Entradas recentes
A predictive analytics framework for early detection of production halts and quality issues
( 2025 ) Matta, Arthur; Matos, Luís Miguel; Silva, Jorge Miguel; Bastos Gomes, Miguel; Pilastri, André; Cortez, Paulo
This study presents a Machine Learning (ML) framework for an Ahead-of-Time (AoT) prediction of production halts and defects in particleboard manufacturing that uses only pre-production input variables. The proposed approach incorporates both Single-Task Learning (STL) and Multi-Task Learning (MTL) paradigms, which are evaluated across three production lines under two modeling strategies: Line-Specific Modeling (LSM) and Line-Agnostic Modeling (LAM). The experimental evaluation benchmarks a lightweight Logistic Regression (LogR) model against three Automated Machine Learning (AutoML) techniques: H2O AutoML, Ludwig, and a Bayesian-optimized Deep Feedforward Network (DFFN). Results show that the STL-LSM combination using LogR achieves the highest overall predictive performance. To enhance model interpretability, we apply two model-agnostic eXplainable Artificial Intelligence (XAI) techniques: SHapley Additive exPlanations (SHAP) and One-Dimensional Sensitivity Analysis (1DSA). These methods generate feature importance rankings across targets and production lines, which are evaluated using quantitative (normalized distance metrics) and qualitative measures (alignment with domain expert insights). The XAI findings reveal a strong consistency between SHAP and 1DSA, with 1DSA requiring a substantially lower computational cost. Moreover, the convergence between model-derived interpretations and expert feedback highlights the practical relevance of the proposed ML framework in supporting data-driven decision-making for particleboard production planning.
Screw Process Anomaly Visualization (SPAV): A Python module for local and global machine learning visualizations for screw tightening anomaly detection
( 2025 ) Moreno, Marta; Rocha, Hugo; Pilastri, André; Moreira, Guilherme; Matos, Luís Miguel; Cortez, Paulo
Modern screwdriver systems generate real-time angle-torque data that form tightening curves that are valuable for quality inspection issues (e.g., detect faulty processes). This work describes the Screw Process Anomaly Visualization (SPAV) Python module, which provides several eXplainable AI (XAI) graphs for Machine Learning (ML) screw tightening results, namely global and local errors, with identification of most probable anomaly angle-torque locations. SPAV integrates seamlessly with the scientific Python ecosystem and is compatible with several ML implementations, including H2O and Keras deep AutoEncoders (AE).
Pre-school education in Portugal: effects of external evaluation on public and private institutions
( 2013 ) Rodrigues, Eduarda
The Portuguese law nº31/2002, of 20th December started a new way of seeing education through external evaluation (Stufflebeam, 2003). With an outsider look, an external team started to help public schools to improve their services by pointing their fragilities but also its stronger points leading each school to find a way to self-improve (Sobrinho, 2003). By the same time, private schools started their own process of evaluation. Different schools are being evaluated by different organizations with similar objectives (but not the same). In the 2000 decade, external evaluation has become a common procedure in every public school and in consequence preschool education is now far more recognised than before (Pacheco, Seabra, Morgado & van Hattum, 2014). The empirical studies we are undertaking attempt to find out what was the impact of external evaluation in both institutions – public and private. To accomplish this objective we decided for a qualitative study based upon content analysis of documents and interviews (Bogdan & Biklen, 1999). The interviews were made to directors and preschool teachers of public and private schools from S. João da Madeira. Despite the preliminary results, we can say that external evaluation gave a great contribution to preschool education prominence either in public as in private institution. If public preschools are pointed as a reference in articulation with other school levels and in curriculum sequence, private schools evaluation leans towards efficiency and accountability criteria.