ABSTRACT
A prominent point of criticism faced by Machine Learning tools is their inability to uncover causality relationships between features and labels because they are mostly focused to capture correlations. Market conditions are known to be time-varying and the relationships between firm characteristics and returns also change from year to year. One solution to this issue may simply be to embrace non-stationarity. Traditional machine learning models aim to uncover relationships between variables but do not usually specify directions for these relationships. One typical example is the linear regression. The most common assumption in machine learning contributions is that the samples that are studied are i.i.d. realizations of a phenomenon that we are trying to characterize. Stationarity is a key notion in financial econometrics: it is much easier to characterize a phenomenon with distributional properties that remain the same through time. This chapter concludes the topic of causality by mentioning a particular type of structural models: structural time series.
