Using the same toy example, we compared several linear domain adaptation methods, listed in Table 1. The toy example is designed in such a way that the projections on the first principal component do not separate well the classes neither in the source nor in the target domain. In addition,...
Keywords: anomaly detection; financial time series; principal component analysis; neural network; missing data; market risk; value at risk1. Introduction In the context of financial risk management, financial risk models are of utmost importance in order to quantify and manage financial risk. Their ...