A1598
Title: Predicting hedge funds returns
Authors: Christos Argyropoulos - University of Essex (United Kingdom) [presenting]
Abstract: Profitability gains are evaluated when investors select hedge funds according to machine learning methods' returns forecasts. Three main techniques are used to forecast individual hedge returns: shrinkage, dimensionality reduction, and artificial neural network, based on an extensive set of variables. An extended set of predictors is used to calculate the forecasts, including hedge fund characteristics, risk factors and other macroeconomic variables. The accuracy of the forecasts is assessed via an out-of-sample asset allocation exercise.
