Font Size: a A A

Research On Intelligent Quantitative Futures Investment Trading Methods

Posted on:2021-02-27Degree:MasterType:Thesis
Country:ChinaCandidate:R Z FengFull Text:PDF
GTID:2428330611953102Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
The securities market has a decisive position in the entire financial market,and how to obtain generous returns in securities market transactions is a topic that financial investment has been discussing.Futures trading,as one of the main forms of trading in the securities market,is a standardized tradable contract based on a public product like oil or a financial asset like bonds.With the continuous growth of futures data,manual analysis is becoming more and more difficult.Due to the rapid development of artificial intelligence technology,the application of machine learning technology to futures trading has become one of the focuses of scholars at home and abroad.First,this paper proposes the integral LSTM algorithm,and uses this algorithm to build a futures price prediction model.The input data of the traditional LSTM algorithm is the futures price data at a single time point,so the hidden layer state value is only the price information at a single time point,and the integral can reflect the total amount of change in a certain period of time.In order to let LSTM The hidden layer state value of the algorithm can pass more price information.In this paper,the integral value of the hidden layer state value of the short-term memory of the LSTM model is used to replace the previous hidden layer state value,so that the hidden layer state value is transmitted in a certain time domain.Information on the futures price increase,thus improving the accuracy of the integral LSTM algorithm in predicting futures prices,and experiments have shown that the integral LSTM algorithm is superior to the original LSTM algorithm in terms of futures price prediction accuracy.Secondly,this article combines the integral LSTM network in deep learning and the Q-learning algorithm in reinforcement learning into a deep Q-Learning algorithm,and uses this algorithm to build an automated futures trading model.The model uses the integral LSTM network as a trading agent.The agent accepts the historical trading information of the futures and analyzes the trading information to generate a trading action.The model trading actions all simulate the trading behavior of real futures investors.There are total selling,buying and holding.After the execution of the trading action,the corresponding revenue will be generated,and the revenue value is used as the reward value of the model to update the network parameters of the model.Finally,the test data was used to test the model's trading ability.The experiment proved that using the deep Q-Learning algorithm for automatic futures trading can obtain good returns.All in all,this article combines the prediction ability of deep learning and the decision-making ability of reinforcement learning to apply to the futures trading market,and has practiced the construction of automated investment models,the improvement of the efficiency of futures trading and the application of artificial intelligence technology in the financial field.
Keywords/Search Tags:deep learning, deep reinforcement learning, price prediction, Q-learning, quantitative trading
PDF Full Text Request
Related items