Intelligent Decision Technologies : Proceedings of the 15th KES-IDT 2023 Conference

This book gathers selected papers from the KES-IDT 2023 Conference, held in Rome, Italy, on June 14-16, 2023. The book presents and discusses the latest research results and generates new ideas in the field of intelligent decision-making. The range of topics discussed is classification, prediction, data analysis, big data, data science, decision support, knowledge engineering and modeling in diverse areas such as finance, cybersecurity, economics, health, management and transportation. The problems in industry 4.0 and IoT are also addressed. The book contains several sections devoted to specific topics, such as intelligent data processing and its applications, high-dimensional data analysis and its applications, multi-criteria decision analysis -- theory and applications, large-scale systems for intelligent decision-making and knowledge engineering, decision technologies and related topics in big data analysis of social and financial issues and decision-making theory for economics Preface KES-IDT 2023 Conference Organization Contents About the Editors Keynote Speeches Data, Information, Knowledge, Wisdom Pyramid Concept Revisited in the Context of Deep Learning 1 Introduction 2 Intelligent Music and Speech Processing 3 Conclusions References Multimedia Industrial and Medical Applications Supported by Machine Learning 1 Introduction 2 Intelligent Autonomous Road Signs for Adaptive Traffic Control 3 Intelligent Monitoring of Wind Turbine Health 4 Multi-biometric Authentication of Banking Clients 5 HCIBRAIN - Assessment of Impaired Cognitive Functions 6 Methodology of Polymodal Allophonic Speech Transcription 7 CyberRadar 8 Conclusions References Main Track Accessibility Measures and Indicators: A Basis for Dynamic Simulations to Improve Regional Planning 1 Introduction 2 A System Theory Perspective 3 Method 4 Results 4.1 Requirements for Simulation Modeling 4.2 Representations of Accessibility 4.3 Accessibility Measures and Indicators 5 Concluding Remarks References Image-Multimodal Data Analysis for Defect Classification: Case Study of Industrial Printing 1 Introduction 2 Objectives and Previous Studies 2.1 Data Set 2.2 Previous Studies 3 Proposal Method #1: Multimodal Analyses 3.1 Pre-process 3.2 Proposal and Evaluations: #1-Multimodal Analyses 4 Proposal and Evaluations: #2: Multimodal Ensemble Model 5 Conclusion References Image-Multimodal Data Analysis for Defect Classification: Case Study of Semiconductor Defect Patterns 1 Introduction 2 Objectives and Previous Studies 2.1 Dataset 2.2 Previous Studies 3 Proposed Methods and Evaluations 3.1 Pre-process: Resize and Data Augmentation 3.2 Proposal Method #1: 2-step Classification with Radon Features 4 Proposal Method #2: Advanced Model of Multimodal Analysis 5 Conclusion References Decision-Making Model for Updating Geographical Information Systems for Polish Municipalities Using the Fuzzy TOPSIS Method 1 Introduction 2 Decision-Making Models 3 Problem Formulation 3.1 Objective and Decision-Making Conditions 3.2 Decision-Making Criteria 3.3 Decision Alternatives 4 Proposed TOPSIS Decision-Making Model 4.1 Fuzzy TOPSIS Method 4.2 Fuzzy TOPSIS Calculations 5 Conclusions References An Ontology-Based Collaborative Assessment Analytics Framework to Predict Groups’ Disengagement 1 Introduction 2 The Collaborative Assessment Analytical Framework (CAAF) Description 3 Assessment Data Ontology Modeling 4 Methodology 5 Results 5.1 Knowing About Groups Disengaged 5.2 Predicting Groups’ Disengagement Depending on Formative Assessment 6 Discussion and Conclusion References Artificial Intelligence Innovation in Daily Life On Sensing Non-visual Symptoms of Northern Leaf Blight Inoculated Maize for Early Disease Detection Using IoT/AI 1 Introduction 2 Materials and Methods 2.1 Experimentation Approach for Data Collection 2.2 IoT Based Data Collection Approach 3 Results and Discussion 3.1 Dataset and Data Preprocessing 3.2 Checking to Stationarity (ADF) 3.3 Identification of Volatiles Patterns 3.4 Ultrasound Emission Patterns 3.5 NPK Consumption Pattern 4 Conclusion References Inference Analysis of Lightweight CNNs for Monocular Depth Estimation 1 Introduction 2 Related Works 3 Proposed Methodology 3.1 Loss Function 3.2 Mask Network 4 Depth Network 5 Analyses of the Predicted Mask 6 Other Network Architecture 7 Summary and Conclusion References Arabic Text-to-Speech Service with Syrian Dialect 1 Introduction 2 Problem Statement 2.1 Phonemes and Mel-spectrogram 3 Background 4 Comparative Analysis 4.1 Tacotron2 4.2 Transformers 5 FastSpeech 5.1 Feed-Forward Transformer 5.2 Length Regulator 5.3 Duration Predictor 6 Service Development and Implementation 6.1 Architecture 6.2 Criteria 6.3 Limitations 6.4 Technologies and Tools 6.5 Model Training 6.6 Service API 7 Evaluation 8 Conclusion and Future Prospects References A Systematic Review of Sentiment Analysis in Arabizi 1 Introduction 2 The Informal Arabic Text and Sentiment Analysis 3 Methodology Used 4 Results 5 Conclusion References Reasoning-Based Intelligent Applied Systems Optimize a Contingency Testing Using Paraconsistent Logic 1 Introduction 2 Backgrounds 2.1 Compliance and Risk Considerations 2.2 ISO 22.301/2013 - Business Continuity Management System 2.3 Logic Paraconsistent Annotated Evidential Eτ 2.4 Para-Analyser Algorithm 3 Methodology 3.1 The Proposal of Data Analysis for Decision-Making 4 Discussions and Results 5 Conclusion References Age-Group Estimation of Facial Images Using Multi-task Ranking CNN 1 Introduction 2 Related Work 3 The Proposed Method 4 Experimental Results 5 Conclusions References Evaluation Instrument for Pre-implementation of Lean Manufacturing in SMEs Using the Paraconsistent Annotated Evidential Logic Eτ Evaluation Method 1 Introduction 2 Theoretical Reference 2.1 Lean Manufacturing 2.2 Paraconsistent Annotated Evidential Logic Eτ 2.3 Degrees of Contradiction and Certainty 2.4 ProKnow-C Method 3 Methodology 3.1 Raw Article Bank Selection 3.2 Criteria for Composing the Indicators 4 Model Building 4.1 Definition of Lean Maturity Categories and Indicators 4.2 Definition of the Questionnaires Made up of the Indicators 4.3 Definition of Experts and Interviewees 4.4 Definition of Measurement Scales for Assessing Lean Maturity 4.5 Expert Evaluation Stage 5 Results and Discussion 6 Conclusion References Recent Development of Multivariate Analysis and Model Selection Modified Cp Criterion in Widely Applicable Models 1 Introduction 2 Main Result 3 Examples of MCp A.1 The Proof of Lemma 1 A.2 The Proof of Lemma 2 A.3 The Proof of Lemma 3 A.4 The Proof of Lemma 4 A.5 The Proof of Theorem 1 A.6 The Proof of Equation in (15) References Geographically Weighted Sparse Group Lasso: Local and Global Variable Selections for GWR 1 Introduction 2 Optimization Algorithm for GWSGL 2.1 Preliminaries 2.2 Solution Update and Main Algorithm 3 Real Data Example 3.1 Data and Method 3.2 Results References Kick-One-Out-Based Variable Selection Method Using Ridge-Type Cp Criterion in High-Dimensional Multi-response Linear Regression Models 1 Introduction 2 Sufficient Conditions for Consistency 3 Numerical Studies References Estimation Algorithms for MLE of Three-Mode GMANOVA Model with Kronecker Product Covariance Matrix 1 Introduction 2 Model and Log-Likelihood Function 3 Proposed Algorithms 3.1 Block-Wise Coordinate Descent Algorithm 3.2 Dual-Block-Wise Coordinate Descent Algorithm 4 Numerical Experiments References Implications of the Usage of Three-Mode Principal Component Analysis with a Fixed Polynomial Basis 1 Three-Mode Principal Component Analysis: 3MPCA 2 Proposed Methods: Three-Mode Principal Component Analysis with a Fixed Orthonormal Polynomial Basis 3 Applications of the Proposed Method References Spatio-Temporal Analysis of Rates Derived from Count Data Using Generalized Fused Lasso Poisson Model 1 Introduction 2 Model 3 Optimization Procedure 3.1 Methods for Minimizing 4 Application and Conclusion References Large-Scale Systems for Intelligent Decision Making and Knowledge Engineering An Ontology Model to Facilitate Sharing Risk Information and Analysis in Construction Projects 1 Introduction 2 Ontology and Risk Management in AEC Industry 3 Research Methodology and Stages 3.1 Steps of Ontology Development 4 Conclusions and Recommendations References Designing Sustainable Digitalization: Crisisology-Based Tradeoff Optimization in Sociotechnical Systems 1 Introduction 2 Crisisology as a Conceptual Framework for Multi-criteria Decision Support in Information Systems Design 3 A Proposed Method for Hierarchical Multi-criteria Choice 4 A Proposed Hierarchical Structure of the Criteria for Selection of IT System Architecture 5 Description of the Decision Support Service 6 Case Study of the Proposed Prototype 7 Conclusion References Decision Making Theory for Economics Calculations by Several Methods for MDAHP Including Hierarchical Criteria 1 Introduction 2 SPCM 3 Numerical Example of Using SPCM for Calculation of MDAHP 4 Results of Calculation by CCM 4.1 Results of Calculation by CCM with the Eigenvalue Method 4.2 Results of Calculation by CCM with the Geometric Mean Method 5 Results of Calculation by GMMDAHP 5.1 Results of Calculation by GMMDAHP with the Eigenvalue Method 5.2 Results of Calculation by GMMDAHP with the Geometric Mean Method 6 Results of Calculation from SPCM by LLSM 7 Results of Calculation from SPCM by the Harker Method 8 Results of Calculation from SPCM by ITSM 9 Conclusion References Utilization of Big Data in the Financial Sector, Construction of Data Governance and Data Management 1 Introduction 1.1 History and Current Status of Data Management in the Financial Sector 1.2 The Era of Ultra-large Data and Corporate Expectations 2 Previous Research on Big Data and Frameworks for Data Management 2.1 Utilization of Big Data in an Organization 2.2 Data Management and Data Governance 3 Utilization of Big Data in Japanese Financial Institutions 4 Trend Survey of Global Big Data Utilization 5 Conclusion References A Block Chart Visualizing Positive Pairwise Comparison Matrices 1 Introduction 2 A Block Chart 3 Adjusting Weights Manually with Using the Block Chart 4 Discussions 5 Conclusion References Applied Methods of Machine and Deep Learning Neural Networks Combinations for Detecting and Highlighting Defects in Steel and Reinforced Concrete Products 1 Introduction 2 Detection and Segmentation Artificial Neural Networks 3 Experiments and Results 4 Conclusions References Modern Methods of Traffic Flow Modeling: A Graph Load Calculation Model Based on Real-Time Data 1 Modern Methods of Traffic Flow Modeling 1.1 Introduction 1.2 A New Traffic Flow Model Prediction 2 Complex Traffic Jams Predicting Method 2.1 Initial Data Processing Algorithm 2.2 Point Scale 2.3 Choosing a Model Training Method 2.4 Model Calculation 2.5 Quality Control 3 Conclusion References Application of Machine Learning Methods for the Analysis of X-ray Images of Luggage and Hand Luggage 1 Introduction 2 Ensemble Approach for Luggage Image Recognition 3 Detection of Prohibited Items of Luggage 4 Conclusions References Author Index