Progress in Artificial Intelligence and Pattern Recognition (Lecture Notes in Computer Science)

Preface
Organization
Keynote Lecture
	The Role of Artificial Intelligence in Industry 4.0
Plenary Talks
	Computation at the Edge of Chaos: Natural Systems as Computation Paradigms
	Quantum Soft Computing: Applied SW & HW of Robust Intelligent Robotic Controllers
	Artificial Intelligence Applications Towards Digital Transformation
	Contents
Pattern Recognition and Applications
Fuzzy Employees Shuttle Bus Routing Problem. A Practical Application
	1 Introduction
	2 School Bus Routing Problem
		2.1 Employees Shuttle Bus Routing Problem
	3 Mathematical Model
		3.1 Parametric Approach
	4 Case of Study for the Transportation of Executives in the CUJAE
	5 Results and Discussion
	6 Conclusions
	References
Oversampling Method Based Covariance Matrix Estimation in High-Dimensional Imbalanced Classification
	1 Introduction
	2 Covariance Matrix Estimation in High-Dimensional Imbalanced Classification
	3 Numerical Data Simulation and Empirical Study
		3.1 Empirical Evaluation
	4 Concluding Remarks and Further Work
	References
Multivariate Cuban Consumer Price Index Database, Statistic Analysis and Forecast Baseline Based on Vector Autoregressive
	1 Introduction
	2 Multivariate Analysis
		2.1 Definitions and Notation
		2.2 Statistical Analysis in Multivariate CPI.
	3 Results and Discussion
		3.1 Exploratory Analysis and Dataset
		3.2 Statistics Analysis
	4 Concluding Remarks and Further Work
	References
Climatic Factors that Impact the Consumption Patterns Tame of Water. A Case of Study
	1 Introduction
	2 Background
	3 Materials and Methods
	4 Results and Discussion
	5 Conclusions
	References
An M-SALDAHP and GIS-Based Approach to Watershed Analysis During the Development of Small and Large-Scale Hydraulic Structure Construction Projects
	1 Introduction
	2 Methodology and Methods
	3 Results and Discussion
	4 Conclusions
	References
Fuzzy Granulation for Feature Extraction in EEG-Based Stress Pattern Recognition
	1 Introduction
	2 Fuzzy Granulation
		2.1 Fuzzy Information Granulation for Time Series
	3 Experiments and Results
		3.1 Fuzzy Clustering
		3.2 Triangular, Trapezoidal and Minmax Granulation
	4 Conclusions
	References
A Parallel Approach for RegularSearch Algorithm
	1 Introduction
	2 Concepts Definition
		2.1 Compatible and Regular Sets
	3 Parallel RegularSearch
		3.1 Execution Example
	4 Experimental Results
	5 Conclusions
	References
An Improved Fault Diagnosis Scheme Based on a Type-2 Fuzzy Classification Algorithms
	1 Introduction
	2 Materials and Methods
		2.1 Type-2 Fuzzy C-Means Algorithm (T2FCM) and Kernelized T2FCM (KT2FCM)
		2.2 Interval Type-2 Fuzzy C-Means Algorithm (IT2FCM) and Kernelized IT2FCM (KIT2FCM)
		2.3 Proposed Methodology
		2.4 Case Study: DAMADICS
		2.5 Design of Experiments
	3 Discussion of Results
	4 Online Recognition Stage
		4.1 Statistical Tests
	5 Conclusions
	References
A New Proposal for Detection and Location of Cyberattacks in Industrial Processes
	1 Introduction
	2 Materials and Methods
		2.1 Main Characteristics of Pythagorean Fuzzy Sets
		2.2 Kernel Pythagorean Fuzzy C-Means Algorithm
		2.3 Proposal of Scheme for Detection and Localization of Cyberattacks
		2.4 Case Study: Two-Tanks
		2.5 Design of Experiments
	3 Analysis and Discussion of Results
		3.1 Comparison with Other Algorithms
	4 Conclusions
	References
Computational Capabilities of Adler Oscillators Under Weak Local Kuramoto-Like Coupling
	1 Introduction
	2 The LAK Model
	3 Comparison to Cellular Automata
	4 Computational Complexity
	5 Conclusions
	References
Echo State Networks for the Prediction of Chaotic Systems
	1 Introduction
	2 Echo State Networks
		2.1 Training an ESN
	3 Results
		3.1 Lorenz System
		3.2 Mackey-Glass System
		3.3 Kuramoto Sivashinsky System
	4 Discussion and Conclusions
	References
Artificial Intelligence, Data Mining and Applications
Weighted t-Distributed Stochastic Neighbor Embedding for Projection-Based Clustering
	1 Introduction
	2 Related Works
	3 The Proposed Method
		3.1 The Traditional t-SNE Algorithm
		3.2 The Weighted t-SNE Algorithm
	4 Experimental Methodology
		4.1 Datasets
		4.2 Plotting the Embeddings
		4.3 Quantitative Comparison
	5 Concluding Remarks
	References
Evaluation of XAI Methods in a FinTech Context
	1 Introduction and Motivation
	2 Problem Description and Motivation
	3 Background and Related Work
		3.1 LIME by Ribeiro et al. 2016 ch13lime2016
		3.2 SHAP by Lundberg and Lee 2017 ch13shap2017
		3.3 CIU by Främling Initially 1996 ch13fraemling2020ciu
		3.4 IG by Sundararajan et al. 2017 ch13axiom2017sund
	4 Methodology
	5 Results
	6 Conclusion
	References
Towards Automatic Principles of Persuasion Detection Using Machine Learning Approach
	1 Introduction
	2 Related Work
	3 Data and Methodology
		3.1 Principles of Persuasion Dataset
		3.2 Learning Phase
	4 Experimental Work
		4.1 Experiments Results
	5 Conclusions
	References
Indirect Condition Monitoring of the Transmission Belts in a Desalination Plant by Using Deep Learning
	1 Introduction
	2 Materials and Methods
		2.1 Pilot RO Desalination Plant Description
		2.2 Description of the Mechanical Vibration Signal Capture System
		2.3 Signal Preprocessing
		2.4 Convolutional Neural Network Modeling
	3 Results and Discussion
	4 Conclusions
	References
A Novel Method for Filtering a Useful Subset of Composite Linguistic Summaries
	1 Introduction
	2 A Short Overview of CLS-QD Model
	3 A Method for Filtering and Assembling the Most Useful Relations
		3.1 Establishing an Attribute Preference Ranking
		3.2 Selecting the Most Useful Evidence Relation
		3.3 Selecting the Most Useful Contrast Relation
		3.4 Selecting the Most Useful Emphasis Relation
		3.5 Assembling the Pe*, Pc* and Ph* Relations
	4 Illustrative Example
	5 Concluding Remarks
	References
Harnessing Key Phrases in Constructing a Concept-Based Semantic Representation of Text Using Clustering Techniques
	1 Introduction
	2 Materials and Methods
		2.1 Key Phrases Extraction
		2.2 Concept Extraction
		2.3 Document Representation
		2.4 Experimental Research
	3 Results and Discussion
	4 Conclusions
	References
A Comparative Study of Deep Learning Methods for Brain Magnetic Resonance Image Reconstruction
	1 Introduction
	2 Methods
		2.1 Data Set
		2.2 Simulation of Aliasing and Blurring in MRI Images
		2.3 Image Reconstruction Schemes
		2.4 Experimented Deep Neural Networks Frameworks
	3 Experimental Results and Discussion
		3.1 Implementation and Training Details
		3.2 Evaluation of the Performance of the Obtained Models
	4 Conclusions
	References
Enhancing Spanish Aspect-Based Sentiment Analysis Through Deep Learning Approach
	1 Introduction
	2 Background
	3 Proposed Method for Spanish ABSA
		3.1 Preprocessing
		3.2 Aspect Extraction
		3.3 Aspect Sentiment Classification
	4 Experimental Results and Discussions
		4.1 Results of the Aspect Extraction Subtask
		4.2 Results of the Polarity Detection Subtask
	5 Conclusions and Future Works
	References
A Hybrid Approach for Spanish Emotion Recognition Applying Fuzzy Semantic Processing
	1 Introduction
	2 Related Works
	3 Proposed Solution
		3.1 Preprocessing
		3.2 Emotion Orientation Computing
		3.3 Semantic Feature Selection
		3.4 Emotion Prediction
	4 Experimental Results
	5 Conclusions and Future Works
	References
A Knowledge-Based User Feedback Classification Approach for Software Support
	1 Introduction
	2 Automatic Classification of Textual Content
		2.1 Machine Learning Algorithms
		2.2 Feature Extraction Techniques
		2.3 Analysis of Related Works
	3 Proposed Solution
		3.1 Pre-processing
		3.2 Extracting Relevance-Oriented Features
		3.3 Predicting Relevance
	4 Results and Discussion
		4.1 Deployment Technologies
		4.2 Experimental Results
	5 Conclusions
	References
A Metaphorical Text Classifier to Compare the Use of RoBERTa-Large, RoBERTa-Base and BERT-Base Uncased
	1 Introduction
		1.1 Metaphor Classification
	2 Materials and Methods
		2.1 LSTM and NLP
		2.2 Corpus of Metaphorical Language
		2.3 Pre-processing
		2.4 Description of the Proposed Solution to the Problem Posed
		2.5 Evaluation
	3 Results
		3.1 Comparison of Classifier Performance Using the Three Pre-trained Language Models
		3.2 Comparison with State-of-the-Art Models
	4 Conclusions and Future Work
	References
Improvements to the IntiGIS Model Related to the Clustering of Consumers for Rural Electrification
	1 Introduction
	2 IntiGIS Model
		2.1 The Proposed Geospatial Clustering Strategy
	3 Application and Results Analysis
	4 Conclusions
	References
Diagnosis of Alzheimer Disease Progression Stage from Cross Sectional Cognitive Data by Deep Neural Network
	1 Introduction
	2 Methods
		2.1 Study of the Degree of Correlation of Cognitive Variables
		2.2 Proposed Deep Neural Network Model
	3 Experimental Results and Discussion
		3.1 Implementation Details
		3.2 Model Training
		3.3 Validation of the Obtained Classifier
		3.4 Extension of the Validation Experiments to the ADNI2, ADNIGO and ADNI3 Cohorts
	4 Conclusions
	References
Robust MCU Oriented KWS Model for Children Robotic Prosthetic Hand Control
	1 Introduction
	2 Deep KWS Model for Evaluation
		2.1 Speech Feature Extraction
		2.2 Deep Learning Model Architectures
	3 Experimentation and Results Discussion
	4 MCU Selection for Model Deployment
	5 Conclusions and Future Work
	References
Detection of Malicious Bots Using a Proactive Supervised Classification Approach
	1 Introduction
	2 Related Works
	3 Malicious Bot Detection Module Based on Machine Learning
		3.1 Stage 2- Detection of Malicious Bot User Instances
	4 Experimental Study
		4.1 Database Description
		4.2 Preprocessing of Data
		4.3 Description of Experiments
		4.4 Discussion of the Results
	5 Conclusions
	6 Future Work
	References
Hybrid Selection of Breast Cancer Risk Factors in Cuban Patients
	1 Introduction
	2 Materials and Methods
		2.1 Feature Selection for Breast Cancer
		2.2 Dataset and Preprocessing Process
		2.3 Feature Selection Process
	3 Experimental Results
		3.1 First Experiment. Filter Method with Non-standardized Weights
		3.2 Second Experiment. Filter Method Using Normalized Weights
		3.3 Third Experiment. Wrapper Method with Backward Elimination
		3.4 Fourth Experiment. Wrapper Method with Forward Selection
		3.5 General Voting
	4 Discussion and Conclusions
	References
Good Negative Sampling for Triple Classification
	1 Introduction
	2 Related Work
		2.1 Traditional Negative Sampling
		2.2 Ontology-Based Negative Sampling
		2.3 Pykeen Negative Sampling
	3 Good Negative Sampling
		3.1 Formal Definitions
	4 Experiments
		4.1 Baselines
		4.2 KG-BERT
		4.3 Datasets
		4.4 Results and Discussion
	5 Conclusions and Future Work
	References
Polarity Prediction in Tourism Cuban Reviews Using Transformer with Estimation of Distribution Algorithms
	1 Introduction
	2 Computational Methodology
		2.1 Model Description
		2.2 Estimation of Distribution Algorithms and Hyper-parameters
		2.3 Dataset for Polarity Detection in Cuban Opinion Reviews
		2.4 Pre-processing and Data Augmentation
	3 Results and Discussion
	4 Conclusions and Further Work
	References
Biometrics, Image, Video and Signals Analysis and Processing
SqueezerFaceNet: Reducing a Small Face Recognition CNN Even More via Filter Pruning
	1 Introduction
	2 Network Pruning Method
	3 SqueezerFaceNet Architecture and Database
	4 Experimental Protocol
	5 Results
	6 Conclusions
	References
Drone Navigation and License Place Detection for Vehicle Location in Indoor Spaces
	1 Introduction
	2 Related Works
		2.1 Navigation
		2.2 Object Detection
	3 Methodology
		3.1 Hardware
		3.2 Software and Data
	4 Experiments and Results
		4.1 Test Cases
		4.2 Speed Factor
		4.3 Light Factor
	5 Conclusions
	References
Comparison of Root Mean Square Index and Hilbert-Huang Transform for Detection of Muscle Activation in a Person with Elbow Disarticulation
	1 Introduction
	2 Determination of the Onset/Offset Thresholds
		2.1 Root Mean Square (RMS)
		2.2 The Hilbert-Huang Transform
	3 Methodology
		3.1 Subjects
		3.2 Data Acquisition
		3.3 Signal Processing
		3.4 Data Analysis
	4 Results and Discussion
	5 Conclusions
	References
MinimalAI: Brain Hemorrhage Detection in Images Through Minimalist Machine Learning Approach
	1 Introduction
	2 Materials and Methods
		2.1 Image Processing
		2.2 Morphological Operations
		2.3 Permutation of Relevance: dMeans
		2.4 Minimalist Machine Learning Paradigm
	3 Proposed Methodology
		3.1 TC Scans Enhancement
		3.2 Minimalist Machine Learning
	4 Results and Discussion
		4.1 Dataset: Brain Hemorrhage CT Scans
		4.2 Competing State-of-the-Art Classifiers
		4.3 Comparative Analysis
		4.4 Conclusions
	References
Enhancing License Plate Recognition in Videos Through Character-Wise Temporal Combination
	1 Introduction
	2 Related Work
	3 Proposal
		3.1 Detection and OCR Methods for LPR
		3.2 Video Frame Combination Strategies
		3.3 Proposed Character-Based Combination Strategy
	4 Experiments
		4.1 Cuban License Plate Dataset
		4.2 Results
	5 Conclusion
	References
A Comparison of Some Gradient Threshold Estimators for the Nonlinear Diffusion of Perona-Malik. A Novel Proposal to Improve Edges Detection in Mammography Images
	1 Introduction
	2 Image Smoothing by Diffusion
		2.1 The Gradient Threshold in the Diffusion Coefficient
	3 Estimating the Gradient Threshold for Mammograms
		3.1 Problems with the Gradient Threshold Estimator KMLS for Mammogram Images
		3.2 KMLS2, a Modified Version of KMLS for AD Smoothing of Mammogram Images
	4 Experimental Environment and Results
		4.1 Results and Discussion
	5 Conclusions
	References
Offline Writer Identification and Verification Evaluation Protocols for Spanish Database
	1 Introduction
	2 Recurrent Neural Networks for Writer Identification
	3 Database and Evaluation Protocols
		3.1 Verification Evaluation Protocol
		3.2 Identification Evaluation Protocol
	4 Experimental Evaluation
		4.1 Implementation Details
		4.2 Model Training
		4.3 Verification Results
		4.4 Identification Results
	5 Conclusion
	References
Authenticity Assessment of Cuban Banknotes by Combining Deep Learning and Image Processing Techniques
	1 Related Work
	2 Security Measures and Features in Cuban Banknotes
	3 Proposed Pipeline
		3.1 Detection and Segmentation
		3.2 Selection of Interesting Regions
		3.3 Security Measures Processing
		3.4 Counterfait/Genuine Decision
	4 Experimental Evaluation
	5 Conclusions
	References
A Robust Preprocessing Method for Measuring Image Visual Quality Using Log-Polar FFT Features
	1 Introduction
	2 Proposed Method
	3 Experimental Results
	4 Conclusions
	References
Effectiveness of Blind Face Restoration to Boost Face Recognition Performance at Low-Resolution Images
	1 Introduction
	2 Related Work
		2.1 Deep Face Restoration
	3 Experimental Setup
		3.1 Blind Face Restoration
		3.2 Face Recognition
		3.3 Databases
		3.4 Implementation Details
	4 Experiments and Discussion
		4.1 Results on Synthesized LR Images
		4.2 Results on Native LR Images
	5 Conclusion
	References
Author Index