An algorithm based on LSTM structure has been created to predict the next day value of a particular stock based on past 60 days historic data. Next step is:
1. Application of this algorithm to real life scenarios and check its returns
2. Fine tune the Algorithm for better performance (out of Scope)
Following are the details of the steps required for the next stage of the case study
1. Implement the current application
2. Modify the existing algorithm:
a. Modify the trading algorithm to take the predicted value (previous and predicted) as input parameter for buy/hold/sell
b. Edit the threshold for buy/hold/sell
3. Connect the existing Algorithm to Zerodha’s Kite API for pure algorithmic trading
4. Backtest the algorithm via “paper trading” from Feb 2018 - Feb 2020.
5. Live testing of the architecture.