Novel Parametric and Non-Parametric Cross-Entropy Models and their Applications in Portfolio Analysis
DOI:
https://doi.org/10.26713/cma.v16i3.3366Keywords:
Divergence Measures, Cross Entropy Models, Portfolio optimization, Risk Measurement, Financial Decision MakingAbstract
For real-world applications in the mathematical sciences, it is important to create adaptable models in probability spaces. Models that are too rigid or too limited typically don't show how unpredictable and complicated these systems really are. To solve this problem, we need families of random models that can make the analytical framework more flexible and strong. This work presents a collection of novel parametric and non-parametric discrete cross-entropy models aimed at improving flexibility while maintaining mathematical precision. The fundamental purpose of these models is to create a framework for creative optimization techniques that may be used across varied situations. The suggested cross-entropy models build on standard probability-based methods by adding new distance metrics that make it easier to judge uncertainty and unpredictability. These models are important in theory and also have evident real-world uses. We specifically concentrate on their use in portfolio analysis, emphasizing risk assessment and the best allocation of assets. By using the created models, we illustrate how investors and decision-makers may more correctly reflect fluctuations in risk and return, leading to better-informed strategies for investing under uncertainty. The research shows that you may use either parametric or non-parametric versions of the models, depending on what data you have and what assumptions you make about the situation. When distributional assumptions are true, parametric forms provide you structured and efficient estimates. When these assumptions are weak or don't exist, non-parametric forms give you more freedom. Together, they provide a complete toolset for addressing uncertainty in complex systems. This paper advances the development of cross-entropy-based models inside probability spaces and demonstrates their efficacy in financial decision-making, especially regarding portfolio risk assessment and optimization
Downloads
References
H. Ahmadzade, R. Gao, H. Naderi, M. Farahikia, Partial divergence measure of uncertain random variables and its application, Soft Computing 24 (2019), 501–512, DOI: 10.1007/s00500-019-03929-0
G.A. Anastassiou, Intelligent Analysis: Fractional Inequalities and Approximations Expanded, Springer, 2021.
A.K. Bera, S.Y. Park, Optimal portfolio diversification using the maximum entropy principle, Econometric Reviews 27(4–6) (2008), 484–512, DOI: 10.1080/07474930801960394
G. Bugár, M. Uzsoki, Portfolio optimization strategies: Performance evaluation with respect to different objectives, Journal of Transnational Management 16(3) (2011), 135–148, DOI: 10.1080/15475778.2011.596773.
F. Camaglia, I. Nemenman, T. Mora, A.M. Walczak, Bayesian estimation of the Kullback-Leibler divergence for categorical systems using mixtures of Dirichlet priors, Physical Review E 109 (2024), 024305, DOI: 10.1103/PhysRevE.109.024305.
Cichocki, S.I. Amari, Families of alpha-, beta- and gamma-divergences: Flexible and robust measures of similarities, Entropy 12(6) (2010), 1532–1568, DOI: 10.3390/e12061532.
J.N. Kapur, Measures of Information and Their Applications, New Age International, 1994.
M. Khalaj, R. Tavakkoli-Moghaddam, F. Khalaj, A. Siadat, New definition of the cross entropy based on the Dempster-Shafer theory and its application in a decision-making process, Communications in Statistics – Theory and Methods 49(4) (2019), 909–923, DOI: 10.1080/03610926.2018.1554123.
S. Kullback, R.A. Leibler, On information and sufficiency, The Annals of Mathematical Statistics 22(1) (1951), 79–86, DOI: 10.1214/aoms/1177729694
H. Markowitz, Portfolio selection, The Journal of Finance 7(1) (1952), 77–91, DOI: 10.1111/j.1540-6261.1952.tb01525.x.
D. Nocetti, Markowitz meets Kahneman: Portfolio selection under divided attention, Finance Research Letters 3(2) (2006), 106–113, DOI: 10.1016/j.frl.2006.03.006.
J. Ou, Theory of portfolio and risk based on incremental entropy, The Journal of Risk Finance 6(1) (2005), 31–39, DOI: 10.1108/15265940510574754.
O. Parkash, P. Kakkar, New information theoretic models, their detailed properties and new inequalities, Canadian Journal of Pure and Applied Sciences 8(3) (2014), 3115–3123.
M. Sarangal, O. Parkash, Weighted parametric divergence models for discrete probability distributions, Gulf Journal of Mathematics 13(2) (2022), 87–93, DOI: 10.56947/gjom.v13i2.721.
I.J. Taneja, On unified (R,S)-information measures, Journal of Information and Optimization Sciences 16(1) (1995), 127–176, DOI: 10.1080/02522667.1995.10699212
V. Torra, Y. Narukawa, M. Sugeno, On the f-divergence for discrete non-additive measures, Information Sciences 512 (2020), 50–63, DOI: 10.1016/j.ins.2019.09.033.
W. Wang, J. Yu, T. Xu, C. Zhao, X. Zhou, On-line abnormal detection of nuclear power plant sensors based on Kullback-Leibler divergence and ConvLSTM, Nuclear Engineering and Design 428 (2024), 113489, DOI: 10.1016/j.nucengdes.2024.113489.
Y. Zhang, J. Pan, K. Li, Z. Chen, X. Liu, J. Wang, On the properties of Kullback-Leibler divergence between multivariate Gaussian distributions, Proceedings of the 37th International Conference on Neural Information Processing Systems, Vol. 2535 (2024), 58152–58165.
Downloads
Published
How to Cite
Issue
Section
License
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a CCAL that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
