The strategy for predicting of stock based on the historical data.
Right now, we have finished the lstm/simple ma and exponential ma models to predict the developing trend of all stocks. But the strategies of buying/holding or selling based on these historial data are not constructed. And I aim at constructing a reinforcement larning based model to suggest the action of buying/selling or holding. To obtain this goal, the to-do lists:
- Construct training environment for RL
- Implement RL (AC/PPO)