You are previewing Predictive Modeling Applications in Actuarial Science.
O'Reilly logo
Predictive Modeling Applications in Actuarial Science

Book Description

Predictive modeling uses data to forecast future events. It exploits relationships between explanatory variables and the predicted variables from past occurrences to predict future outcomes. Forecasting financial events is a core skill that actuaries routinely apply in insurance and other risk-management applications. Predictive Modeling Applications in Actuarial Science emphasizes life-long learning by developing tools in an insurance context, providing the relevant actuarial applications, and introducing advanced statistical techniques that can be used to gain a competitive advantage in situations with complex data. Volume 2 examines applications of predictive modeling. Where Volume 1 developed the foundations of predictive modeling, Volume 2 explores practical uses for techniques, focusing on property and casualty insurance. Readers are exposed to a variety of techniques in concrete, real-life contexts that demonstrate their value and the overall value of predictive modeling, for seasoned practicing analysts as well as those just starting out.

Table of Contents

  1. Cover
  2. Half title
  3. Series
  4. Title
  5. Copyright
  6. Table of Contents
  7. Contributors
  8. Preface
  9. Acknowledgments
  10. 1 Pure Premium Modeling Using Generalized Linear Models
    1. 1.1 Introduction
    2. 1.2 Data Characteristics
    3. 1.3 Exploratory Data Analysis
    4. 1.4 Frequency Modeling
    5. 1.5 Severity Modeling
    6. 1.6 Pure Premium
    7. 1.7 Validation
    8. 1.8 Conclusions
    9. References
  11. 2 Applying Generalized Linear Models to Insurance Data
    1. 2.1 Introduction
    2. 2.2 Comparing Model Forms
    3. 2.3 The Dataset and Model Forms
    4. 2.4 Results
    5. Appendix 2.A Proof of Equivalence between Pure Premium Model Forms
    6. Conclusion
    7. Appendix 2.B The Gini Index
    8. References
  12. 3 Generalized Linear Models as Predictive Claim Models
    1. 3.1 Review of Loss Reserving
    2. 3.2 Additional Notation
    3. 3.3 GLM Background
    4. 3.4 Advantages of GLMs
    5. 3.5 Diagnostics
    6. 3.6 Example
    7. 3.7 Conclusion
    8. References
  13. 4 Frameworks for General Insurance Ratemaking
    1. 4.1 Introduction
    2. 4.2 Data
    3. 4.3 Univariate Ratemaking Framework
    4. 4.4 Multivariate Ratemaking Frameworks
    5. 4.5 Model Comparisons
    6. 4.6 Conclusion
    7. References
  14. 5 Using Multilevel Modeling for Group Health Insurance Ratemaking
    1. 5.1 Motivation and Background
    2. 5.2 Data
    3. 5.3 Methods and Models
    4. 5.4 Results
    5. 5.5 Conclusions
    6. Acknowledgments
    7. Appendix
    8. References
  15. 6 Clustering in General Insurance Pricing
    1. 6.1 Introduction
    2. 6.2 Overview of Clustering
    3. 6.3 Dataset for Case Study
    4. 6.4 Clustering Methods
    5. 6.5 Exposure-Adjusted Hybrid (EAH) Clusering Method
    6. 6.6 Results of Case Study
    7. 6.7 Other Considerations
    8. 6.8 Conclusions
    9. References
  16. 7 Application of Two Unsupervised Learning Techniquesto Questionable Claims
    1. 7.1 Introduction
    2. 7.2 Unsupervised Learning
    3. 7.3 Simulated Automobile PIP Questionable Claims Data and the Fraud Issue
    4. 7.4 The Questionable Claims Dependent Variable Problem
    5. 7.5 The PRIDIT Method
    6. 7.6 Processing the Questionable Claims Data for PRIDIT Analysis
    7. 7.7 Computing RIDITS and PRIDITS
    8. 7.8 PRIDIT Results for Simulated PIP Questionable Claims Data
    9. 7.9 How Good Is the PRIDIT Score?
    10. 7.10 Trees and Random Forests
    11. 7.11 Unsupervised Learning with Random Forest
    12. 7.12 Software for Random Forest Computation
    13. 7.13 Some Findings from the Brockett et al. Study
    14. 7.14 Random Forest Visualization via Multidimensional Scaling
    15. 7.15 Kohonen Neural Networks
    16. 7.16 Summary
    17. References
  17. 8 The Predictive Distribution of Loss Reserve Estimates over a Finite Time Horizon
    1. 8.1 Introduction
    2. 8.2 The CAS Loss Reserve Database
    3. 8.3 The Correlated Chain Ladder Model
    4. 8.4 The Predictive Distribution of Future Estimates
    5. 8.5 The Implications for Capital Management
    6. 8.6 Summary and Conclusions
    7. References
  18. 9 Finite Mixture Model and Workers’ Compensation Large-Loss Regression Analysis
    1. 9.1 Introduction
    2. 9.2 DGLM and FMM
    3. 9.3 Data
    4. 9.4 Traditional Distribution Analysis
    5. 9.5 Univariate and Correlation Analyses
    6. 9.6 Regression Analysis
    7. 9.7 Conclusions
    8. References
  19. 10 A Framework for Managing Claim Escalation Using Predictive Modeling
    1. 10.1 Introduction
    2. 10.2 Loss Development Models
    3. 10.3 Additional Data for Triage Models
    4. 10.4 Factor Selection
    5. 10.5 Modeling Method
    6. 10.6 Conclusions
    7. 10.7 Further Research Opportunities
    8. Appendix: Penalized Regression
    9. References
  20. 11 Predictive Modeling for Usage-Based Auto Insurance
    1. 11.1 Introduction to Usage-Based Auto Insurance
    2. 11.2 Poisson Model for Usage-Based Auto Insurance
    3. 11.3 Classification Trees
    4. 11.4 Implementing UBI Models with a Traditional Rating Plan
    5. 11.5 Summary and Areas for Future Research
    6. Acknowledgments
    7. References
  21. Index