Intelligent Integrated Systems of Automated Design Laboratory (IISAD)

THE LOGICAL AND PROBABILISTIC RISK MODELS IN BANKING,
BUSINESS AND QUALITY
(in Russian)

Solojentsev E.D., Karassev V.V., Solojentsev V.E.

St-Petersburg, "Nauka", 1999

In this book the general risk task is formulated. It includes the numerical risk evaluation as probability of failure, classification by risk value, analysis of object's risk and risk of set of objects, the assignment of price for risk and risk management.

The theory of logical and probabilistic (LP) simulation, not known by mathematician and economists, were used and developed. The logical addition of initiating events instead of traditional arithmetical addition of certain values is applied. The events can have logical connections AND, OR, NOT.

Risk object is described by large number of signs, every sign has up to 10 gradations. Signs-events are considered as independent, gradations-events forms groups of uncompatible events. Events has clear probabilistic sense.

LP risk models for banks, business, quality and insurance are constructed as associative, on basis of sense, and they are hypotheses of the failure scenario. Failure takes place if there is one, or two, -, or there are all initiating events.

The methods of identification (training) of P risk model by statistical data are offered. As result, the probabilities of signs-events and gradations-events are determined. The method of structural identification or improvement of L risk model for increasing of its precision is offered.

LP risk model has shown twice more accuracy and seven times more stability in classification of credits on "good" and "bad" than known western techniques and software, based on neural networks and methods of trait recognition.

CONTENTS

Introduction

Section 1. THE INTEGRATED TASK OF FAILURE RISK EVALUATION

1. The general features of tasks of failure risk evaluation
2. The table "object - signs" and Data Mining
3. Classification of objects by rank and recognition methods
4. Classification of objects with using of neural networks
5. Logic and probabilistic evaluation of risk in engineering
6. Concept formulation of the integrated task of failure risk

Section 2. THE THEORY OF LP-SIMULATION OF FAILURE RISK

7. The general rules of LP-theory
8. The basic definitions of failure risk LP-evaluation
9. The methodic of the LP-evaluation of failure risk
10. The examples of risk LP-models with connections OR, NOT, AND
11. Classification of objects by the risk and price for the risk
12. The analyses of the failure risk

Section 3. THE IDENTIFICATION OF FAILURE RISK LP-MODELS

13. The serious formulation of training task of failure risk B-model
14. The mathematical formulation of training task of failure risk B-model
15. The algorithm of optimization
16. The method of the casual search
17. The method of slight increments
18. The method of the casual search without account of groups of non-compartible events
19. The illustrations to the identification of risk B-models
20. The training of risk B-model by method of least squares
21. The training of risk B-model by method of maximum probable

Section 4. THE RESEARCHES OF IDENTIFICATION OF RISK LP-MODELS

22. The determination of the rated number of the good objects Ngc
23. The choosing of the average value of risk Pz
24. Way out from impasses
25. The search of global extreme of object function Fmax
26. The connection of accuracy of calculated probabilities and training parameters
27. The control by speed of training
28. The setting of primary approximation for probabilities
29. The training and checking sets
30. The accuracy of object classification
31. About increase of accuracy of object classification
32. The stability of object classification

Section 5. LP-MODELS OF FAILURE RISK IN BANKS

33. Credit risk of individuals
34. Credit risk of companies
35. Analyses of a risk LP-model and a bank credit activity
36. The control by state and development of bank by risk criterion

Section 6. LP-MODELS OF FAILURE RISK IN BUSINESS

37. LP-model and signs of the manager fraud
38. LP-model and signs of the office worker fraud
39. LP-model and signs of the fraud with investment
40. Logical and arithmetical addition of event probabilities
41. LP-models of failure risk of penetration into new market
42. LP-models of a risk of projects, financing of several banks

Section 7. LP-MODELS OF A RISK OF QUALITY LOSS (RQL)

43. The risk of quality loss in quality Japanese conception
44. The example of RQL because of manufacture inaccuracy
45. The example of RQL because of failure of furnish components
46. The example of RQL for product "bridge"
47. The training of the risk LP-model for structure-complex system
48. The construction of risk LP-model for human and machine system

Section 8. SOFTWARE FOR FAILURE RISK LP-MODELS

49. Software for calculation of credit risks and training of risk LP-models
50. Software for building and analyses of risk LP-models
51. Software for orthogonolization L-functions on base of algebra of touples

Conclusion

Literature

13
14
15
16
17
17

20
23
25
26
28
29

31
31
33
34
34
35
35
37
41

42
43
43
45
46
47
48
48
50
51
54

56
58
60
67

71
73
75
77
79
82

84
85
87
88
90
96

99
106
108

111

113

We are looking for:

1) Partners for publishing this book in English. We'll consider any your suggestions;

2) Partners for publishing this book in Russian in large editions;

3) We would like to cooperate with interested people in goals of joint projects, commercial activity, development of new techniques and software for risk evaluation and quality management, publishing and consulting service.

We welcome any your notes, suggestions, comments and wishes.

Please, contact us:

Phone: 7(812)321-47-66 Fax: 7(812)321-47-71
E_mail: risk@sapr.ipme.ru

IISAD Main Page

Tel.: (+7-812)-321-47-66; Fax: (+7-812)-321-47-71;
E-mail: risk@sapr.ipme.ru

Last updated: July, 18, 2005