Intelligent Data Engineering and Automated Learning – IDEAL 2023: 24th International Conference, Évora, Portugal, November 22–24, 2023, Proceedings (Lecture Notes in Computer Science)

Preface
Organization
Contents
Main Track
Optimization of Image Acquisition for Earth Observation Satellites via Quantum Computing
	1 Introduction
	2 Background
	3 Mathematical Formulations of the Problem
		3.1 Classical Formulation of the SIASP
		3.2 Formulations of the SIASP for its Quantum Solving
	4 Experimentation
		4.1 Benchmark
		4.2 Setup and Results
	5 Conclusions and Further Work
	References
Complexity-Driven Sampling for Bagging
	1 Introduction
	2 Methodology
	3 Experiments
	4 Conclusions and Future Work
	References
A Pseudo-Label Guided Hybrid Approach for Unsupervised Domain Adaptation
	1 Introduction
	2 Methodology
		2.1 Model Architecture
		2.2 Training Process
	3 Experimental Results
	4 Conclusions
	References
Combining of Markov Random Field and Convolutional Neural Networks for Hyper/Multispectral Image Classification
	1 Introduction
	2 Background
		2.1 Markov Random Field and Gibbs Distribution
		2.2 Convolutional Neural Networks (CNNs)
	3 Materials and Methods
		3.1 MRF Filter Bank
		3.2 The CNN Structure
		3.3 Datasets
	4 Experiments and Results
	5 Conclusions
	References
Plant Disease Detection and Classification Using a Deep Learning-Based Framework
	1 Introduction
	2 Material and Method
		2.1 Dataset
		2.2 Data Pre-Processing
		2.3 Proposed Model
	3 Experiment and Results
		3.1 System Setup
		3.2 Training Regime
		3.3 Evaluation Protocols
		3.4 Result and Discussions
	4 Conclusion and Future Scope
	References
Evaluating Text Classification in the Legal Domain Using BERT Embeddings
	1 Introduction
	2 Related Works
		2.1 Natural Language Processing for Legal Document Analysis and Classification
		2.2 Embedding Models
		2.3 Classification of Legal Texts Using BERT Embeddings
	3 Materials and Methods
		3.1 Preprocessing
		3.2 Generation of Embeddings
		3.3 Classification Models
		3.4 Application of Classifiers
		3.5 Evaluation of Results
		3.6 Databases
	4 Results and Discussion
	5 Conclusions
	References
Rapid and Low-Cost Evaluation of Multi-fidelity Scheduling Algorithms for Hyperparameter Optimization
	1 Introduction
		1.1 Rising Prevalence of Hyperparameter Optimization
		1.2 Multi-fidelity Hyperparameter Optimization Schedulers
		1.3 Pre-tabulated Benchmarks are Not a Generic Solution for Evaluating HPO Schedulers
		1.4 Problem: Scheduler Evaluation Time is Slow
	2 A Platform for Hyperparameter Optimization Search Simulation (PHOSS)
		2.1 Surrogate Modelling Approach
		2.2 Implementation Details
	3 Related Work
	4 Results and Contributions
	References
The Applicability of Federated Learning to Official Statistics
	1 Introduction
	2 Background
		2.1 Federated Learning
		2.2 Privacy Challenges with Machine Learning
		2.3 Frameworks
	3 Simulations
		3.1 Medical Insurance Data
		3.2 Fine Dust Pollution
		3.3 Mobile Radio (LTE)
		3.4 Key Observations
	4 Implications for Official Statistics
	5 Conclusion
	References
Generating Wildfire Heat Maps with Twitter and BERT
	1 Introduction
	2 Related Work
		2.1 Fire Burst Detection
		2.2 Language Models
	3 Proposed Approach
		3.1 Data Fetching
		3.2 Classification
		3.3 Geoparsing
	4 Experimentation and Results
		4.1 Classifier
		4.2 Geoparser
	5 Conclusions and Future Work
	References
An Urban Simulator Integrated with a Genetic Algorithm for Efficient Traffic Light Coordination
	1 Introduction
	2 Urban Simulator
	3 Genetic Algorithm
		3.1 Representation
		3.2 Fitness
		3.3 Evolutionary Operations
		3.4 Genetic Algorithm Parameters
	4 Experiments
		4.1 Parameters Analysis
	5 Conclusions
	References
GPU-Based Acceleration of the Rao Optimization Algorithms: Application to the Solution of Large Systems of Nonlinear Equations
	1 Introduction
	2 Related Work
	3 The Sequential Rao Algorithms
	4 GPU Acceleration of the Rao Algorithms
	5 Numerical Experiments
	6 Results and Discussion
	7 Conclusion
	References
Direct Determination of Operational Value-at-Risk Using Descriptive Statistics
	1 Introduction and Motivation
	2 Literature Review
	3 Optimal Methodology Determination for VaR
		3.1 Data and Pre-processing
		3.2 Candidate VaR Assessment Models
		3.3 Success Metrics
		3.4 Optimisation Using a Proxy
	4 Results
		4.1 Model Prediction Results
		4.2 The Effect of the Proxy
	5 Discussion
	6 Conclusion
	References
Using Deep Learning Models to Predict the Electrical Conductivity of the Influent in a Wastewater Treatment Plant
	1 Introduction
	2 State of the Art
	3 Materials and Methods
		3.1 Data Collection
		3.2 Data Exploration
		3.3 Data Preparation
		3.4 Transformers
		3.5 LSTMs
		3.6 Evaluation Metrics
	4 Experiments
	5 Results and Discussion
	6 Conclusions
	References
Unsupervised Defect Detection for Infrastructure Inspection
	1 Introduction
	2 Related Work: Defect Detection in Civil Infrastructures
	3 Methodology
		3.1 Variational Autoencoder
		3.2 Anomaly Segmentation via Grad-CAMs
		3.3 Inference
	4 Experimental Setting
		4.1 Dataset and Evaluation Metrics
		4.2 Implementation Details
	5 Results
		5.1 Classification
		5.2 Unsupervised Defect Segmentation
	6 Conclusion
	References
Generating Adversarial Examples Using LAD
	1 Introduction
		1.1 Logical Analysis of Data
		1.2 Binarization
		1.3 Support Set Generation
		1.4 Pattern Generation
		1.5 LAD Classifier
	2 Related Work
	3 Performance Evaluation
		3.1 Adversarial Attack Generation
		3.2 Experimental Results on Adversarial Dataset
	4 Conclusion
	References
Emotion Extraction from Likert-Scale Questionnaires
	1 Introduction
	2 Related Work
	3 The Emotion Dimension of Likert-Scale Questionnaires
	4 Emotion Extraction on a Drop-Out Risk Assessement Instrument
	5 Discussion and Future Work
	6 Conclusion
	References
Recent Applications of Pre-aggregation Functions
	1 Introduction
	2 The Role Choquet Integral as (pre-) aggregation
	3 Methodology
		3.1 Studies Selection
	4 Analysing the Recent Application of Pre-aggregation Functions
	5 Conclusion
	References
A Probabilistic Approach: Querying Web Resources in the Presence of Uncertainty
	1 Introduction
	2 Uncertain Web Resources
		2.1 Definition
	3 Programmatic Representation of Uncertain Resources
	4 Composing Uncertain Web Resources
	5 The Operators
	6 Evaluation of a Request Within an Uncertain Resource
	7 Experimental Study and Analysis of the Results
	8 Conclusion
	References
Domain Adaptation in Transformer Models: Question Answering of Dutch Government Policies
	1 Introduction
	2 Related Work
	3 PolicyQA: A Dutch Government Policies Question and Answers Dataset
	4 Experimental Setup
		4.1 Datasets
		4.2 Extractive Question Answering Techniques
		4.3 Domain Adaptation
		4.4 Evaluation Metrics
	5 Results
	6 Discussion
	7 Conclusions and Future Work
	References
Sustainable On-Street Parking Mapping with Deep Learning and Airborne Imagery
	1 Introduction
		1.1 Our Contribution
	2 Related Work
	3 Dataset
	4 Data Processing and Detection Framework
		4.1 Data Pre-Processing Step
		4.2 Image Annotation Step
		4.3 Post-processing Steps
		4.4 Creating the Parking Space Map
	5 Experiments
		5.1 Experimental Setup
		5.2 Evaluation of the Experiments
	6 Conclusion and Future Work
	References
Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks
	1 Introduction
	2 Methods
		2.1 Adaptive Signed Random Walk (ASRW)
		2.2 Extending the ASRW With a Hebbian Learning-Inspired Strategy
		2.3 Community Detection Based on the Final Weight Matrix
	3 Results
		3.1 Synthetic Benchmarks
		3.2 Correlation-Based Functional Connectivity Networks Estimated from fMRI Data
	4 Conclusions and Future Work
	References
Extracting Automatically a Domain Ontology from the “Book of Properties” of the Archbishop’s Table of Braga
	1 Introduction
	2 Background
	3 The “Book of Properties”
	4 Learning the Ontology
	5 Conclusions and Future Work
	References
Language Models for Automatic Distribution of Review Notes in Movie Production
	1 Introduction
	2 Proposed Method
		2.1 Label Estimation
		2.2 Tokenization
		2.3 Classification
	3 Experiments and Results
	4 Conclusions and Future Work
	References
Extracting Knowledge from Incompletely Known Models
	1 Introduction
	2 Related Work
		2.1 Data Augmentation
		2.2 Model Extraction
		2.3 Explainability in Neural Networks
	3 Proposed Framework
		3.1 Data Augmentation Module
		3.2 Neural Network Generation Module
		3.3 Explainability Module
	4 Experiments
	5 Conclusions
	References
Threshold-Based Classification to Enhance Confidence in Open Set of Legal Texts
	1 Introduction
	2 Proposed Strategy
		2.1 Motivation for One-vs-Rest
		2.2 Threshold Scale
	3 Study Experimental
		3.1 Dataset
		3.2 Preprocessing
		3.3 Environment and Supporting Tools
		3.4 Training
		3.5 Threshold Scale Application
		3.6 Results
		3.7 Discussions
	4 Conclusion
	References
Comparing Ranking Learning Algorithms for Information Retrieval Systems
	1 Introduction
	2 Methodology
		2.1 Considered Tools
		2.2 Datasets
		2.3 Parameter Setup
	3 Obtained Results
		3.1 Time Consumption Analysis
		3.2 Training Behavior
		3.3 Evaluations
	4 Conclusions
	References
Analyzing the Influence of Market Event Correction for Forecasting Stock Prices Using Recurrent Neural Networks
	1 Introduction
	2 Preliminaries Concepts
		2.1 Market Event Correction Factor
		2.2 Recurrent Neural Network
		2.3 Data Adaptability in Windowed Dataset
	3 Related Works
	4 Methodology
	5 Obtained Results
		5.1 The Learning Rate
		5.2 Tunning Hyperparameters
		5.3 Proposed Model
	6 Conclusion
	References
Measuring the Relationship Between the Use of Typical Manosphere Discourse and the Engagement of a User with the Pick-Up Artist Community
	1 Introduction
	2 Data and Resources
	3 Methodology
	4 Results
	5 Conclusion
	References
Uniform Design of Experiments for Equality Constraints
	1 Introduction
	2 Latin Hypercubes
	3 Design of Experiments for Equality Constraints
	4 Case Studies
		4.1 Two-Dimensional Comparison
		4.2 Three-Dimensional Pareto Front
	5 Conclusion
	References
Globular Cluster Detection in M33 Using Multiple Views Representation Learning
	1 Introduction
	2 Dataset
		2.1 Data Collection
		2.2 Data Pre-processing
		2.3 Data Transformations
	3 The Proposed YOLO-GC Model
		3.1 Problem Setting
		3.2 YOLO-GC Architecture
	4 Experimental Results
	5 Conclusion
	References
Segmentation of Brachial Plexus Ultrasound Images Based on Modified SegNet Model
	1 Introduction
	2 The Modified SegNet Model
		2.1 The Structure of SegNet Model
		2.2 Our Improvements
	3 Training Details
		3.1 The Dataset of Brachial Plexus Ultrasound Images
		3.2 Data Preprocessing
	4 Experiments
		4.1 Parameter Settings
		4.2 Implementation Details
	5 Experimental Results and Discussion
	6 Conclusions
	References
Unsupervised Online Event Ranking for IT Operations
	1 Introduction
	2 Related Work
	3 Proposed Methods
		3.1 Atomic Events Detection
		3.2 Complex Events Detection
		3.3 Clustering for IT Ops
	4 Evaluation Framework
	5 Results
	6 Discussion
	7 Conclusion
	References
A Subgraph Embedded GIN with Attention for Graph Classification
	1 Introduction
	2 Related Works
	3 Proposed Method
		3.1 Node Selection and Subgraph Partitioning Algorithm
		3.2 Graph Embedding Network
		3.3 Attention-Based Readout for Classification
	4 Experimental Results
		4.1 Datasets
		4.2 Baseline Methods
		4.3 Result Analysis
	5 Concluding Remarks
	References
A Machine Learning Approach to Predict Cyclists’ Functional Threshold Power
	1 Introduction
	2 Research Background and Related Work
	3 Machine Learning Model to predict a cyclist’s FTP
		3.1 Data and Data Preprocessing
		3.2 Actual vs Estimated FTP
		3.3 Predicting FTP with Multiple Regression Models
	4 Evaluation of our Machine Learning Model
	5 Conclusion
	References
Combining Regular Expressions and Supervised Algorithms for Clinical Text Classification
	1 Introduction
	2 Materials and Methods
		2.1 Clinical Texts and Pre-processing
		2.2 Classification Algorithm Based on Regular Expressions
		2.3 Feature Space Based on Regular Expressions
		2.4 Decision Function
	3 Experimental Results
	4 Conclusions and Future Work
	References
Modeling the Ink Tuning Process Using Machine Learning
	1 Introduction
	2 Literature Review
		2.1 Related Work
	3 Material and Methods
		3.1 Business and Data Understanding
		3.2 Data Preparation
		3.3 Modeling
		3.4 Evaluation
	4 Results and Discussion
	5 Conclusions
	References
Depth and Width Adaption of DNN for Data Stream Classification with Concept Drifts
	1 Introduction
	2 Related Work
	3 Network Structure and Learning Strategy of ANSN
		3.1 Network Structure
		3.2 Classification Mechanism
		3.3 Learning Strategy for Network Depth Adaption
		3.4 Learning Strategy for Network Width Adaption
		3.5 ANSN pseudocode
	4 Experiment
		4.1 Dataset
		4.2 Baseline
		4.3 Experimental Setup
		4.4 Results
	5 Conclusion
	References
FETCH: A Memory-Efficient Replay Approach for Continual Learning in Image Classification
	1 Introduction
	2 Problem Formulation
	3 Related Work
		3.1 Literature Review
		3.2 Greedy Sampler and Dumb Learner
	4 Approach
		4.1 Fixed Encoder and Trainable Classification Head
		4.2 Compressor and Decompressor
		4.3 Calculation of the Storage Consumption
	5 Implementation Details
	6 Experiments
		6.1 Tradeoff Between Storage and Performance
		6.2 Ablation: Effect of the Fixed Encoder
		6.3 Ablation: Effect of Compression
		6.4 Performance on a Fixed Memory Budget
		6.5 Comparison with Other Approaches
	7 Conclusion
	References
Enhanced SVM-SMOTE with Cluster Consistency for Imbalanced Data Classification
	1 Introduction
	2 Border-Based SMOTE
	3 Materials and Methods
		3.1 Datasets
		3.2 Proposed Method
	4 Experimental Results and Discussion
		4.1 Artificial Datasets
		4.2 UCI Datasets
		4.3 Hyperspectral Dataset
	5 Conclusions and Future Work
	References
Preliminary Study on Unexploded Ordnance Classification in Underwater Environment Based on the Raw Magnetometry Data
	1 Introduction
	2 The Virtual Environment and the Dataset
	3 Experiments
		3.1 Dataset Processing and Evaluation Scheme
		3.2 Results
	4 Conclusions
	References
Efficient Model for Probabilistic Web Resources Under Uncertainty
	1 Introduction
	2 Uncertain Web Resources
		2.1 Definition
		2.2 Model of an Uncertain Web Resource
	3 Programmatic Representation of Uncertain Resources
		3.1 Parser JSON Probabilistic Model
	4 HTTP Request on Uncertain Resources
		4.1 Query Evaluation: Algorithm 1
	5 Experimental Study and Analysis of the Results
	6 Conclusion
	References
Unlocking the Black Box: Towards Interactive Explainable Automated Machine Learning
	1 Introduction
	2 The Need for Transparency to Trust in AI and in AutoML
	3 Explainable Artificial Intelligence
	4 Visual Analytics for AutoML
		4.1 Towards Interactive Explainable AutoML
		4.2 Key Components of Visual Analytics for AutoML
	5 Conclusion
	References
Machine Learning for Time Series Forecasting Using State Space Models
	1 Introduction
	2 State Space Models
	3 Time Series Forecasting Methodology
	4 Regularization Framework for State Space Models
	5 Model Selection with K-Fold Cross-Validation Based on Imputed Data
	6 State Space Alternating Least Squares (SSALS) Algorithm
	7 Validation with Simulated and Real Data
		7.1 Small Scale Simulation with Noisy Data
		7.2 Application to Prey-Predator Real Data
	8 Conclusion
	References
Causal Graph Discovery for Explainable Insights on Marine Biotoxin Shellfish Contamination
	1 Introduction
	2 Background
	3 Data Preprocessing
	4 Results and Discussion
	5 Conclusion and Future Work
	References
Special Session on Federated Learning and (Pre) Aggregation in Machine Learning
Adaptative Fuzzy Measure for Edge Detection
	1 Introduction
	2 Preliminaries
	3 Experimental Setup
		3.1 Dataset and Evaluation of the Proposal
		3.2 Framework and Results
	4 Conclusions and Future Work
	References
Special Session on Intelligent Techniques for Real-World Applications of Renewable Energy and Green Transport
Prediction and Uncertainty Estimation in Power Curves of Wind Turbines Using ε-SVR
	1 Introduction
	2 State-of-Art
	3 Methodology
		3.1 Epsilon Support Vector Regression (ε-SVR)
		3.2 Uncertainty Estimation in ε-SVR
		3.3 Gaussian Process Regression (GPR)
	4 Dataset and Preprocessing
	5 Model and Results
	6 Conclusions and Future Works
	References
Glide Ratio Optimization for Wind Turbine Airfoils Based on Genetic Algorithms
	1 Introduction
	2 Numerical Optimization Procedure
	3 Discussion of Results
	4 Conclusions and Future Works
	References
Special Session on Data Selection in Machine Learning
Detecting Image Forgery Using Support Vector Machine and Texture Features
	1 Introduction
	2 The Proposed Method for Digital Image Forgery Detection
		2.1 Preprocessing
		2.2 Feature Extraction
		2.3 Classification
	3 Simulation Results
		3.1 Data
		3.2 Size Dependent Model
		3.3 Size Independent Model
	4 Conclusion
	References
Instance Selection Techniques for Large Volumes of Data
	1 Introduction
	2 Data Preprocessing
	3 Instance Selection
	4 The Proposed Methodology
	5 Experimental Results
	6 Conclusions
	References
Author Index