With the outbreak of the SARS-CoV-2 o COVID-19 pandemic, multiple studies of risk factors and their influence on patient deaths have been developed. However, little attention is often paid to analyzing patients in risk groups despite the fact that they have been infected and inpatients can survive. In this article, with the dataset available from the Ministery of the health of Mexico, this paper proposes the use of the latent topic extraction algorithm Latent Dirichlet Allocation (LDA) for the study of COVID-19 survival factors in Mexico. The results let us conclude that in the year before strategies for prevention and control of COVID-19, the latent topics support that patients without comorbidities have a low risk of death, compared with the period of 2021, wherein in spite of having some risk factors patients can survive.

Deep learning has achieved remarkable performance in skin lesion classification, but its lack of interpretability often remains a critical barrier to clinical adoption. In this study, we investigate the spatial properties of saliency-based model explanations, focusing on symmetry and other morphometric features. We benchmark five deep learning architectures (ResNet-50, EfficientNetV2-S, ConvNeXt-Tiny, Swin-Tiny, and MaxViT-Tiny) on a nine-class skin lesion dataset from the International Skin Imaging Collaboration (ISIC) archive, generating saliency maps with Grad-CAM++ and LayerCAM. The best-performing model, Swin-Tiny, achieved an accuracy of 78.2% and a macro-F1 score of 71.2%. Our morphometric analysis reveals statistically significant differences in the explanation maps between correct and incorrect predictions. Notably, the transformer-based models exhibit highly significant differences ($p<0.001$) in metrics related to attentional focus (Entropy and Gini), indicating that their correct predictions are associated with more concentrated saliency maps. In contrast, convolutional models show less consistent differences, and only at a standard significance level ($p<0.05$). These findings suggest that the quantitative morphometric properties of saliency maps could serve as valuable indicators of predictive reliability in medical AI.

COVID-19 has become the most significant pandemic in recent years. Today, Mexico has recorded millions of infections and deaths since the pandemic started. Around the world, machine learning methods have been used to understand, predict or develop strategies to manage the virus and the pandemic. Although algorithms provide good results, it is necessary to understand why a model makes specific predictions with a particular data set. To explain this question, we apply Explainable Artificial Intelligence (XAI) in this paper. With this, it is possible to understand the characteristics that influence the model decisions when denoting between deaths and survivors. As a case of study, the positive cases detected during the winter season of 2020-2021 and 2021-2022 were considered. In this season, respiratory diseases increased considerably, and in the study period, they influenced the increase in positive cases and the spread of COVID-19. Preliminary results suggest that age is essential when using a Random Forest model. Preliminary results suggest that age is essential when determining the prognosis of a patient infected by COVID-19 in winter seasons.