Producción científica y docente
Producción científica reciente
Does Bigger Mean Better? A Domain-Aware Comparison of Lexicon-Based, TF-IDF, and Transformer Models for Sentiment Analysis
Sentiment analysis systems range from training-free lexicons to compact supervised models and fine-tuned transformer encoders, but fair comparison requires identical test examples and explicit computational accounting. (1) Background: We reassess three sentiment-classification paradigms under matched binary-label evaluation on IMDb movie reviews and Sentiment140 tweets. (2) Methods: After duplicate and label audits, VADER, TF–IDF with validation-tuned Logistic Regression, and three-epoch GPU-fine-tuned DistilBERT were evaluated on the same 5065 IMDb and 10,000 Sentiment140 held-out records within each corpus. Accuracy, macro-F1, confusion matrices, paired McNemar tests, bootstrap intervals, training time, and inference throughput were computed from saved predictions and execution logs. (3) Results: DistilBERT reached 92.00% accuracy on IMDb and 83.41% on Sentiment140, compared with 91.06% and 78.77% for Logistic Regression and 69.83% and 66.54% for VADER. DistilBERT exceeded Logistic Regression by 0.95 percentage points on IMDb (paired 95% CI 0.08–1.76; p=0.0351 ) and by 4.64 points on Sentiment140 (CI 3.85–5.41; p<0.001 ), but required GPU fine-tuning and longer measured training. (4) Conclusions: The relative advantage of a transformer depends on the dataset and available computation. Logistic Regression remains a compact, accurate alternative; results are restricted to these two within-domain configurations and do not demonstrate cross-domain transfer or domain causation.
Rethinking the development of teacher autonomy: the heteronomy of teaching as a relational perspective on self-determined learning
Introduction: Teacher education often promotes professional autonomy while assessing teachers' compliance with predefined competencies. However, less attention has been paid to the relational conditions that enable beginning teachers to develop independent professional judgment. This study examines how newly qualified teachers describe the relationships, institutional experiences, and learning practices that shape their professional development. It introduces teacher heteronomy as a conceptual perspective on how interaction with others contributes to the formation of autonomous pedagogical judgment.Methods: The study employed an exploratory mixed-methods design within the Erasmus+ DigitalTA project (2023–2024). Data were collected through an online questionnaire completed by 327 teachers from Spain and 15 countries or territories in the Americas, and a focus group involving six early-career teachers in Spain. Survey data were analysed using descriptive statistics, while open-ended responses and focus-group discussions underwent inductive categorisation and thematic analysis. Findings from both sources were compared.Results: Participants identified integration into teaching teams as their main expectation of school leadership (68.2%). Communication difficulties with colleagues (44.3%) and lack of confidence (44.0%) were common challenges. University studies (43.7%) and postgraduate education (41.3%) were frequently reported learning sources, while 30.9% highlighted learning in the workplace as a key source of professional development and self-evaluation was the most selected reflective strategy (72.8%). Pedagogical and socio-emotional competences were prominent priorities. Participants also called for more school-based practice, updated curricula, mentoring, and stronger professional support.Discussion: The findings support a relational understanding of professional learning, highlighting the importance of colleagues, mentors, institutional experiences, and reflective practices in beginning teachers' development. The article proposes teacher heteronomy as a conceptual framework in which engagement with others contributes to autonomous professional judgment, rather than implying dependence or obedience. In dialogue with heutagogy, this perspective suggests strengthening initial teacher education through supervised placements, mentoring, collaborative inquiry, and opportunities for self-determined learning. Longitudinal research is needed to assess these relationships empirically.
SPE-ECG: An Ontology-Guided Knowledge Representation Framework for Enhancing LLM-Based Code Generation
Large Language Models (LLMs) can translate natural-language instructions into code, but when the input is unstructured clinical protocol text, the generated code may omit doses, monitoring intervals, or safety thresholds that the text states. Such omissions matter because they can propagate into software that supports clinical monitoring. We propose SPE-ECG (Structured Prompt Engineering for Enhanced Code Generation), a transparent, rule-based pipeline that converts unstructured domain text into a JSON prompt using text cleaning, part-of-speech tagging, bigram matching, and hand-built domain dictionaries, without fine-tuning any model. In a case study of 43 rheumatoid arthritis DMARD monitoring protocols, we compared raw and structured prompts on three LLMs (GPT-3.5, LLaMA 2, and Gemma), scoring the generated Python code using version 3.15.1 with BLEU against the raw-prompt code of the other two models. Mean BLEU rose for all three models (relative gains of 55.1%, 32.5%, and 19.9%); the paired differences were significant for GPT-3.5 and LLaMA 2 but not, after correction for multiple comparisons, for Gemma. These results suggest that lightweight input structuring can change how closely different models’ outputs agree, and that it could serve as a low-cost, auditable preparation step for protocol-to-code tools built on free or open models. BLEU does not measure correctness, no ablation or state-of-the-art baseline was run, and clinical validation is required before any use.
Cross-Dataset Generalization of Deep Learning for Retinal Vessel Segmentation in Fundus Images: External Evaluation from DRIVE to CHASE_DB1
Automated retinal vessel segmentation can support quantitative analysis of fundus photographs and facilitate measurement of vascular characteristics associated with ocular and systemic disease. Although deep-learning methods achieve high segmentation performance on benchmark datasets, their reliability may decrease when applied to images obtained under different acquisition conditions. This study evaluated the cross-dataset generalizability of a U-Net segmentation architecture incorporating an ImageNet-pretrained ResNet34 encoder. Sixteen images from the DRIVE dataset were used for model training and four were retained for in-domain validation. Images were resized to 256 × 256 pixels and processed using a consistent normalization pipeline, while spatial and photometric augmentation was applied during training. Without target-domain retraining or domain adaptation, the trained model was subsequently evaluated on the independent CHASE_DB1 dataset. Segmentation performance was assessed using the Dice coefficient, intersection over union (IoU), and sensitivity, together with qualitative examination of predicted vessel maps. On the held-out DRIVE images, the model achieved a Dice coefficient of 0.7263, IoU of 0.5703, and sensitivity of 0.7201. On CHASE_DB1, performance decreased to 0.6269, 0.4573, and 0.5502, respectively. The disproportionately larger reduction in sensitivity indicated an increased frequency of missed vessel pixels under cross-dataset evaluation. Qualitative examination showed greater fragmentation and omission of fine and peripheral vascular structures in CHASE_DB1 images. These findings demonstrate that within-dataset performance may overestimate the robustness of retinal vessel segmentation models when transferred to a different imaging domain. External evaluation should therefore form an important component of retinal imaging artificial intelligence research. Future work should investigate higher-resolution processing, vessel-width-specific evaluation, multi-domain training, domain-adaptation strategies, and larger multicentre datasets to improve model generalizability.
Cost-sensitive random forest with adaptive threshold optimization for imbalanced financial fraud detection
Background: Financial transaction fraud detection is a challenging binary classification problem characterized by severe class imbalance, rare fraudulent cases, and asymmetric misclassification costs. In real-world banking systems, failing to detect fraudulent transactions can lead to substantial financial losses, whereas false alarms mainly increase operational review costs. Therefore, fraud detection models should not be optimized only for overall accuracy, as accuracy-oriented classifiers may favor the majority non-fraud class and fail to identify minority fraud cases effectively. To address this issue, this study proposes a cost-sensitive ML framework for financial transaction fraud detection under class imbalance and asymmetric financial risk.Methods: In this study, a cost-sensitive RF framework with adaptive probability threshold optimization is proposed for binary fraud classification. Misclassification costs are incorporated during model training through class-weighted learning, allowing the classifier to assign greater importance to minority fraud cases. In addition, instead of using the default probability threshold of 0.5, the decision threshold is optimized using validation data to minimize the expected misclassification cost and improve minority-class detection. The proposed framework was evaluated on two publicly available financial fraud datasets with different transaction characteristics: the Credit Card Fraud Detection dataset and the Financial Fraud Prediction dataset. Its performance was compared with conventional ML baselines using fraud-oriented evaluation metrics, including precision, recall, F1-score, false-negative rate, false-positive rate, and expected misclassification cost.Results: The proposed framework achieved strong and consistent performance across both datasets. On the Credit Card Fraud Detection dataset, it achieved a precision of 94.4%, recall of 85.7%, and F1-score of 89.8%. On the Financial Fraud Prediction dataset, it achieved a precision of 93.1%, a recall of 83.8%, and an F1-score of 88.2%. Compared with conventional fixed-threshold ML models, the adaptive threshold strategy reduced false-negative predictions while maintaining high precision. These results indicate that the proposed cost-sensitive framework provides a more balanced decision strategy for fraud detection under highly imbalanced transaction distributions.Conclusion: The findings demonstrate that combining class-weighted learning with adaptive threshold optimization can improve fraud detection performance under asymmetric misclassification costs. The proposed framework offers a practical and interpretable decision-support approach for banking fraud detection systems, particularly in settings where reducing undetected fraudulent transactions is more important than maximizing overall accuracy alone.