Abstract:Recently defined expectile regions capture the idea of centrality with respect to a multivariate distribution, but fail to describe the tail behavior while it is not at all clear what should be understood by a tail of a multivariate distribution. Therefore, cone expectile sets are introduced which take into account a vector preorder for the multi-dimensional data points. This provides a way of describing and clustering a multivariate distribution/data cloud with respect to an order relation. Fundamental properties of cone expectiles including dual representations of both expectile regions and cone expectile sets are established. It is shown that set-valued sublinear risk measures can be constructed from cone expectile sets in the same way as in the univariate case. Inverse functions of cone expectiles are defined which should be considered as rank functions rather than depth functions. Finally, expectile orders for random vectors are introduced and characterized via expectile rank functions.
What problem does this paper attempt to address?
### Problems the paper attempts to solve
This paper attempts to solve the problems of centrality and tail - behavior description in multivariate data distributions. Specifically:
1. **Centrality description**: Existing multivariate expectile regions can capture the centrality of multivariate distributions, but cannot effectively describe tail behavior. This is because for multivariate distributions, what is "tail" is not clear.
2. **Tail - behavior description**: To overcome this problem, the paper introduces cone expectile sets, which describe and cluster multivariate distributions / data clouds by considering the vector preorder of multi - dimensional data points. This provides a method for describing multivariate distributions according to the order relationship.
3. **Risk measure construction**: The paper shows how to construct set - valued sublinear risk measures from cone expectile sets, similar to the method in the univariate case.
4. **Inverse function definition**: The inverse functions of cone expectiles are defined, and these inverse functions should be regarded as rank functions rather than depth functions.
5. **Random vector ordering**: The expectile orders of random vectors are introduced and characterized by expectile rank functions.
### Key concepts and methods
- **Cone Expectile Sets**: Cone expectile sets are introduced by considering the vector preorder of multi - dimensional data points. These sets can theoretically describe the data clouds of multivariate distributions and have good mathematical properties, such as dual representations and convexity.
- **Dual Representations**: The dual representations of cone expectile sets and expectile regions are established. These representations not only provide theoretical support but also methods for practical calculations.
- **Set - Valued Risk Measures**: Set - valued sublinear risk measures are constructed from cone expectile sets, and these measures are of great significance in financial risk management.
- **Rank Functions**: The inverse functions of cone expectiles are defined, and these functions can be used for the ordering and classification of data points.
### Application areas
- **Financial risk management**: The risk measures constructed through cone expectile sets can help financial analysts better understand and manage multivariate financial risks.
- **Multi - Criteria Decision Making**: Cone expectile sets and rank functions can be applied to multi - criteria decision making, helping decision - makers find the optimal solution in multi - dimensional data.
- **Data analysis**: These methods can be used to describe and analyze the centrality and tail - behavior of multivariate data, especially when dealing with complex data sets.
### Summary
By introducing cone expectile sets and related concepts, the paper solves the problems of centrality and tail - behavior description in multivariate data distributions, and provides new tools and methods for financial risk management, multi - criteria decision making and other fields.