Frontiers in Statistical Quality Control 13

Preface
Contents
Contributors
*-1.5pc Statistical Process Control
Use of Conditional False Alarm Metric in Statistical Process Monitoring
1 Introduction
2 Static Use of CFAR
3 Dynamic Use of CFAR
4 Implementation Methods
5 Other Applications
6 Conclusions
References
Design Considerations and Trade-offs for Shewhart Control Charts
1 Introduction
2 Parameter Estimation
3 Design Criteria
3.1 Bias Criterion
3.2 Exceedance Probability Criterion
4 Distributional Assumptions
4.1 Parametric Methods
4.2 Nonparametric Methods
5 Sample Size and Strictness
6 Concluding Remarks
References
On the Calculation of the ARL for Beta EWMA Control Charts
1 Introduction
2 The Beta Distribution and the EWMA Control Chart
3 Numerical Methods to Calculate EWMA ARL
4 Comparison Study
5 Conclusions
References
Flexible Monitoring Methods for High-yield Processes
1 Introduction
2 The Zero-Inflated Conway-Maxwell-Poisson (ZICOM-Poisson) Distribution
3 Monitoring Methods Based on ZICOM-Poisson Distribution
3.1 The p-λ CUSUM Chart
3.2 The CRL-ZTCOMP CUSUM Chart
4 Performance Evaluations
4.1 Performance Measure
4.2 IC and OOC Parameters for the Simulation Study
4.3 Derivation of Control Limits
5 Results and Discussions
6 An Example
7 Summary, Conclusions, and Recommendations
References
An Average Loss Control Chart Under a Skewed Process Distribution
1 Introduction
2 The ML Control Chart
2.1 The Skew-Normal Distribution
2.2 The Loss Function
2.3 The Design of a Median Loss Control Chart
3 Performance Measurement of the ML Chart
4 The Average Loss Control Chart
4.1 The Distribution of Average Loss
4.2 The Design of an Average Loss Control Chart
5 Performance Measurement of the ALSN Control Chart
6 Conclusions
References
ARL-Unbiased CUSUM Schemes to Monitor Binomial Counts
1 Introduction
1.1 A Handful of CUSUM Charts and Schemes for Independent Counts
1.2 A Few CUSUM Charts and Schemes for Autocorrelated Counts
1.3 On ARL-Unbiased Charts for Discrete Output
2 The ARL of CUSUM Charts and Schemes for i.i.d. Counts
2.1 Upper and Lower One-Sided CUSUM Charts
2.2 The Two-Sided CUSUM Scheme
2.3 Relating the ARL of One-Sided CUSUM Charts and the Two-Sided CUSUM Scheme
3 The ARL-Unbiased Two-Sided CUSUM Scheme for i.i.d. Binomial Output
3.1 The Control Limits and Randomization Probabilities
3.2 Preliminary Results
4 The ARL of Two-Sided CUSUM Schemes for Binomial AR(1) Counts
4.1 Overall ARL Functions
4.2 Further Preliminary Results
5 Final Thoughts
References
Statistical Aspects of Target Setting for Attribute Data Monitoring
1 Introduction
2 Robust Point Estimation of p
2.1 Weight-Trimmed Estimate for p
2.2 Confidence Bounds for p
2.3 Bias and Robustness Issues
3 Target Derivation Process
3.1 The Processing Parameters
3.2 The Yardstick
3.3 Decisions
4 Example of Implementation
5 Discussion
References
MAV Control Charts for Monitoring Two-State Processes Using Indirectly Observed Binary Data
1 Introduction
2 Indirect Evaluation of States Using State-Related Observational Data
2.1 Classifiers for Binary Classification—Logistic Regression
2.2 Combined Classifiers
3 Monitoring the State-Related Observational Data
4 Application—The Case of Bipolar Disorder
5 Conclusions and Future Research
References
Monitoring Image Processes: Overview and Comparison Study
1 Introduction
2 Image Analysis
2.1 Digital Image Fundamentals
2.2 Statistical Image Analysis
3 Monitoring Procedures for the Pixel Process in the Time Domain
3.1 Model
3.2 Residual Charts
3.3 Control Charts Based on the GLR Approach
4 Comparison Study
5 Conclusions
References
Parallelized Monitoring of Dependent Spatiotemporal Processes
1 Introduction
2 Spatial Dependence Models
3 Monitoring Procedure for High-Resolution Images
3.1 Out-of-Control Model
3.2 Parallel Monitoring
3.3 Control Characteristic and Multivariate EWMA Chart
4 Monte Carlo Simulation Study
4.1 Calibration
4.2 Performance of the Proposed Parallelized Chart
5 Conclusion
References
Product's Warranty Claim Monitoring Under Variable Intensity Rates
1 Introduction
2 Modelling the Warranty Claims
3 Control Chart for Monitoring Claims with Heterogeneity Poisson Intensities
4 Performance Evaluation
5 A Real Data Example
6 Conclusions
References
A Statistical (Process Monitoring) Perspective on Human Performance Modeling in the Age of Cyber-Physical Systems
1 Introduction
2 Background
3 Wearables for Occupational Fatigue Management
3.1 Importance of the Domain
3.2 An Illustrative Example
3.3 Opportunities for Statistical (Process Control) Research
4 The Use of On-Board Vehicular Sensors to Capture Changes in Driver's Safety Performance
4.1 Importance of the Domain
4.2 An Illustrative Example
4.3 Opportunities for Statistical (Process Control) Research
5 Biometric-Driven Computer Security
5.1 Importance of the Domain
5.2 An Illustrative Example
5.3 Data Analysis
5.4 Opportunities for Statistical (Process Control) Research
6 Concluding Remarks
References
Monitoring Performance of Surgeons Using a New Risk-Adjusted Exponentially Weighted Moving Average Control Chart
1 Introduction
2 Real Dataset and Review of Risk-Adjustment Mechanism
3 Risk-Adjusted EWMA Chart
4 Comparison of Average Run Lengths
5 Analyses of Surgeons' Data
6 Conclusions
References
Exploring the Usefulness of Functional Data Analysis for Health Surveillance
1 Introduction
2 Functional Data Analysis
3 FDA Based Statistical Process Monitoring
3.1 Control Chart for a Whole Curve
3.2 Control Chart for Individual Observations
3.3 Performance Measurement
4 Case Study
4.1 Data Description
4.2 Fitted MOI by Functional Data Analysis
4.3 Implementation of Control Chart
5 Conclusions
References
Rapid Detection of Hot-Spot by Tensor Decomposition with Application to Weekly Gonorrhea Data
1 Introduction
2 Data Description
3 Proposed Model
3.1 Tensor Algebra and Notation
3.2 Our Proposed SSR-Tensor Model
3.3 Estimation of Hot-Spots
4 Hot-Spot Detection
4.1 Detect When the Hot Spot Occurs
4.2 Localize Where and Which the Hot Spot Occur?
5 Optimization Algorithm
5.1 Procedure of Our Algorithm
5.2 Computational Complexity
6 Simulation
6.1 Generative Model in Simulation
6.2 Hot-Spot Detection Performance
7 Case Study
7.1 When the Temporal Changes Happen?
7.2 In Which State and Week Do the Spatial Hot-spots Occur?
References
An Approach to Monitoring Time Between Events When Events Are Frequent
1 Introduction
2 Monitoring TBE for Homogeneous Processes
3 Monitoring Non-homogeneous Processes
4 Example of Application
5 Diagnosing the Nature of the Outbreak
6 Conclusions
References
*-1.5pc Selected Topics from Statistical Quality Control
Analysis of Measurement Precision Experiment with Ordinal Categorical Variables
1 Introduction
2 Data
3 Methods
3.1 ISO 5725
3.2 Ordinal Analysis of Variance (ORDANOVA)
3.3 Attribute Agreement Analysis (AAA)
3.4 Item Response Theory (IRT)
4 Results and Discussion
4.1 Graphical Presentation
4.2 Estimation of Precision Measures
4.3 Estimation of Toxicity
5 Conclusions
References
Assessing a Binary Measurement System with Operator and Random Part Effects
1 Introduction
2 Fixed Operator Effects
2.1 Models
2.2 Estimation
2.3 Example
2.4 Choice of Plan
2.5 Three or More Operators
3 Random Operator Effects
3.1 Models
3.2 Estimation
3.3 Example
3.4 Choice of Plans
4 Discussion
References
Concepts, Methods, and Tools Enabling Measurement Quality
1 Preamble
2 Introduction
3 Measurement
4 Measurement Uncertainty
5 Calibration
6 Traceability
6.1 Traceability for Counts and Qualitative Measurands
7 Measurement Models
7.1 Measurement Equations
7.2 Observation Equations
8 Evaluating Measurement Uncertainty
8.1 NIST Uncertainty Machine
9 Mutual Consistency and Consensus Building
10 Summation and Conclusions
References
Assessing Laboratory Effects in Key Comparisons with Two Transfer Standards Measured in Two Petals: A Bayesian Approach
1 Introduction
2 Statistical Model for Measurement Data
3 Bayesian Inference Based on the Reference Prior
3.1 Reference Prior for the Marginal Model
3.2 Posterior for Laboratory Effects
4 Analysis of CCM.M-K7 Data
5 Conclusions
6 Appendices
6.1 Derivation of the Fisher Information Matrix
6.2 Derivation of the Posterior
References
Quality Control Activities Are a Challenge for Reducing Variability
1 Introduction
2 Value Chain in Production Processes
3 Three Kinds of Variability of Outcome in the Value Chain
4 Four Approaches to Reduce the Variability
4.1 Four Approaches in Causal Model
4.2 Approach A: Taking Action Against Effect
4.3 Approach B: Taking Action Against Cause
4.4 Approach C: Observing the Situation of Causes and Taking Action Against Effect According to the Situation
4.5 Approach D: Taking Action Against Causal Relationship
5 Four Approaches to Reduce the Three Kinds of Variability Overlooking the Value Chain
6 Further Discussions and Conclusive Remarks
References
Is the Benford Law Useful for Data Quality Assessment?
1 Introduction
2 Mathematical Basics of Benford's Law
3 Statistical Goodness-of-Fit Tests
4 Our Benchmark Data
5 Performance Analysis
6 Summary
References