Data Mining and Big Data: 7th International Conference, DMBD 2022, Beijing, China, November 21–24, 2022, Proceedings, Part II

Preface
Organization
Contents – Part II
Contents – Part I
Identification and Recognition Methods
Complementary Convolutional Restricted Boltzmann Machine and Its Applications in Image Recognition
	1 Introduction
	2 Theory of Foundations
		2.1 Restricted Boltzmann Machine
		2.2 Deep Belief Nets
		2.3 Convolutional Neural Nets
	3 Complementary Convolutional Deep Belief Nets
		3.1 Complementary Restricted Boltzmann Machine
		3.2 Analysis for Complementary Restricted Boltzmann Machine
		3.3 Complementary Convolutional Restricted Boltzmann Machine
		3.4 Complementary Convolutional Deep Belief Nets
	4 Experiments
	5 Conclusion
	References
Text-Independent Speaker Identification Using a Single-Scale SincNet-DCGAN Model
	1 Introduction
	2 Related Work
		2.1 Single-Scale SincNet Structure
		2.2 Deep Convolutional Generative Adversarial Networks
	3 The Proposed Speaker Identification System
	4 Experiment
		4.1 Datasets
		4.2 Experimental Setup
		4.3 Results Analysis
	5 Conclusion
	References
Genome-Wide Feature Selection of Robust mRNA Biomarkers for Body Fluid Identification
	1 Introduction
	2 Methods
		2.1 Data Availability and Pre-processing
		2.2 Tissue Specificity Index
		2.3 Identification of Candidate Biomarkers
		2.4 Identification of Reference Genes
		2.5 Testing on Public Datasets
		2.6 Deconvolution Analysis
	3 Experimental Results
		3.1 Principal Component Analysis of Six Body Fluid-related Tissues
		3.2 Feature Selection for Tissue-specific Biomarkers
		3.3 Identification of Reference Genes for Normalization
		3.4 Comparison with Previously Reported mRNA Biomarkers
		3.5 Generalization on External Datasets
		3.6 Mixture Component Analysis with Identified Biomarkers
	4 Discussion
	5 Conclusion
	References
HOS-YOLOv5: An Improved High-Precision Remote Sensing Image Target Detection Algorithm Based on YOLOv5
	1 Introduction
	2 Related Work
		2.1 Object Detection
		2.2 Tiny Object Detection
		2.3 Evaluation Metric in Object Detection
	3 Method
		3.1 Review of YOLOv5
		3.2 HOS Backbone Network
		3.3 Introduce DotD Algorithm
	4 Experiments
		4.1 Dataset
		4.2 Implementation Details
		4.3 Experimental Results
		4.4 Ablation Experiment
	5 Conclusion
	References
A Multi-module 3D U-Net Learning Architecture for Brain Tumor Segmentation
	1 Introduction
	2 Proposed Methods
		2.1 Residual Inception Blocks
		2.2 Loss Function
	3 Evaluations and Analyses
		3.1 Method of Training
		3.2 Comparative Analysis of the Proposed Model and Standard U-Net and Without Inception Module
		3.3 Ablation Study
		3.4 Comparative Analysis with Other Methods
		3.5 Discussions
	4 Conclusion
	References
A Better Linear Model Than Regression-Line for Data-Mining Applications
	1 Introduction
	2 Translation-Invariance Property of an OPDL
	3 Effect of Translation on RL
	4 Effect of Rotation on RL
	5 Effect of Rotation on an OPDL
		5.1 Minimizing aspd When m =
	6 Conclusion
	References
Research on Hot Spot Mining Technology for Network Public Opinion
	1 Introduction
	2 Related Work
	3 Methodology
		3.1 Word2Vec
		3.2 Single-Pass Algorithm
		3.3 WPK-Means++ Algorithm Implementation
	4 Experimental Setup
		4.1 Dataset
		4.2 Data Pre-processing
		4.3 Text Representation
	5 Experimental Result
		5.1 Check Accuracy Rate
		5.2 Check the Full Rate
	6 Conclusion
	References
Research Hotspots, Emerging Trend and Front of Fraud Detection Research: A Scientometric Analysis (1984–2021)
	1 Introduction
	2 Data and Method
		2.1 Data
		2.2 Method
	3 Result and Discussion
		3.1 Scientific Outputs of Fraud Detection Research
		3.2 Characteristics of International Collaboration
		3.3 Characteristics of Institutional Cooperation
		3.4 Publication Distribution and Co-citation Network Analysis
		3.5 Character of Subject Categories
		3.6 Research Hotspots and Emerging Trends of Fraud Detection
	4 Conclusion
	References
Optimization Methods
An Algorithm of Set-Based Differential Evolution for Discrete Optimization Problem
	1 Introduction
	2 Preliminary
		2.1 Particle Swarm Optimization
		2.2 Set-Based Particle Swarm Optimization
		2.3 Differential Evolution
	3 Set-Based Differential Evolution
	4 Numerical Experiments
	5 Conclusions
	References
Multi-objective Optimization Technique for RSU Deployment
	1 Introduction
	2 Problem Description and System Model
		2.1 Multi-RSU Cooperative Data Transmission Network Framework
		2.2 Transmission Delay Model
		2.3 RSU Number Model
		2.4 Problem Formulation
	3 Algorithm Background
	4 Experiment and Simulation
		4.1 Simulation Set up
		4.2 Experiment Data
	5 Conclusion and Future Work
	References
Knowledge Learning-Based Brain Storm Optimization Algorithm for Multimodal Optimization
	1 Introduction
	2 Background Knowledge
		2.1 Brain Storm Optimization Algorithm
		2.2 Multimodal Optimization Problem
	3 Knowledge Learning for Brain Storm Optimization
		3.1 Knowledge-Driven Strategies for Optimization Algorithms
		3.2 Knowledge Learning Strategy of Brain Storm Optimization
	4 Experimental Results and Analysis
		4.1 Algorithm Parameter Settings
		4.2 Results
	5 Conclusion and Future Work
	References
Market Investment Methods
Non-local Graph Aggregation for Diversified Stock Recommendation
	1 Introduction
	2 Related Work
		2.1 Stock Prediction
		2.2 Graph Neural Networks
	3 Method
		3.1 Problem Statement
		3.2 Architecture
		3.3 Spatial-Temporal Embedding
		3.4 Non-local Graph Aggregation
		3.5 Diversified Stock Recommendation
	4 Experiments
		4.1 Settings
		4.2 Performance Comparison
		4.3 Ablation Study
		4.4 Signal Analysis
	5 Conclusion
	References
Novel Sentiment Analysis from Twitter for Stock Change Prediction
	1 Introduction
	2 Related Work
		2.1 System Design
		2.2 Data Collection
		2.3 Pre-processing
		2.4 Tokenizing Text Mood by XLNet
		2.5 Sentiment Analysis by FinBert
		2.6 Comparing Sentiment Analysis Results of XLNet and FinBert
		2.7 Cross-Validation of XLNet and FinBert Time Series for High-Impact Sociocultural Events
		2.8 The Lag of Public Sentiment on Events
		2.9 Model Training and Prediction
	3 Conclusions and Future Work
	References
A Novel Investment Strategy for Mixed Asset Allocation Based on Entropy-Based Time Series Prediction
	1 Introduction
		1.1 Background
		1.2 Problem Analysis
	2 Objective Empowerment Multi-objective Programming Investment Strategy Based on ARIMA
		2.1 Symbol, Definitions and Assumptions
		2.2 Data Preprocessing
		2.3 ARIMA Model
		2.4 Multi-objective Programming Model
		2.5 Entropy Weight Method
	3 Results and Analysis
		3.1 Case Study
		3.2 The Discussion of Optimal Strategy
		3.3 Sensitivity Analysis
	4 Conclusion
	References
The Framework of Hammer Credit Rating System for Enterprises in Capital Markets of China with International Standards
	1 The Background and Related Issues
	2 The Framework of CAFÉ Risk Assessment System by Using Bigdata Method
		2.1 The Framework for the Construction of CAFÉ Risk Assessment System
		2.2 The Extraction of Risk Features Based on AI Algorithm Under the Bigdata Framework
		2.3 The Construction of Credit Transition Matrix by Hammer (CAFÉ) System for Financial Markets in China
		2.4 The Applications of Hammer (CAFÉ) Assessment System on Capital Markets in China
	3 The Conclusion with Brief Discussion and Remark
	Appendix: The Framework for the Extraction of Features Based on Gibbs Algorithms
	References
Community Detection and Diagnosis Systems
A Self-adaptive Two-Stage Local Expansion Algorithm for Community Detection on Complex Networks
	1 Introduction
	2 The Proposed Algorithm SALEA
		2.1 The General Framework of SALEA
		2.2 Self-adaptive Strategy
		2.3 Self-adaptive Local Expansion Stage
		2.4 Self-adaptive Global Merger Stage
		2.5 Complexity Analysis
	3 Experimental Results
		3.1 Experimental Settings
		3.2 Experiments on Real-World Networks
		3.3 Experiments on LFR Networks
	4 Conclusion and Future Work
	References
Supervised Prototypical Variational Autoencoder for Shilling Attack Detection in Recommender Systems
	1 Introduction
	2 Related Works
		2.1 Traditional Maching Learning-Based Models
		2.2 Deep Neural-Based Models
		2.3 Prototypical Network
	3 Preliminaries
		3.1 Problem Definition
		3.2 Feature Generation
	4 Approach
		4.1 Framework
		4.2 Variational Auto-encodering Embedding
		4.3 Iterative Prototype-Classification Assignment
		4.4 Summary and Implementation Details
	5 Experiments
		5.1 Dataset
		5.2 Evaluating Metrics
		5.3 Baselines
		5.4 Experimental Details
		5.5 Results
	6 Conclusion
	References
Knowledge Graph Based Chicken Disease Diagnosis Question Answering System
	1 Introduction
	2 Related Works
		2.1 Automated Poultry Disease Diagnosis
		2.2 Intelligent Question Answering System
		2.3 Knowledge Graph Based QA System
	3 Knowledge Graph Construction
	4 Question Answering System Based on Knowledge Graph
		4.1 Overview
		4.2 Intent Recognition
		4.3 Entity Recognition
	5 Experiments
		5.1 Datasets
		5.2 Experiment Setup
		5.3 Evaluation Metrics
		5.4 Experimental Results
		5.5 Demonstration
	6 Conclusion
	References
Therapeutic Effects of Corticosteroids for Critical and Severe COVID-19 Patients
	1 Introduction
	2 Materials and Methods
		2.1 Study Design and Participants
		2.2 Data Collection
		2.3 Inclusion and Exclusion Criteria
		2.4 Statistical Analysis
	3 Results
		3.1 Characteristic and K-M Survival Curves
		3.2 Univariate and Multivariable Cox Regression
		3.3 Differences in Physiological Indicators
	4 Discussion
	5 Limitations
	References
Big Data Analysis
Secure Cross-User Fuzzy Deduplication for Images in Cloud Storage
	1 Introduction
	2 Related Work
	3 Proposed Secure Cross-User Fuzzy Deduplication Scheme
		3.1 The Framework
		3.2 The Proposed Scheme
	4 Performance Evaluations
		4.1 Deduplication Performance
		4.2 Security Evaluation
	5 Conclusion
	References
Blockchain-Based Integrity Auditing with Secure Deduplication in Cloud Storage
	1 Introduction
	2 Definitions and Preliminaries
		2.1 System Model
		2.2 Threat Model
		2.3 Design Goals
	3 The Proposed Scheme
		3.1 Construction of BIAD
		3.2 Security Analysis
	4 Performance Evaluation
		4.1 Storage Overhead
		4.2 Communication Overhead
		4.3 Computation Overhead
	5 Conclusion
	References
Name Disambiguation Based on Entity Relationship Graph in Big Data
	1 Introduction
	2 The Name Disambiguation Method Based on Commonly Used Author Information
		2.1 Construct Entity Relationship Graph
		2.2 Calculate Connection Strength between Authors
	3 Experiments
		3.1 Experiments Data Sets
		3.2 Baselines
		3.3 Evaluation Indicators
		3.4 Results
	4 Conclusions
	References
Ontology-Based Metadata Model Design of Data Governance System
	1 Introduction
	2 Metadata Model Technology
		2.1 Metadata Model Classification
		2.2 Structure Metadata Model
		2.3 Discovery Metadata Model
		2.4 Management Metadata Model
		2.5 The Data Reference Model of FEA
	3 Metadata System Architecture
		3.1 Data Governance System Requirements
		3.2 Based on Ontology Metadata Modeling Method
		3.3 Metadata Model System
	4 Experimental Results and Analysis
	5 Conclusion
	References
Ontology-Based Combat Force Modeling and Its Intelligent Planning Using Genetic Algorithm
	1 Introduction
	2 Related Works
	3 Ontology-Based Combat Force Modeling
		3.1 The Entity Modeling
		3.2 The Relationship Modeling
	4 The Combat Force Planning Optimization by Genetic Algorithm
		4.1 The Operational Task Modeling
		4.2 The Proposed Optimization Method Using Genetic Algorithm
	5 Conclusion
	References
Research on Multi-channel Retrieve Mechanism Based on Heuristic
	1 Introduction
	2 Related Work
	3 Method
		3.1 Text Retrieve Channel
		3.2 Semantic Retrieve Channel
		3.3 Intention Retrieve Channel
	4 Experiment
		4.1 Text Retrieve Channel
		4.2 Semantic Retrieve Channel
	5 Conclusion
	References
Big-Model Methods
PoetryBERT: Pre-training with Sememe Knowledge for Classical Chinese Poetry
	1 Introduction
	2 Related Work
		2.1 Pre-training Language Models
		2.2 Knowledge-Enriched Pre-training Models
		2.3 Ancient Chinese Domain Tasks
	3 PoetryBERT
		3.1 Overall Architecture
		3.2 Construction of SKG-Poetry
		3.3 Two Stage Encoder Components
		3.4 Pre-training of Classical Chinese Poetry
	4 Experiments
		4.1 Pre-training Dataset
		4.2 Parameter Settings of PoetryBERT
		4.3 Fine-Tuning on Downstream Tasks
	5 Experimental Analysis
		5.1 Ablation Study
		5.2 Case Study of Poetry Translation
	6 Conclusion
	References
Image Hide with Invertible Network and Swin Transformer
	1 Introduction
	2 Related Work
		2.1 Steganography and Image Hiding
		2.2 Invertible Neural Network
		2.3 Swin Transformer
	3 Approach
		3.1 Network Architecture
		3.2 Wavelet Domain Hiding
		3.3 Loss Function
	4 Experiment
		4.1 Database and Experimental Setting
		4.2 Evaluation Metrics
	5 Conclusion
	References
Modeling and Analysis of Combat System Confrontation Based on Large-Scale Knowledge Graph Network
	1 Introduction
	2 General Framework for Combat System Analysis
	3 Modeling of the Combat System Knowledge Graph
		3.1 Combat System Representation Based on Knowledge Graph
		3.2 Mapping Modeling of Large Scale Combat System of System
	4 Multilevel Analysis of Combat System Effectiveness
		4.1 Capability Evaluation Process of Networked Combat System
		4.2 Modeling of Operational Effectiveness Evaluation Combat System
		4.3 Analysis and Calculation of Multi-level Typical System Combat Capability
	5 System Modeling and Collaborative Analysis Application
	6 Summary and Conclusions
	References
Generating Adversarial Examples and Other Applications
Generating Adversarial Malware Examples for Black-Box Attacks Based on GAN
	1 Introduction
	2 Architecture of MalGAN
		2.1 Overview
		2.2 Generator
		2.3 Substitute Detector
	3 Training MalGAN
	4 Experiments
		4.1 Experimental Setup
		4.2 Experimental Results
		4.3 Comparison with the Gradient Based Algorithm to Generate Adversarial Examples
		4.4 Retraining the Black-Box Detector
	5 Conclusions
	References
Defending Adversarial Examples by Negative Correlation Ensemble
	1 Introduction
	2 Related Work
		2.1 Adversarial Examples
		2.2 Adversarial Robustness of Ensembles
		2.3 Negative Correlation Learning
	3 The Proposed Method
		3.1 On Gradient Directions of Members
		3.2 On the Gradient Magnitudes of Members
		3.3 The Proposed NCEn
	4 Experiments
		4.1 Experimental Setup
		4.2 Defence Performance
		4.3 Transferability Between Members
	5 Conclusion
	References
Accurate Decision-Making Method for Air Combat Pilots Based on Data-Driven
	1 Introduction
	2 The Decision-Making Process of Pilot Expert System
	3 Construction of Deep Network Based on Data-Driven
		3.1 Network Model Selection
		3.2 Battlefield Data Sampling Dimensions
		3.3 Loss Function and Accuracy Function
		3.4 Construction of Deep Learning Network
	4 Experiments and Results
		4.1 Data Scales and Preprocessing
		4.2 Training Result
		4.3 Data Scales Compression
	5 Conclusion
	References
Establishment of Empirical Expression of Atmospheric Scattering Coefficient for Line-of-Sight Ultraviolet Propagation in Coastal Area
	1 Introduction
	2 Applicability Verification of MODTRAN in Coastal Area Introduction
	3 Establishment of Empirical Expression of Atmospheric Scattering Coefficient of LOS Ultraviolet Propagation in Coastal Area
		3.1 Calculation of Atmospheric Scattering Coefficient
	4 Establishment of Empirical Expression of Atmospheric Scattering Coefficient in Coastal Area
	5 Conclusion
	References
Author Index