Database and Expert Systems Applications: 32nd International Conference, DEXA 2021, Virtual Event, September 27–30, 2021, Proceedings, Part I ... Applications, incl. Internet/Web, and HCI)

This two-volume set, LNCS 12923 and 12924, constitutes the thoroughly refereed proceedings of the 5 th International Conference on Database and Expert Systems Applications, DEXA 2021. Due to COVID-19 pandemic, the conference was held virtually. The 37 full papers presented together with 31 short papers in these volumes were carefully reviewed and selected from a total of 149 submissions. The papers are organized around the following topics: big data; data analysis and data modeling; data mining; databases and data management; information retrieval; prediction and decision support. Preface Organization Abstracts of Keynote Talks Privacy in the Era of Big Data, Machine Learning, IoT, and 5G Don’t Handicap AI without Explicit Knowledge Extreme-Scale Model-Based Time Series Management with ModelarDB Big Minds Sharing their Vision on the Future of AI (Panel) Contents – Part I Contents – Part II Big Data Reference Architecture for Running Large Scale Data Integration Experiments 1 Introduction 2 Reference Architecture: Logical Level 3 Reference Architecture: Implementation Level 4 Summary References Subgroup Discovery with Consecutive Erosion on Discontinuous Intervals 1 Introduction 2 Related Works 3 SD-CEDI Algorithm 3.1 Preliminaries 3.2 SD-CEDI 4 Experimental Results 4.1 Best Quality on Raw Dataset 4.2 Best Quality on Resized Datasets 4.3 Time Comparison 5 Conclusion References Fast SQL/Row Pattern Recognition Query Processing Using Parallel Primitives on GPUs 1 Introduction 2 Related Work 2.1 Implementation of SQL/RPR Based on Spark SQL 2.2 SQL Query Processing on GPUs 3 SQL/Row Pattern Recognition 4 Implementation of SQL/RPR on GPUs 4.1 Partition 4.2 Sorting 4.3 Pattern Definition Using Regular Expression 4.4 Pattern Matching 5 Evaluation 5.1 Method Used for Comparison 5.2 Experiment Setup 5.3 Experimental Results 6 Conclusion References Scalable Tabular Metadata Location and Classification in Large-Scale Structured Datasets 1 Introduction 2 Definitions 2.1 Non-relational Table Representation 3 Methodology 3.1 Ensemble Architecture 3.2 Large-Scale Evaluation Architecture 4 Experimental Study 4.1 Training the Models 4.2 Training Data 4.3 Test Data 5 Evaluation 6 Related Work 7 Conclusion References Unified and View-Specific Multiple Kernel K-Means Clustering 1 Introduction 2 Background 2.1 Kernel k-means (KKM) 2.2 Multiple Kernel k-means (MKKM) 3 The Proposed Method 3.1 A Variant of Kernel k-means 3.2 Unified and View-Specific Multiple Kernel Clustering 3.3 Optimization 4 Experiments and Analysis 4.1 DataSets 4.2 Compared Algorithms 4.3 Experimental Setup 4.4 Clustering Performance 4.5 Kernel Structure and Parameter Sensitivity Study 5 Conclusion References Data Analysis and Data Modeling Augmented Lineage: Traceability of Data Analysis Including Complex UDFs 1 Introduction 2 Related Work 3 Data Model 4 Augmented Lineage 5 Augmented Lineage Derivation 5.1 Segment 5.2 Tracing Query 5.3 Augmented Lineage Derivation Procedure 6 Implementation of Augmented Lineage Derivation 7 Experiment 8 Conclusions and Future Work References Neural Ordinary Differential Equations for the Regression of Macroeconomics Data Under the Green Solow Model 1 Introduction 2 The Green Solow Model 3 Related Work 4 Methodology 4.1 Baseline Method 4.2 Proposed Models 5 Performance Evaluation 5.1 Experimental Setup 5.2 Results and Discussion 6 Conclusion References A Quantum-Inspired Neural Network Model for Predictive BPaaS Management 1 Introduction 2 Related Work 3 Resource Usage Prediction Model 3.1 Multi-layer Perceptron for Resource Usage Prediction 3.2 Quantum Genetic Algorithm for the Neural Network's Training 4 Experiments 4.1 Evaluation of the Prediction Model 4.2 Evaluation of the Placement Quality 5 Conclusion and Future Work References Predicting Psychiatric Diseases Using AutoAI: A Performance Analysis Based on Health Insurance Billing Data 1 Introduction 2 Method 2.1 Data 2.2 AutoAI Frameworks, Development Environment, and Configuration 3 Results 3.1 Watson AutoAI 3.2 Auto-Sklearn 4 Discussion 5 Conclusion References Improving Billboard Advertising Revenue Using Transactional Modeling and Pattern Mining 1 Introduction 2 Proposed Framework of the Problem 3 Proposed Billboard Allocation Framework 4 Performance Evaluation 5 Conclusion References Sarcasm Detection for Japanese Text Using BERT and Emoji 1 Introduction 2 Related Works 3 Proposed Method 3.1 Definition of Sarcasm 3.2 Overview of Proposed Method 3.3 BERT: Bidirectional Encoder Representations from Transformers 3.4 Word2Vec 4 Experiment 4.1 Dataset and Preprocessing 4.2 Experimental Setup 4.3 Experimental Results 5 Conclusion References Sigmalaw PBSA - A Deep Learning Model for Aspect-Based Sentiment Analysis for the Legal Domain 1 Introduction 2 Related Work 2.1 Legal Information Extraction 2.2 Existing Aspect-Based Sentiment Analysis Models 3 Methodology 3.1 Word Embedding Layer 3.2 RNN Layer 3.3 Position Aware Attention Mechanism 3.4 Graph Convolution Network 3.5 Sentiment Classification 3.6 Model Training 4 Experiments 4.1 Dataset 4.2 Parameter Setting 4.3 Word Embedding Models Comparison 4.4 RNN Models Comparison 4.5 Overall Performance 4.6 Ablation Study 5 Conclusion References BERT-Based Sentiment Analysis: A Software Engineering Perspective 1 Introduction 1.1 BERT 2 Related Work 3 Proposed Work 3.1 Data Augmentation 3.2 Fine-Tuning BERT Model 3.3 Ensemble BERT Models 3.4 Compressed Model 4 Experiments 4.1 Training and Testing 5 Results and Discussion 6 Conclusion References A Stochastic Block Model Based Approach to Detect Outliers in Networks 1 Introduction 2 Model Definition 3 Experiments References Medical-Based Text Classification Using FastText Features and CNN-LSTM Model 1 Introduction 2 Related Works 3 Materials and Methods 3.1 Dataset 3.2 Feature Extraction 3.3 Machine Learning Algorithms 3.4 CNN-LSTM 4 Experimentations and Results 4.1 Experimental Flow 4.2 Analysis and Discussion of the Results 5 Conclusion References Data Mining Diversified Pattern Mining on Large Graphs 1 Introduction 2 Related Work 3 Graphs, Patterns and Pattern Mining 3.1 Graph Pattern Matching 3.2 Diversified Top-k Pattern Mining 4 Diversified Top-k Pattern Mining 5 Experimental Study 6 Conclusion References EHUCM: An Efficient Algorithm for Mining High Utility Co-location Patterns from Spatial Datasets with Feature-specific Utilities 1 Introduction 2 Problem Definition 3 The EHUCM Algorithm 3.1 The Search Space 3.2 Feature-Object Neighbor Tree (FONT) 3.3 Two Pruning Strategies 3.4 The EHUCM Algorithm 4 Experimental Evaluation 5 Conclusions References BERT-Based Multi-Task Learning for Aspect-Based Opinion Mining 1 Introduction 1.1 Problem Definition 1.2 Contributions 2 Literature Review 3 Proposed Approach 3.1 Multi-Task Learning Model 3.2 Aspect Term and Category Related Opinion Polarity 4 Steps of BERT-MTL 4.1 Steps of BERT-MTL for ATE and ACD Algorithm 4.2 Steps of BERT-MTL for Fine and Coarse-Grained Algorithm 5 Experimental Evaluation 5.1 Dataset 5.2 Hyper-Parameters 5.3 Result Analysis 6 Conclusions and Future Work References GPU-Accelerated Vertex Orbit Counting for 5-Vertex Subgraphs 1 Introduction 2 Preliminaries 2.1 Orbit Counting 2.2 EVOKE 3 Proposed Method 3.1 {3, 4}-Vertex Orbit Counting 3.2 5-Vertex Orbit Counting 4 Experiments 4.1 Experimental Setup 4.2 Execution Time Comparison 5 Related Work 6 Conclusion References Databases and Data Management Efficient Discovery of Partial Periodic-Frequent Patterns in Temporal Databases 1 Introduction 1.1 Background and Related Work 1.2 Motivation 1.3 The Contributions of This Paper 2 Proposed Algorithm 3 Experimental Results 3.1 Experimental Setup 4 Conclusions and Future Work References Database Framework for Supporting Retention Policies 1 Introduction 2 Retention Definitions and Requirements 3 Related Work 4 Policy Setting 5 Policy Execution 6 Experiments 7 Conclusion and Future Work References Internal Data Imputation in Data Warehouse Dimensions 1 Introduction 2 Preliminaries 3 Internal Data Imputation for Dimensions 4 Experimental Assessments 4.1 Datasets and Experimental Method 4.2 Intra-dimensional Imputation Experiments 4.3 Inter-dimensional Imputation Experiments 5 Conclusion and Future Work References Purging Data from Backups by Encryption 1 Introduction 1.1 Motivation 2 Background 2.1 Compliance Terminology 2.2 Database Backups and Types 2.3 Related Work 3 Our Process 3.1 Defining Policies 3.2 Encryption Process 3.3 Encryption on Update 3.4 Purging Process 3.5 Restore Process 4 Experiments 5 Discussion 5.1 Implementation 5.2 ACID Guarantees 5.3 Future Work 6 Conclusion References Information Retrieval Improving Quality of Ensemble Technique for Categorical Data Clustering Using Granule Computing 1 Introduction 2 Background and Notations 2.1 Information Granulation 2.2 Hybridization of Fuzzy Sets and Rough Sets 3 The Proposed Techniques 3.1 Information Granulation Selection Criteria 3.2 Fuzzy Rough Refinement Matrix 4 Experiments and Results 4.1 Evaluation of the Proposed Selection Criteria IGQS 4.2 Evaluation of the Proposed Refinement Techniques 4.3 Statistical Significance Tests 5 Conclusions and Future Work References Online Optimized Product Quantization for Dynamic Database Using SVD-Updating 1 Introduction 2 Preliminaries 2.1 Approximate k-Nearest Neighbor Search 2.2 Vector Quantization, Product Quantization, and Its Variants 2.3 Online Product Quantization for Dynamic Data 2.4 Other Methods Based on PQ 3 Proposed Method 3.1 Updating Codebook 3.2 Update Rotation Matrix Using SVD-Updating 3.3 Orthogonality 3.4 Computational Complexity 4 Experiments 4.1 Experimental Setting 4.2 Effects of Low-Rrank Approximation 4.3 Comparison with Non-online Methods 4.4 Comparison with Online Methods 5 Conclusion References Querying Collections of Tree-Structured Records in the Presence of Within-Record Referential Constraints 1 Introduction 2 Defining the Data Model 3 Within-Record Constraints 4 Querying Tree-Structured Data 4.1 Flattening Nested Data 4.2 Navigating Through References References Dealing with Plethoric Answers of SPARQL Queries 1 Introduction 2 Motivating Example 3 Related Work 4 Preliminaries and Problem Statement 4.1 Basic Notions 4.2 Notions of MFIS and XSS 5 Computing MFIS and XSS 5.1 Baseline 5.2 General Properties 5.3 Cardinality-Based Property 6 Experimental Evaluation 6.1 Results 7 Conclusion References Prediction and Decision Support Feature Selection and Software Defect Prediction by Different Ensemble Classifiers 1 Introduction 2 Related Works 3 Methods and Results 3.1 The Dataset 3.2 Research Methods: Feature Selection 3.3 Research Methods: Classification 3.4 Research Methods: The Ensemble of Classifiers 4 Conclusions and Future Work References Traffic Flow Prediction Through the Fusion of Spatial-Temporal Data and Points of Interest 1 Introduction 2 Related Work 2.1 Convolution Neural Network-Based Traffic Prediction Method 2.2 Graph Neural Network-Based Traffic Prediction Method 3 Preliminaries 4 Method 4.1 Gated Fusion Network 4.2 Hierarchical Adaptive Graph Convolution 4.3 Gated Convolution Module 4.4 Multi-task Learning Loss 5 Experiments 5.1 Experimental Settings 5.2 Performance Results (RQ1) 5.3 Ablation Experiments (RQ2) 5.4 Evaluation on Framework Settings (RQ3) 5.5 Evaluation on Spatial-Temporal Fusion Methods (RQ4) 6 Conclusion and Future Work References Predicting Student Performance in Experiential Education 1 Introduction 2 Preliminaries, Key Concepts and Theories 3 Model and Methods 4 Experiments, Results and Discussion References Log-Based Anomaly Detection with Multi-Head Scaled Dot-Product Attention Mechanism 1 Introduction 2 Related Works 2.1 Log Parsing 2.2 Log-Based Anomaly Detection 3 LogAttention Design 3.1 Overview 3.2 Log Parsing 3.3 Log Pattern Vectorization 3.4 Log Anomaly Detection 4 Evaluation 4.1 Datasets and Criteria 4.2 Overall Results and Analysis 4.3 Experiments on Log Sequences with Different Window Lengths 4.4 Experiments on Log Pattern Vectorization Method 5 Conclusion References Enhancing Scan Matching Algorithms via Genetic Programming for Supporting Big Moving Objects Tracking and Analysis in Emerging Environments 1 Introduction 2 Background Concepts 3 Genetic Pre-alignment Phase 4 A Novel Family of Enhanced Hybrid Scan Matching Algorithms 5 Experimental Study 5.1 Genetic Pre-alignment: Performance Analysis 5.2 Proposed Hybrid Algorithms: Performance Analysis 6 Conclusions and Future Work References A Stacking Approach for Cross-Domain Argument Identification 1 Introduction 2 Related Work 2.1 Argument Mining 2.2 Ensemble Learning 3 Contribution 3.1 Problem Statement 3.2 Corpora Description 3.3 Classical Machine Learning Model - SVM 3.4 Transfer Learning Model (DistilBERT-Based) 3.5 Overall Model (SVM + DistilBERT) 4 Evaluation 5 Conclusion References Property Analysis of Stay Points for POI Recommendation 1 Introduction 2 Related Work 3 Method 3.1 Preliminaries 3.2 Weighted HITS-Based Algorithm 4 Experiments and Results 5 Conclusion References Beacon Technology for Retailers - Tracking Consumer Behavior Inside Brick-and-Mortar-Stores 1 Introduction 2 Theoretical Background 2.1 Challenges of Brick-and-Mortar Stores 2.2 Beacon Technology 2.3 Applications of Beacon Technology in Brick-and-Mortar Stores 2.4 Challenges of Beacon Technology in Brick-and-Mortar Stores 3 Recommendations for Implementation 3.1 Prerequisites for Adapting Beacon Technology 3.2 Recommendations for Establishing Beacon Technology 4 Conclusion References Author Index