內容大鋼
This book discusses the detection of multiple confounders in causal inference; the influence of standardization of non-confounders on the precision of causal effects; the estimation of causal effects with missing data. It provides the basic theory and method of causal inference for beginners. The theoretical significance is that using mathematical models to describe the assumptions of causal inference and using feasible algorithms to eliminate non-confounding factors, which could improve the estimation precision of causal effects. The theoretical research results can be applied in epidemiology, biomedicine, food science, economics, information science and other fields. The purpose of this book is to introduce the latest research results of the author and her collaborators on the confounders, nonconfounders, causal effects estimation and Bayesian statistics in causal inference in the past few years. The book consists of six chapters. The first chapter mainly introduces the confounders, non-confounders, compressibility, homogeneity, and subpopulation without confounding, the detection and elimination of confounders, and the relationship between them in causal inference; Chapter two discusses whether to standardize the non-confounders in the estimation of causal effects and relative risks, and the precision changes with and without standardization for a non-confounder. Chapter three discusses the estimation of causal effects with missing data; Chapter four discusses how to establish a Bayesian hierarchical model for the real data; Chapter five discusses the estimation of traffic load in wireless cellular networks. Based on the temporal and spatial characteristics of different base stations, a shrinkage estimation method is proposed to make variable selection. Chapter six discusses the identification and authentication of radio frequency signals.
目錄
Preface
Chapter 1 Confounder and Non-confounder
1.1 Relations among Homogeneity， Collapsibility and Non-confounding
1.1.1 Basic Knowledge and Related Research
1.1.2 Non-confounding， Homogeneity and Collapsibility
1.1.3 Uniformly Non-confounding
1.2 Uniformly Non-confounding of Causal Distribution Effects
1.2.1 Introduction
1.2.2 Concepts of Non-confounding over Multiple Covariates
1.2.3 Condition for Non-confounding over Multiple Covariates
1.2.4 Discussion and Future Plans
1.3 How to Detect Multiple Confounders
1.3.1 Introduction
1.3.2 Concepts
1.3.3 Uniform Non-confounding over Multiple Covariates
1.3.4 Non-confounding in Subpopulations
1.3.5 Some Examples
References
Chapter 2 Whether to Adjust for a Non-confounder
2.1 Whether to Adjust for a Non-confounder
2.1.1 Related Research and Controversy
2.1.2 Confounding Bias， Confounder and Standardization
2.1.3 Estimates of Hypothetical Proportion
2.1.4 Expectation and Variances of Estimates
2.1.5 Proofs of Theorems
2.2 The Estimation of Log Relative Risk with Non-confounder
2.2.1 Introduction of Basic Knowledge
2.2.2 Asymptotical Estimators
2.2.3 Counterexamples of Theorem 2.2.3
2.2.4 Conclusion and Discussion
References
Chapter 3 Binomial Parameter Estimation with Missing Data
3.1 Binomial Proportion Estimation with Missing Data
3.1.1 Introduction
3.1.2 The Model and Notations
3.1.3 Asymptotic Variance Estimation
3.1.4 Simulation Results
3.1.5 Concluding Remarks
3.2 Using Auxiliary Data for Binomial Parameter Estimation with Missing Data
3.2.1 Basic Knowledge and Related Research
3.2.2 Concepts and Some Notations
3.2.3 Maximum Likelihood Estimator of Variance
3.2.4 Simulation Study
3.2.5 Application of Real Data
3.2.6 Concluding Remarks
References
Chapter 4 A Bayesian Approach and its Applications
4.1 Forecasting of COVID-19 Onset Cases
4.1.1 Related Works
4.1.2 Construction of the Model
4.1.3 Results of Statistical Analysis
4.1.4 Conclusion and Discussion
4.2 A Bayesian Approach to Forecast Chinese Foodborne Diseases
4.2.1 Related Basic Knowledge
4.2.2 Data Material and Statistical Analysis
4.2.3 Main Results
4.2.4 Conclusion and Discussion
References
Chapter 5 Statistical Analysis and Inference of Traffic Load
5.1 Traffic Load Prediction Based on Shrinkage Estimation
5.1.1 Introduction
5.1.2 Data Source
5.1.3 Model Construction and Parameters Prediction
5.1.4 Discussion
5.2 Spatial-temporal Analysis of Traffic Load in Mobile Cellular Network
5.2.1 Introduction
5.2.2 Model Construction
5.2.3 Conclusion and Discussion
References
Chapter 6 Identification and Authentication of Radio Frequency Signal
6.1 Identification and Authentication for Wireless Transmission Security
6.1.1 Introduction
6.1.2 Experimental Process
6.1.3 Statistic Fingerprint Generation
6.1.4 Classification Methods
6.1.5 Results and Discussion
6.1.6 Conclusion
6.2 Radio Frequency Signal Identification Using Transfer Learning Based on LSTM
6.2.1 Overview of Basic Knowledge
6.2.2 Related Works
6.2.3 Methods
6.2.4 Experimental Process
6.2.5 Conclusion
References
Color Figures
[