Data Science and SDGs: Challenges, Opportunities and Realities

Description

The book presents contributions on statistical models and methods applied, for both data science and SDGs, in one place. Measuring and controlling data of SDGs, data driven measurement of progress needs to be distributed to stakeholders. In this situation, the techniques used in data science, specially, in the big data analytics, play an important role rather than the traditional data gathering and manipulation techniques. This book fills this space through its twenty contributions. The contributions have been selected from those presented during the 7th International Conference on Data Science and Sustainable Development Goals organized by the Department of Statistics, University of Rajshahi, Bangladesh; and cover topics mainly on SDGs, bioinformatics, public health, medical informatics, environmental statistics, data science and machine learning. The contents of the volume would be useful to policymakers, researchers, government entities, civil society, and nonprofit organizations for monitoring and accelerating the progress of SDGs. Foreword Preface Pictures Conference Contents Editors and Contributors SDGs in Bangladesh: Implementation Challenges and Way Forward 1 SDGs and Bangladesh’s Engagement 2 Brief Note on Implementation of SDGs 3 The “Whole of Society” Approach and Coordination Among Different Stakeholders in the Implementation of SDGs in Bangladesh 4 Successes Achieved So Far in SDGs Implementation in Bangladesh 5 Revisiting Challenges of SDGs Implementation in Bangladesh 6 LDC Graduation, International Cooperation, Resource Mobilization, and Capacity Development in Achieving SDG Targets 7 Future Roadmaps References Some Models and Their Extensions for Longitudinal Analyses 1 Introduction 2 Regression and Multivariate Techniques 2.1 The Growth Curve Model 3 Some Extensions of the Growth Curve Model 3.1 Random Effects Growth Curve Model 3.2 Measurement Errors 3.3 Spline Growth Model 4 Models Based on Finite Mixtures 4.1 Introduction 4.2 Data Analysis: Analysis of Drinking Profiles 4.3 Extension: Semiparametric Mean Model 5 Clustering Techniques for Categorical Longitudinal Data: Factory Downsizing References Association of IL-6 Gene rs1800796 Polymorphism with Cancer Risk: A Meta-Analysis 1 Introduction 2 Methods and Materials 2.1 Search Strategy 2.2 Eligibility Criteria 2.3 Data Extraction 2.4 Statistical Analysis 3 Results 3.1 Study Characteristics 3.2 Quantitative Synthesis 3.3 Source of Heterogeneity 4 Publication Bias 5 Discussions References Two Level Logistic Regression Analysis of Factors Influencing Dual Form of Malnutrition in Mother–Child Pairs: A Household Study in Bangladesh 1 Introduction 2 Methods 2.1 Outcome Variable 2.2 Independent Variables 2.3 Statistical Analysis 3 Results 3.1 Prevalence of Nutritional Status Among Mother–Child Pairs 3.2 Two-Level Binary Logistic Regression Model 4 Discussion 4.1 Prevalence of Under Nutrition 4.2 Effect of Socioeconomic and Demographic Factors on Under Nutrition 5 Conclusion 6 Limitation of the Study References Divide and Recombine Approach for Analysis of Failure Data Using Parametric Regression Model 1 Introduction 2 Parametric Regression Model 3 Divide and Recombine 4 Example 5 Simulation Study 6 Conclusion Appendix: Programming Codes in R References Performance of Different Data Mining Methods for Predicting Rainfall of Rajshahi District, Bangladesh 1 Introduction 2 Materials and Methods 2.1 Data Collection 2.2 Data Pre-Processing 2.3 Predictive Models 2.4 Model Evaluation 3 Results and Discussions 4 Conclusions References Generalized Vector Autoregression Controlling Intervention and Volatility for Climatic Variables 1 Introduction 2 Materials 3 Methods 4 Results 5 Discussion 6 Conclusion References Experimental Designs for fMRI Studies in Small Samples 1 Introduction 2 Estimability Issues 3 Choice of N and Dn for Given K* 4 Comparison of Design Sequences 5 Structural Balance in and Orthogonality of a Design Sequence 6 Use of Hadamard Matrices References Bioinformatic Analysis of Differentially Expressed Genes (DEGs) Detected from RNA-Sequence Profiles of Mouse Striatum 1 Introduction 2 Materials and Methods 2.1 RNA-Seq Data Collection 2.2 Methods for Identification of DEGs 2.3 DESeq2 2.4 edgeR 2.5 Limma 2.6 Methods for Functional Analysis of DEGs 2.7 PPI Analysis of DEGs 2.8 GO Enrichment and KEGG Pathway Analysis of DEGs 2.9 miRNAs-Target Gene Interactions of DEGs 2.10 Downstream Analysis of DEGs 3 Results 3.1 Identified DEGs 3.2 PPI Analysis of DEGs 3.3 GO Enrichment Analysis of DEGs 3.4 KEGG Pathway Analysis of DEGs 3.5 miRNA–mRNA Network Construction for DEGs 3.6 Downstream Analysis for DEGs 4 Discussion 5 Conclusions References Role of Serum High-Sensitivity C-Reactive Protein Level as Risk Factor in the Prediction of Coronary Artery Disease in Hyperglycemic Subjects 1 Introduction 2 Materials and Methods 2.1 Selection of Study Participants 2.2 Ethical Permission 2.3 Blood Pressure Measurement 2.4 Collection of Specimen 2.5 Biochemical Investigations 2.6 Statistics 3 Results 3.1 Descriptive Characteristics of Study Participants 3.2 Association of Fasting Plasma Glucose and hs-CPR with Cardiac Risk Factors 3.3 Serum hs-CRP Level and FPG 4 Discussion 5 Conclusion References Identification of Outliers in Gene Expression Data 1 Introduction 2 Existing Outlier Detection Methods for the Gene Expression Data 3 Proposed Outlier Detection Method 4 Monte Carlo Comparison of Different Outlier Detection Methods 5 Conclusions References Selecting Covariance Structure to Analyze Longitudinal Data: A Study to Model the Body Mass Index of Primary School-Going Children in Bangladesh 1 Introduction 2 Materials and Methods 3 Results and Discussion 4 Conclusion References Statistical Analysis of Various Optimal Latin Hypercube Designs 1 Introduction 2 Experimental Result and Discussion 3 Conclusion References Erlang Loss Formulas: An Elementary Derivation 1 Introduction 2 An Elementary Derivation 3 Answering Questions 1–3 4 Application 5 Conclusion References Machine Learning, Regression and Optimization 1 Introduction 1.1 Programming Versus Machine Learning 1.2 Data Science and Machine Learning 1.3 Some Essential Numerical Optimization Concepts and Techniques for Machine Learning 1.4 Types of Learning 1.5 Basic Steps of Machine Learning 2 Linear Regression Analysis and Least-Squares Techniques 2.1 Multivariable Linear Regression 2.2 Least Squares Solution of the Regression Model 2.3 Solution of the Least-Squares Problem in Optimization Setting 2.4 Properties of the Least-Squares Estimator 3 Computational Algorithms for Computing Linear Least-Squares Solution 3.1 The Cholesky Method for Least Squares Solution 3.2 The QR Factorization Method for Least-Squares Solution 3.3 The SVD Method for Least-Squares Solution 4 Nonlinear Regression Models 4.1 The Gauss–Newton Method 4.2 The Levenberg–Marquardt Method 4.3 Results of Comparison of the Three Methods on a Toy Example (Heath 2018) 5 Numerical Experiments 5.1 Predicting House Price in the City of Boston 5.2 Predicting Number of Coronavirus (COVID-19) Cases in the State of New York 5.3 Portfolio Optimization 6 Concluding Remarks References