We introduce new general measures of dependence structures that remain invisible when only traditional analysis is employed. This line of research constitutes significant contribution since it opens new routes for measurement of dependence in economic variables [1].
We focus on usefulness of machine learning techniques in economics and finance. At its core, one may perceive machine learning as a general statistical analysis that economists can use to capture complex relationships that are hidden when using simple linear methods. The ability of machine learning techniques to find relationships in data seems well-suited for financial applications. In a paper [1], we evaluate the economic gains of using deep learning methods for the construction of optimal portfolios.
The classical asset pricing frameworks is critical for asset valuation, they are all dominated by the expected utility models. Recently, many researchers find this idealization to be overly restrictive feeling that the expected utility model should be generalized. Collecting more and more data, and witnessing the shift in the behaviour of agents, new theoretical approaches need to be developed.