men and women. This work aims to assemble classification models to detect a case of melanoma with high accuracy based on a Convolutional Neural Networks system. The methodology considers training 21 models for image classification, with the best assembly performance of EfficientNet and VGG-19 architectures, the data augmentation technique was used to the images to improve its performance. The results show 92.85% of accuracy, 71.50% of sensitivity, and 94.89% of specificity, with an improvement of 0.06% in accuracy and specificity. The assembly of the classification models achieved higher accuracy in melanoma skin cancer image classification.

The objective of this research is to reduce the dropout rate of students in the Faculty of Systems Engineering and Informatics of the Universidad Nacional Mayor de San Marcos – FISI-UNMSM, through the implementation of an intelligent system with a data mining approach and the autonomous learning algorithm (decision trees) that predicts which students are at risk of dropping out. It was developed in Python and the free software Weka, for this purpose student data was collected from 2014 to 2020. This solution increases the availability and the level of satisfaction of the faculty; in the learning process, an accuracy percentage of 90.34% and precision of 95.91% was obtained, so the data mining model is considered valid. In addition, it was found that the variables that most influenced students in making the decision to abandon their studies were the historical weighted average, the weighted average of the last cycle and the number of credits passed.

The timely diagnosis of brain tumors is currently a complicated task. The objective was to build an image classification model to detect the existence or not of brain tumors by adding a classification header to a ResNet-50 architecture. The CRISP-DM methodology was used for data mining. A dataset of 3847 brain MRI images was used, 2770 images for training, 500 for validation, and 577 for testing. The images were resized to a 256 × 256 scale and then a data generator is created that is responsible for dividing pixels by 255. The training was performed and then the evaluation process was carried out, obtaining an accuracy percentage of 92% and a precision of 94% in the evaluation process. It is concluded that the proposed CNN model composed of a head with a ResNet50 architecture and a seven-layer convolutional network achieves adequate accuracy, becoming an efficient and complementary proposal to other models developed in previous works.

The prediction of students' academic performance is an area of great concern for universities and educational institutions since academic performance is one of the most important aspects of the learning process. To analyze this behavior, this study makes a critical analysis of the topic of interest and aims to review, analyze and summarize the latest research advances related to the prediction of academic performance. The systematic literature review method is applied to answer three questions: (1) What factors are determinants in predicting students' academic performance? (2) what methods are used to predict students' academic performance? and (3) what are the objectives and interests in predicting students' academic performance? After conducting the study of 50 outstanding articles, as results, we found that academic factor is the guideline for predicting academic performance; supervised machine learning is the most used technique, highlighting support vector machine, random forests, and neural networks; the most outstanding objectives for the application of prediction were: Student performance with 53%, risk of failure with 14%, search for student knowledge with 12%, avoid dropout with 12%, and decision making with 10%.

Neurodegenerative disorders, notably Alzheimer’s, pose an escalating global health challenge. Marked by the degeneration of brain neurons, these conditions lead to a gradual decline in nerve cells. Worldwide, over 55 million people grapple with dementia, with Alzheimer’s prominently impacting the aging demographic. The primary hurdle to early Alzheimer’s detection is the widespread lack of awareness. The main goal is to design and implement an artificial intelligence system using deep learning (DL) to detect Alzheimer’s disease (AD) through medical images and classify them into various stages, such as non-demented, moderate dementia, mild dementia, and very mild dementia. The dataset contains 6400 magnetic resonance images in .jpg format, with standardized dimensions of 176 × 208 pixels. To demonstrate the advantages of data augmentation and transformation techniques, four scenarios were created: two without these techniques, utilizing the Adam and SGD optimizers, and two with these techniques, also employing the Adam and SGD optimizers, respectively. The main results revealed that scenarios utilizing these techniques exhibited more stable performance when validated with a new dataset. Scenario 3, using the Adam optimizer, achieved a weighted average accuracy of 91.83%, whereas scenario 4, employing the SGD optimizer, reached 87.58% accuracy. In contrast, scenarios 1 and 2, which omitted these techniques, obtained low accuracies below 55%. It is concluded that classifying AD with a DL model exceeding 90% accuracy is feasible. This is the importance of utilizing data augmentation and transformation techniques to improve generalizability to input image variations, which is a consistent factor in the healthcare sector.