Algorithm trading strategy based on GARCH and LSTM models

The prediction of the volatility of stock has been a topic of great interest in the financial realm. In this paper, we aim to create new trading strategies based on the volatility within the financial market and choose GARCH and LSTM models to forecast the volatility separately. We then discuss and set two methods in the research: one is the strategy of buying at high volatility and selling at low volatility, and the other is a comparison strategy on volatility and implied volatility as our two trading strategies, then both of which are compared with a buy-and-hold strategy. Our two volatility strategies are backtested to determine the relative threshold parameters. The findings of the back-test show that both have satisfactory performance in buying and holding, with the comparison strategy on volatility and implied volatility performing extraordinarily.