China
Africa
Explore chapters and articles related to this topic
Optimization algorithms for multiple-asset portfolios with machine learning techniques
Published in Noura Metawa, M. Kabir Hassan, Saad Metawa, Artificial Intelligence and Big Data for Financial Risk Management, 2023
In this chapter, we characterize the trading risk for emerging equity markets by using a multivariate liquidity-adjusted value at risk (L-VaR) technique and optimization algorithms that focus on the modeling of optimum L-VaR under the notion of illiquid and adverse market conditions and by exercising different dependence measures (correlation factors) and liquidity closeout periods.1 The overall aim of this chapter is to construct different structured equity portfolios, which includes stock markets indices of the Gulf Cooperation Council (GCC) region, and to evaluate the risk characteristics of those portfolios besides examining a robust iterative optimization algorithmic process for computing efficient and coherent2 economic capital.3 To that end, we apply a general trading risk modeling algorithm that accounts for the characteristics of the series of equity price returns—for example, fat tails (leptokurtosis), skewness, correlation parameters, and closeout horizons—and effectively forecasts market risk within a short time horizon. As such, the robust modeling technique and non-linear quadratic optimization algorithms implemented in this chapter are based on the Al Janabi model (Al Janabi, 2008; Madoroba and Kruger, 2014).
Modeling with Probability
Published in William P. Fox, Robert E. Burks, Modeling Change and Uncertainty, 2022
William P. Fox, Robert E. Burks
In real-world scenarios, the assumption of a constant rate (or probability per unit time) is rarely satisfied. For example, the rate of incoming phone calls differs according to the time of day. But if we focus on a time interval during which the rate is roughly constant, such as from 2 to 4 p.m. during workdays, the exponential distribution can be used as a good approximate model for the time until the next phone call arrives. Similar caveats apply to the following examples, which yield approximately exponentially distributed variables:The time until a radioactive particle decays, or the time between clicks of a Geiger counter.The time it takes before your next telephone call.The time until default (on payment to company debt holders) in reduced form credit risk modeling.
A Case Study of Enterprise Machine Learning Frame Work for Investment Platforms
Published in Durgesh Kumar Mishra, Nilanjan Dey, Bharat Singh Deora, Amit Joshi, ICT for Competitive Strategies, 2020
Rao Casturi, Rajshekhar Sunderraman
The area of credit risk modeling [13] [1] became important due to the cyclical nancial instabilities [9]. The nancial institutions want to predict the following loses in their portfolios or to the investments they did. Since the introduction of the Basel II Capital Accord (Basel Committee on Banking Supervision, 2004) over a decade ago, qualifying nancial institutions have been able to derive their own internal credit risk models under the advanced internal ratingsbased approach (A-IRB) without relying on regulator’s xed estimates. Under the IBR the banks or nancial institutions should follow the four key parameters. Probability of Default (PD), Loss Given Default (LGD), Exposure At Default (EAD), Maturity (M) PD: The likelihood that a loan will not be repaid and will therefore fall into default in the next 12 months. LGD: The estimated economic loss, express as a percentage of exposure, which will be incurred if an obligor goes into default. we can say LGD equals: 1 – recovery rate. in other words ()LGD = (1 −RecoveryRate)
Concept Drift Monitoring and Diagnostics of Supervised Learning Models via Score Vectors
Published in Technometrics, 2023
Kungang Zhang, Anh T. Bui, Daniel W. Apley
In order to demonstrate the performance of our score-based concept drift approach and illustrate its usage, we now present a real data example, which involves credit risk modeling with data collected over the period 2003–2008 from a major financial company, during which time the subprime mortgage crisis happened. In Section B.1, supplementary materials we provide a simulation example where concept drift results in no change in expected error rate (and thus, error-based methods cannot detect it), but our score-based approach is able to effectively detect it. In Section B.2, supplementary materials, we provide more comprehensive simulation results comparing the drift detection performances of our score-based approach and alternative existing methods. The main conclusion is that our score-based approach achieves better performance across every example that we considered, often substantially so. In Section B.4, supplementary materials, we also provide another real data example, the Capital Bikeshare rental data from 2010 to 2020 during which time the “sharing economy” steadily expanded, to demonstrate retrospective and prospective analysis for a regression problem. In Section C, supplementary materials, we use simulation examples to investigate the performance of the Fisher decoupling approach in diagnosing which parameters have changed and how they have changed.
Optimal Model for Electricity Retailer Considering Demand Response and Risk Management through Stochastic Formulation
Published in Electric Power Components and Systems, 2022
Kourosh Apornak, Soodabeh Soleymani, Faramarz Faghihi, Babak Mozafari
The hypotheses of the present study can be categorized according to the following:Determining significant relationship between CVaR risk modeling and Retailer’s Expected profitInvestigating the Impact of DR Programs on Retailer ProfitsAnalysis of price and power changes due to price elasticity of different types of customersStudy for the impact of profit standard deviation on the profit of electricity retailer in possible values of the confidence level (α)Examining the effect of confidence level (α) in risk retailer’s profit maximization