The Recent Advances in Transdisciplinary Data Science. First Southwest Data Science Conference, SDSC 2022 Waco, TX, USA, March 25–26, 2022 Revised Selected Papers

Preface
Organization
The Convergence of BI and AI: With Applications for the Better Management of Higher Education (Keynote)
Contents
Business and Social Data Science
Forecasting Stock Excess Returns with SEC 8-K Filings
	1 Introduction
	2 Data Preprocessing, Text Vectorization, and Imbalanced Learning Forecast
		2.1 Text Vectorization
	3 Dimension Reduction Stacking
	4 Results
		4.1 Dimension Reduction Stacking Under the BERT Vectorization
		4.2 Stock Excess Return Forecast Under Imbalance Resampling
	5 Discussion and Conclusion
	References
A Fast Initial Response Approach to Sequential Financial Surveillance
	1 Introduction
	2 Geometric Random Walk with Log-Student's t Increments
	3 Robust Model Calibration
		3.1 Minimum Covariance Determinant (MCD)
		3.2 Minimum Density Power Divergence (MDPD)
	4 Statistical Process Monitoring
		4.1 Taut String (TS) Fast Initial Response (FIR) Chart
	5 Example
	6 Summary and Conclusions
	References
Evaluation of High-Value Patents in the Reverse Innovation: A Theory and Case Study from China
	1 Introduction
	2 Literature Review and Theoretical Foundation
		2.1 Reverse Innovation of Local Companies
		2.2 High-Value Patents
		2.3 Evaluation of Patent Value
		2.4 The Connotation of High-Value Patent in the Reverse Innovation
	3 The Methods of Selection and Evaluation of High-Value Patents
		3.1 The Selection of High-Value Patents in the Reverse Innovation
		3.2 The Process of Evaluation of High-Value Patents
	4 Case Study
		4.1 BYD Company and Its Reverse Innovation
		4.2 Selection of High-Value Patents in BYD'S Reverse Innovation
		4.3 Discussion
	5 Conclusions
	References
Research on the Influence of Different Fairness Reference Points on the Supply Chain Enterprises
	1 Introduction
	2 Model Description
	3 Decisions Under Different Fairness Reference Points
		3.1 SUpplier’s Profit Reference Point
		3.2 Nash Bargaining Solution Reference Point
		3.3 FIrm’s Contribution Reference Point
	4 Comparing the Optimal Order Quantities
	5 Numerical Analysis
		5.1 The Impact of Parameter
		5.2 The Impact of Parameters  and 
	6 Conclusion
	Appendix
	References
Health and Biological Data Science
EGRE: Calculating Enrichment Between Genomic Regions
	1 Introduction
	2 Results
		2.1 EGRE is a Customizable Tool to Calculate Enrichment Between Pairs of Genomic Interval Files
		2.2 GC Matched Permutations Improve the Accuracy of Enrichment Results
		2.3 Parallelization Using Multithreading Reduces the Elapsed Time for Analysis
	3 Methods
		3.1 Implementation of EGRE
		3.2 Implementation of GC Matching for Shuffled Genomic Regions
		3.3 Functional Genomics and Simulated Data Used for Experiments
		3.4 Benchmark Analysis
		3.5 Code Availability
	4 Conclusions
	References
THSLRR: A Low-Rank Subspace Clustering Method Based on Tired Random Walk Similarity and Hypergraph Regularization Constraints
	1 Introduction
	2 Method
		2.1 Sparse Low-Rank Representation
		2.2 Hypergraph Regularization
		2.3 Tired Random Walk
		2.4 Objective Function of THSLRR
		2.5 Optimization Process and Spectral Clustering of THSLRR Method
	3 Results and Discussion
		3.1 Evaluation Measurements
		3.2 scRNA-seq Datasets
		3.3 Parameters Setting
		3.4 Comparative Analysis of Clustering
		3.5 Visualize Cells Using t-SNE
		3.6 Gene Markers Prioritization
	4 Conclusion
	References
Traceability Analysis of Feng-Flavour Daqu in China
	1 Introduction
	2 Materials and Methods
		2.1 Sample Collection
		2.2 Microbial High-Throughput Sequencing Analysis of Traceable Samples
		2.3 Traceability Analysis of Brewing Microorganisms
	3 Results
		3.1 Analysis of Fungal Community Diversity in Feng-Flavour Daqu and Its Environment
		3.2 Diversity Analysis of Bacteria and Environmental Communities in Feng-Flavour Daqu
		3.3 Microbial Traceability Analysis of Feng-Flavour Daqu
	4 Discussion and Conclusion
	References
Visualization of Functional Assignment of Disease Genes and Mutations
	1 Introduction
	2 Identification and Visualization of Biological Process Functions
	3 Data, Experiments, and Result
		3.1 Gene and Mutation Data
		3.2 Result with UniProt Humsavar Mutations Data
		3.3 Results with the ClinVar Mutation Database
	4 Results and Discussion
	5 Conclusions
	References
Applied Data Science, Artificial Intelligence, and Data Engineering
Characterizing In-Situ Solar Wind Observations Using Clustering Methods
	1 Introduction: The Solar Wind
	2 K-Means: Spatially Separate Clusters
		2.1 ICME Characterization
	3 UMAP: Latent Features
	4 Conclusion
	References
An Improved Capsule Network for Speech Emotion Recognition
	1 Introduction
	2 The Proposed Methods
	3 Datasets and Feature Extraction
		3.1 Baseline Datasets
		3.2 Augmented Datasets
		3.3 Feature Extraction
	4 Experiments and Analysis
		4.1 Experimental Setup
		4.2 Model Parameter Configuration
		4.3 Data Division Methods
		4.4 Experimental Results and Analysis
	5 Conclusion
	References
Distributed Query Processing and Reasoning Over Linked Big Data
	1 Introduction
	2 Related Work
	3 Method
		3.1 Data Storage Format, Query Planning, Evaluation, and Optimization
	4 Evaluation
		4.1 System Setup
		4.2 Dataset
		4.3 Results
	5 Conclusion
	References
Normal Equilibrium Fluctuations from Chaotic Trajectories: Coupled Logistic Maps
	1 Introduction
	2 Discussion
	3 Conclusions
	References
Matching Code Patterns Across Programming Language
	1 Introduction
	2 ReSSA
	3 Parser Construction
	4 Conclusion
	References
Partitionable Programs Using Tyro V2
	1 Introduction
	2 Related Work
		2.1 Tyro
		2.2 Similar Tools
		2.3 Graphical Program Representations
	3 Tyro V2
		3.1 Program Analyzer
		3.2 Operation Translator
		3.3 Code Generator and Verifier
		3.4 Example Translated Program
	4 Graph Based Extraction
	5 The Model
	6 Conclusion and Future Work
	References
Analyzing Technical Debt by Mapping Production Logs with Source Code
	1 Introduction
	2 Background and Related Work
		2.1 Technical Debt
		2.2 Code Smells
		2.3 Architectural Degradation
		2.4 Static Code Analysis
		2.5 Tools
		2.6 Log Analysis
	3 Method
	4 Case Study
		4.1 Threats to Validity
	5 Conclusion and Future Work
	References
EB-FedAvg: Personalized and Training Efficient Federated Learning with Early-Bird Tickets
	1 Introduction
	2 Related Works
	3 Design of EB-fedAvg
		3.1 Training Algorithm
		3.2 Structured Pruning
		3.3 Early Bird Ticket Masked Weights
		3.4 Aggregate Heterogeneous Weight Masks
		3.5 Generate Personalized Model
	4 Evaluation
		4.1 Datasets, and Models
		4.2 Hyper-Parameter Setting
		4.3 Compared Methods
		4.4 Evaluation Metrics
		4.5 Training Performance
	5 Conclusion
	References
Author Index