Quantitative Transaction Research Based On Attention-LSTM-OPTRSRS Model

With the rapid development of artificial intelligence and data mining,the quantitative investment generated by the cross research of finance,statistical theory and computer technology has attracted much attention from the global financial community.In essence,quantitative investment is a new investment method based on a series of statistical theories,through computer analysis,to deeply mine the information behind the data,and obtain the trading stocks and trading time.Therefore,this paper presents a quantitative trading research that is based on the Attention-LSTM-OPTRSRS model.Nine factors with different styles are selected as input variables for the model.The Attention model is introduced to allocate the weight of each factor when inputting the model.The LSTM model is used to predict all constituent stocks of the CSI 300 index,and the five stocks with the highest cumulative earnings are selected.The rationality of stock selection is proved through various performance indicators including Mean Absolute Error,Mean Squared Error,Mean Squared Logarithmic Error,Mean Absolute Percentage Error,Maximum Withdrawal Rate,Sharp Ratio,and Volatility.For the selected stocks,this paper gradually calculates the slope index of the day,performs standardization optimization,box-cox optimization,price-volume optimization,box-cox stop loss optimization,and price-volume stop loss optimization,and obtains the Optimized Resistance Support Relative Strength.The advantages of the Optimized Resistance Support Relative Strength over traditional strategies such as the Double Moving Average strategy and Moving Average Convergence Divergence strategy are demonstrated through comparison.In the quantitative stock selection part,it is recommended to use the AttentionLSTM model with nine factor inputs;In the quantitative timing part,for different investors,different suggestions are put forward according to the investor’s risk preference: For those with a high risk appetite,box-cox optimization strategy is recommended,and for risk averse,it is recommended to adopt box-cox optimization stop loss strategy.