Can Representation Bias Help the Returns Forecast and Portfolio Selection?

The purpose of this work is to figure out whether the representation bias can help the returns forecasting and portfolio selection. Based on the representation bias theory, first of all, a new stock hierarchy system is constructed, and the influence of the criteria on the stocks is measured by adapted fuzzy AHP. Then a new weighting and computing method to the horizontal and vertical representation returns is proposed and tested. The results show that the representation bias information can be useful to the returns forecast as well as the portfolio selection. Finally, the horizontal and vertical representation returns are combined as to construct a comprehensive representation return, and both its advantages and disadvantages are discussed in some way.