CDS Rate Construction Methods by Machine Learning Techniques

To price and risk-manage OTC derivatives, banks should estimate counterparty default risks based on liquidly quoted CDS rates,which aren't available for the vast majority of counterparties.Thus,banks construct proxy CDS rates... [ view full abstract ]

To price and risk-manage OTC derivatives, banks should estimate counterparty default risks based on liquidly quoted CDS rates,which aren't available for the vast majority of counterparties.Thus,banks construct proxy CDS rates using methods ignoring counterparty-specific default risks.Our CDS Proxy methods using Machine Learning Techniques achieve high accuracy based on tests from 156 classifiers out of 8 most popular classifier families including feature correlations on classification. It is a first systematic study of CDS Proxy Construction using Machine Learning Techniques,first systematic classifier comparison study based exclusively on financial market data and can extend for the construction of proxies for other financial variables.