Database and Expert Systems Applications: 34th International Conference, DEXA 2023, Penang, Malaysia, August 28–30, 2023, Proceedings, Part II (Lecture Notes in Computer Science)

The two-volume set, LNCS 14146 and 14147 constitutes the thoroughly refereed proceedings of the 34th International Conference on Database and Expert Systems Applications, DEXA 2023, held in Penang, Malaysia, in August 2023. The 49 full papers presented together with 35 short papers were carefully reviewed and selected from a total of 155 submissions. The papers are organized in topical sections as follows: Part I: Data modeling; database design; query optimization; knowledge representation; Part II: Rule-based systems; natural language processing; deep learning; neural networks. Preface Organization Abstracts of Keynote Talks Physics-Informed Machine Learning Data Integration Revitalized: from Data Warehouse through Data Lake to Data Mesh Contents – Part II Contents – Part I Rule-Based Systems RD-Classifier: Reduced Dimensionality Classifier for Alzheimer’s Diagnosis Support System 1 Introduction 2 Related Work 3 Proposed Model 3.1 Data Preparation 3.2 RD-Classifier 3.3 Evaluation 4 Experimentation 4.1 Experimental Environment Setup 4.2 Dataset 4.3 Results and Discussion 5 Conclusion References User Interaction-Aware Knowledge Graphs for Recommender Systems 1 Introduction 2 Related Work 2.1 KG-Based Recommendation 2.2 Graph Neural Network Techniques 3 Problem Formulation 4 Methodology 4.1 Learn User Representation 4.2 Learn Item Representation 5 Experiments 5.1 Evaluation Datasets and Metrics 5.2 Baselines 5.3 Experimental Setup 5.4 Results 5.5 Ablation Study of UIKR 6 Conclusion and Future Work References How Does the System Perceive Me? — A Transparent and Tunable Recommender System 1 Introduction 2 Related Work 2.1 Explainable Recommendation 2.2 Interactive Recommendation 2.3 Tag-Based Recommendation 3 Proposed System 3.1 Item Modeling 3.2 User Modeling 3.3 Rating Prediction 3.4 User Feedback 4 Experimental Evaluation 4.1 Datasets 4.2 Evaluation Metrics 4.3 Design Choices 4.4 Case Studies 5 Conclusion References MERIHARI-Area Tour Planning by Considering Regional Characteristics 1 Introduction 2 Related Work 3 MERIHARI Area Tour Planning 3.1 Area Clustering 3.2 Extraction of Representative Spots 3.3 Courses of Area Visiting 3.4 MERIHARI Area Plan 4 Experiments 4.1 Dataset 4.2 Result of Area Clustering 4.3 Result of Extraction of Representative Spots and Generate Courses 4.4 Result of Generate MERIHARI Area Plan 4.5 Evaluation Methods 4.6 Experimental Results 5 Conclusion References Financial Argument Quality Assessment in Earnings Conference Calls 1 Motivation 2 Related Work 2.1 Computational Argument Quality Assessment 2.2 Text Quality in Finance and Business Communication 3 Argument Quality Dimensions 3.1 At the Level of Argument 3.2 At the Level of Argument Unit 4 Data Creation 4.1 Annotation Study 4.2 Inter Annotator Agreement 4.3 FinArgQuality Data Statistics 5 Discussion and Conclusions References Explaining Decisions of Black-Box Models Using BARBE 1 Introduction 2 Related Work 3 Associative Classification 4 BARBE: Black-Box Association Rule-Based Explanation 4.1 Shortcoming of Other Methods 4.2 Explanations by BARBE 4.3 How Does BARBE Work? 5 Experiments 5.1 Experiments Setup 5.2 Experiments' Metrics 5.3 Comparison with Other Explainers 5.4 Faithfulness to the Black-Box 6 Experiments on BARBE for Text 6.1 Results 6.2 Comparison with Other Explainers 7 Conclusion and Perspectives References Efficient Video Captioning with Frame Similarity-Based Filtering 1 Introduction 2 Related Work 3 Preliminaries 3.1 Tensor Dot (TD) 3.2 Chamfer Similarity (CS) 3.3 LSTM for the Sequence-to-Sequence Models 4 Proposed Method 4.1 Frame-to-Frame Similarity 4.2 Frame Selection 4.3 Feature Extraction 4.4 LSTM for Video Caption Generation 5 Results and Discussion 5.1 Data Sets 5.2 Evaluation Metric 5.3 Results 5.4 Training Time and Number of Parameters 6 Conclusion References Trace-Based Anomaly Detection with Contextual Sequential Invocations 1 Introduction 2 Related Works 3 TICAD Design 3.1 Trace Pre-processing 3.2 Anomaly Detection 4 Evaluation 4.1 Datasets and Criteria 4.2 Preparation Experiments 4.3 Experiments on Trace-Level Anomaly Detection 5 Conclusion References Fusing Fine-Grained Information of Sequential News for Personalized News Recommendation 1 Introduction 2 Related Work 3 Personalized News Recommendation 3.1 News Encoder 3.2 Clicked News Optimizer and User Encoder 3.3 Click Predictor and Model Training 4 Experiments 5 Conclusion and Future Work References A Finite-Domain Constraint-Based Approach on the Stockyard Planning Problem 1 Introduction 2 The Stockyard Planning Problem 3 Basics of Constraint Programming 4 Abstraction of the Problem 5 A Constraint Optimization Problem for the SPP 6 Conclusion and Future Work References Data Analytics Framework for Smart Waste Management Optimisation: A Key to Sustainable Future for Councils and Communities 1 Introduction 2 Proposed Data Analytics Framework 3 Implementation 3.1 Dataset 3.2 Waste Generation Forecasting 3.3 Results and Discussion 4 Conclusion References Natural Language Processing Hierarchy-Aware Bilateral-Branch Network for Imbalanced Hierarchical Text Classification 1 Introduction 2 Related Work 2.1 Hierarchical Text Classification 2.2 Data Imbalance in Classification 3 Problem Definition 4 Methodology 4.1 Data Sampler 4.2 Text Encoder 4.3 Cumulative Learning 4.4 Feature Propagation 4.5 Training and Inference 5 Experiments 5.1 Datasets 5.2 Implementation Details 5.3 Comparison Models 5.4 Results and Discussion 5.5 Label Granularity Performance Study 5.6 Ablation Study 6 Conclusion References Multi-Feature and Multi-Channel GCNs for Aspect Based Sentiment Analysis 1 Introduction 2 Related Work 3 Proposed Approach 3.1 Embedding and Encoding Layer 3.2 Aspect GCN Module 3.3 Type RGAT Module 3.4 Sem-Syn GCN Module 3.5 Feature Fusion 3.6 Classifier 3.7 Training 4 Experiments 4.1 Dataset 4.2 Experimental Setting 4.3 Baseline Methods 4.4 Main Results 4.5 Ablation Study 4.6 Case Study 5 Conclusion References Knowledge Injection for Aspect-Based Sentiment Classification 1 Introduction 2 Related Work 2.1 Hybrid Models for ABSC 2.2 Knowledge Injection in Neural Networks 3 Data 4 Methodology 4.1 LCR-Rot-hop++ 4.2 Knowledge Injection 5 Results 6 Conclusion References Towards Ensemble-Based Imbalanced Text Classification Using Metric Learning 1 Introduction 2 Related Work 2.1 Text Representations 2.2 Imbalanced Classification 3 MLEnsemble: Metric Learning-Based Ensemble for Imbalanced Text Classification 3.1 MLBagging 3.2 MLBoosting 3.3 MLStacking 3.4 MLBoostacking 4 Experimental Evaluation 4.1 Settings 4.2 Results 4.3 Lessons Learned 5 Conclusion References Target and Precursor Named Entities Recognition from Scientific Texts of High-Temperature Steel Using Deep Neural Network 1 Introduction 2 Related Works 3 Methodology 3.1 Data Collection and Processing 3.2 Dataset Preparation for TP-NER Model Training 3.3 TP-NER: Deep Neural Network Model Development 3.4 Evaluation Metrics 4 Result and Discussion 5 Conclusion and Future Work References Enabling PII Discovery in Textual Data via Outlier Detection 1 Introduction 2 Related Work 3 PII Discovery and Masking 4 Experimental Evaluation and Results 5 Conclusion References Deep Learning An Efficient Embedding Framework for Uncertain Attribute Graph 1 Introduction 2 Related Works 3 The Proposed Methods 3.1 Problem Definition 3.2 Embedding Framework for Uncertain Attribute Graph 4 Experiments 4.1 Datasets and Experiment Setup 4.2 Performance Comparisons 5 Conclusion References Double-Layer Attention for Long Sequence Time-Series Forecasting 1 Introduction 2 Related Work 2.1 Traditional Time Series Forecasting Models 2.2 Transformer-Based Time Series Models 3 The Proposed Model: DEPformer 3.1 Problem Definition 3.2 DEPformer Model 4 Experiments 4.1 Datasets 4.2 Baselines 4.3 Hyper-parameter Tuning 4.4 Results and Analysis 4.5 Model Analysis and Discussion 5 Conclusion References Multi-core Adaptive Merging of the Secondary Index for LSM-Based Stores 1 Introduction and Motivation 2 Related Work 3 Adaptive Indexing 3.1 Adaptive Merging Without Database Modification 3.2 Database Modification 4 Experiments 5 Conclusion References CAGAIN: Column Attention Generative Adversarial Imputation Networks 1 Introduction 2 Related Work 3 Problem Statement 4 Method 4.1 The CAGAIN Model 4.2 Loss Functions 5 Experiments 5.1 Experimental Setup 5.2 Experimental Results 6 Conclusion and Future Work References CF-SAFF: Collaborative Filtering Based on Self-attention Mechanism and Feature Fusion 1 Introduction 2 Related Work 2.1 CF Model 2.2 Self-attention Mechanism 2.3 Graph Neural Networks 3 The Proposed Model: CF-SAFF 3.1 Symbolic Description 3.2 Attention Mechanism Attempts 3.3 CF-SAFF Model 4 Experiments 4.1 Datasets and Data Preparation 4.2 Evaluation Indicators 4.3 Baselines 4.4 Experimental Setup 4.5 Performance Comparison 4.6 Model Analysis and Discussion 5 Conclusion References Except-Condition Generative Adversarial Network for Generating Trajectory Data 1 Introduction 2 Background 3 Except-Condition Generative Adversarial Network (ExGAN) 3.1 Condition and Exception 3.2 Trajectory Data Model 4 Experimental Results 5 Conclusion References Next POIs Prediction for Group Recommendations: Influence-Based Deep Learning Model 1 Introduction 2 Problem Definition 3 Methodology 3.1 User Representation 3.2 Group Representation 3.3 The Interaction Learning Module 4 Experimental Evaluation 5 Conclusion References Interpreting Deep Text Quantification Models 1 Introduction 2 Background and Related Work 2.1 Performance Metrics for Quantification Models 2.2 LRP 3 Proposed Work: Interpreting QuaNet 3.1 LRP-Based Algorithm to Calculate Relevance Scores 3.2 Method to Calculate the Amount of the Contribution of Each Input to QuaNet 4 Experiment Design and Results 4.1 Datasets 4.2 Experimental Settings 4.3 RQ1: What Input Features are Important to the Final Class Distribution Prediction of QuaNet? 4.4 RQ2: Does Sorting L by Pr(c|x) Increase the Performance of QuaNet? 5 Conclusion and Future Work References NExtGCN: Modeling Node Importance of Graph Convolution Network by Neighbor Excitation for Recommendation 1 Introduction 2 Method 2.1 Input Layer 2.2 NExtGCN Layer 2.3 Prediction Layer 3 Experiments 3.1 Experimental Settings 3.2 Performance Comparison 3.3 Ablation of NExtGCN 4 Conclusions References Dual Congestion-Aware Route Planning for Tourists by Multi-agent Reinforcement Learning 1 Introduction 2 Related Work 3 Multi-agent Reinforcement Learning Environment 4 Dual-Congestion Aware Routes Planning Model 4.1 Multi-agent Reinforcement Learning Implementation 4.2 Dual-Congestion Mechanism 5 Experimental Evaluation 5.1 Environment Setting 5.2 Baseline Setting 5.3 Evaluation Metrics 5.4 Experimental Results 6 Conclusion References Subspace Clustering Technique Using Multi-objective Functions for Multi-class Categorical Data 1 Introduction 2 The Proposed Multi-objective Technique 2.1 Cao Clustering Algorithm 2.2 Step One: Finding and Analyzing the Clusters 2.3 Step Two: Reassigning Non-clustered Objects 3 Experiments and Results 4 Conclusion and Future Work References Neural Networks Multi-task Graph Neural Network for Optimizing the Structure Fairness 1 Introduction 2 Related Work 3 Methodology 3.1 Overview of Architecture 3.2 Position Partition 3.3 Link Matrix Construction 3.4 Multi-task Graph Neural Network 4 Experimental Analysis and Discussion 4.1 Data Description 4.2 Performance of GNN-OSF 4.3 Ablation Experiments 4.4 Effect of Different Similarity Distances Threshold t 4.5 Effect of Different Link Weight 4.6 Impact of Multi-task Learning on Node Representation 4.7 Performance of Fraud Detection 5 Conclusion References Few-Shot Multi-label Aspect Category Detection Utilizing Prototypical Network with Sentence-Level Weighting and Label Augmentation 1 Introduction 2 Related Work 3 Methodology 3.1 Overview 3.2 Word-Level Attention with Label Augmentation 3.3 Sentence-Level Attention 3.4 Query Attention 3.5 Training Objective 4 Experiments 4.1 Dataset 4.2 Baseline Models 4.3 Evaluation Metrics 4.4 Experimental Settings 4.5 Experimental Results and Discussions 5 Conclusion References Toward Healthy Aging: Temporal Regression for Disability Prediction and Warning Decision-Making 1 Introduction 2 Related Work 3 Problem Statement 4 Our Approach 4.1 Encoding 4.2 Training 4.3 Prediction 4.4 Decision-Making 5 Experiments 5.1 Data 5.2 Baselines 5.3 Variants 5.4 Evaluation 5.5 Results 5.6 Factors' Impact 6 Conclusions References A Label Embedding Method via Conditional Covariance Maximization for Multi-label Classification 1 Introduction 2 Preliminaries 3 A Novel Label Embedding Approach via Conditional Covariance Maximization 3.1 Conditional Covariance Operator and Its Simplified Form 3.2 A Novel Label Embedding Method for Multi-label Classification 4 Experiments 4.1 Four Data Sets and Two Evaluation Metrics 4.2 Three Compared Methods 4.3 Experimental Settings 4.4 Performance Comparison and Analysis 5 Conclusions References Integrally Private Model Selection for Deep Neural Networks 1 Introduction 2 Model Comparison Attack for DNNs 2.1 Framework 2.2 Intruders Approach 2.3 Integral Privacy 3 -Integrally Private Model Selection for DNNs 4 Experimental Results 4.1 Discussion 4.2 Limitations 5 Conclusion and Future Work References Gaussian Process Component Mining with the Apriori Algorithm 1 Introduction 2 Gaussian Processes 3 Method 4 Preliminary Experiments 4.1 Synthetic Data 4.2 Real Data 5 Conclusion References Learnable Filter Components for Social Recommendation 1 Introduction 2 Framework Overview 2.1 Problem Definition 2.2 Model Architecture 3 Experiment 3.1 Dataset 3.2 Evaluation Index 3.3 Baseline 3.4 Parameter Setting 3.5 Experimental Results and Analysis 4 Conclusion and Future Work References Efficient Machine Learning-Based Prediction of CYP450 Inhibition 1 Introduction 2 Proposed Model for the CYP450 Isoforms Inhibition Prediction 2.1 CYP450 Bioassay Data Collection 2.2 Data Preprocessing 2.3 Molecular Descriptors Calculation 2.4 Experimentation 3 Results and Discussion 4 Conclusion References A Machine-Learning Framework for Supporting Content Recommendation via User Feedback Data and Content Profiles in Content Managements Systems 1 Introduction 2 Related Work 3 Methodology 3.1 System Architecture 3.2 User Preference Matrix Generation 3.3 Matrix Factorization and Post Similarity Calculation 4 Experiments 4.1 Result Analysis 5 Conclusions and Future Work References Fine-Tuning Pre-Trained Model for Consumer Fraud Detection from Consumer Reviews 1 Introduction 2 Methodology 2.1 Dataset Construction 2.2 Model Training 3 Experiment 3.1 Data Preparation 3.2 Experimental Setup 3.3 Experimental Evaluation 3.4 Empirical Study 4 Conclusion References Deep Multi-interaction Hidden Interest Evolution Network for Click-Through Rate Prediction 1 Introduction 2 Related Work 3 The Proposed Method 3.1 Embedding Layer 3.2 Hidden Interest Extraction Layer 3.3 Item-to-Item Sub-network and User-to-Item Sub-network 3.4 MLP and Loss Function 4 Experiments 4.1 Datasets and Experimental Setup 4.2 Baseline and Result 5 Conclusion References Temporal Semantic Attention Network for Aspect-Based Sentiment Analysis 1 Introduction 2 Methodology 2.1 Embedding 2.2 Global Semantic Feature Network 2.3 Interact Dual Attention 2.4 Output 3 Experiment 3.1 Experiment Settings 3.2 Performance Comparison 3.3 Case Study 4 Conclusion References Celestial Machine Learning 1 Introduction 2 Background and Related Work 3 Methodology 4 Performance Evaluation 5 Conclusion References Author Index