Artificial Neural Network Approach to Counterparty Credit Risk and XVA
Artificial Neural Network Approach to Counterparty Credit Risk and XVA
复制标题
交易对手信用风险和 XVA 的人工神经网络方法
DOI：
10.2139/ssrn.3312944
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Sven Welack
中科院分区：
文献类型：
--
作者：
Sven Welack
采用人工神经网络的新方法，以提高现有的Monte Carlo风险引擎的基础设施。利用人工神经网络从现有的利率互换预期风险配置文件中检索贸易和市场数据，使其能够用作数据控制框架和风险解释应用程序的一部分。还利用人工神经网络来预测预期风险，模仿Monte Carlo风险引擎，在更快的执行速度下显示出类似的准确性。
Novel approaches employing an Artificial Neural Networks to enhance the infrastructure of existing Monte Carlo Risk engines are presented. An Artificial Neural Network is utilized to retrieve trade- and market data from existing Expected Exposure profiles of interest rate swaps which enables its usage as part of data control frameworks and exposure explain applications. An Artificial Neural Network is also utilized to predict Expected Exposure mimicking a Monte Carlo Risk engine showing similar accuracy at faster speeds of execution.