This repository contains codes and part of samples in my PhD thesis Multiple Testing Procedures on the Mean-Variance Efficiency of Factor Pricing Models, including three methods named AdaFAT, SOAP and listNN.
In the r folder, the adaptive factor-adjusted multiple testing procedure (AdaFAT) and the shrinkage operator for alternatives proportion (SOAP) are written in R codes. A simulation is given here as an example and model parameters are calibrated by Chinese A-share market from Jan. 2014 to Dec. 2018.
In the torch folder, the feedforward neural network for constructing micro portfolios (listNN) is implemented in PyTorch. Empirical data in the torch/sample folder has been masked and is used to display the format of inputs.