Как я могу спроектировать искусственный интеллект для прогнозирования запасов с различными входными переменными?

---

I am new in AI tech and currently I am interested in the stock prediction using AI . I saw some of the example which is using LSTM for stock prediction (input = historical prices in certain time period e.g. price at t=-60 to t=0), output = price at t=1). However, it is just using the history price pattern to predict the future price. I believed using the historical price itself would not be enough to predict the stock price since the stock price move depend on the human reaction on different variables ( such as the trading volume, the stock market index, the RSI …etc) . So what method or algorithm should be used in order to to let the machine predict the stock price based on the multiple historical data (price, volume, index, RSI…etc) . I am thinking if there is anyway I can put a series of data (price, volume, index, RSI…etc) at each of the time slot) from t=-60 to t=0, and use the weighting on all this historical variables to predict the stock instead of just the historical price

Что я уже пробовал:

ЛСТМ.............................................

---