Statistical and Soft Computing Approaches in Insurance Problems


Sancho Salcedo-Sanz (Editor)
Department of Signal Theory and Communications, Universidad de Alcalá, Madrid, Spain

Mercè Claramunt-Bielsa (Editor)
Universidad de Barcelona, Spain

Jose Luis Vilar-Zanón (Editor)
Universidad Complutense de Madrid, Spain

Antonio Heras (Editor)
Universidad Complutense de Madrid, Spain

Series: Computer Science, Technology and Applications
BISAC: MAT008000

This book reviews the application of different statistical and Soft-Computing (SC) techniques in insurance-related problems. The book has been divided into 5 chapters with the following structure: Chapter 1 provides a comprehensive review of SC techniques in insurance problems. This chapter includes a full description of different SC techniques, such as fuzzy logic-based approaches, evolutionary computation approaches and neural computation techniques. A review of the application of these methods in insurance problems and a case study in a real problem completes this first chapter. Chapters 2 and 3 describe two different real applications of SC techniques in insurance. Chapter 2 details an algorithm close to fuzzy logic (Rough Set) that has been applied to insurance problems in the last few years. In this chapter, the authors describe the main concepts related to the algorithm and discuss different application of the algorithm in the insurance sector.

Chapter 3 is devoted to a state-of-the-art technique in neural computation (Support Vector Machines, SVM). The authors describe the hybridization of SVM with genetic algorithms and Random Forest approaches in a crucial problem of car insurance: the detection of the most important risk factors affecting the claims. Chapter 4 presents a new mathematical tool based on modal intervals that allow us to process interval data (reflecting the fact that the data we are working with is not exact) and to interpret semantically the results obtained. The authors also include some examples related to the world of economics and insurance. Chapter 5 concentrates on two well-known risk measures: the Value at Risk and the Tail Value at Risk. The authors present a new analytical expression of the Tail Value at Risk using the Normal-Power approximation and they analyze its precision. (Imprint: Novinka )





Table of Contents


Chapter 1. A Review of Computational Intelligence Algorithms in Insurance Applications
(S. Salcedo-Sanz, L. Cuadra, A. Portilla-Figueras, S. Jimenez-Fernandez and E. Alexandre, Department of Signal Theory and Communications, Universidad de Alcala, Madrid, Spain)

Chapter 2. Rough Sets in Insurance Sector
(M.J. Segovia-Vargas and Z. Diaz-Martinez, Department of Financial Economics and Accounting I, Universidad Complutense de Madrid, Spain)

Chapter 3. Prediction of Claims and Selection of Risk Factors in Automobile Insurance using Support Vector Machines, Genetic Algorithms and Classification Trees
(A. Heras-Martinez, C. Bousano-Calzon and P. Tolmos Rodriguez-Pinero, Department of Signal Theory and Communications, Universidad Carlos III, Madrid, Spain, and others)

Chapter 4. Tail Value at Risk. An Analysis with the Normal-Power Approximation
(A. Castaner, M.M. Claramunt and M. Marmol, Department of Economic, Financial and Actuarial Mathematics, Universitat de Barcelona, Spain)
Free Download Available

Chapter 5. Financial Applications of Modal Interval Analysis
(R. Adillon and L. Jorba, Department of Economic, Financial and Actuarial Mathematics, Universitat de Barcelona, Spain)



Additional information