Deep Deterministic Policy Gradients for Portfolio management

  • ณฐพงษ์ เมืองไพศาล UTCC
  • สมพร ปั่นโภชา
Keywords: Deep Deterministic Policy Gradients, Machine learning, Portfolio management


The purpose of this study is to adapt Deep Deterministic Policy Gradients with Asset Allocation in SET50 Portfolio. Then measure annual return, standard deviation and Sharpe ratio of portfolio data from 4 January 2010 to 31 December 2020 compare with Equal weight portfolio that we use as benchmark. Base on this study results show Deep Deterministic Policy Gradient portfolio can outperform benchmark. At Learning rate 0.001 Annual return 0.0018 Volatility 0.2363 Sharpe ratio 0.01 compared to Equal weight portfolio Annual return -0.0278 Volatility 0.2289 Sharpe ratio -0.12


Most read articles by the same author(s)

1 2 3 4 5 > >>