Adaptive Trading System Based on LSTM Neural Network

【Author】 Wang, Yue; Wang, Shuyue; Tang, Nan; Kumar, Priyan Malarvizhi; Hsu, Ching-Hsien

【Source】ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING

【影响因子】2.807

【Abstract】The growing difficulty and competitive assets in the capital market are a significant obstacle to the finance sector that cannot attain excellent results under all economic conditions through a rigid trading system (TS) built by established financial experts. Increasing shareholders make a conscious effort to develop a structured approach to stock valuation and predictive analytics that is denoted as a trading system. Since the implementation of online trading, the total amount of day-to-day purchases on the trading platform has been substantially increased, with the subsequent term of market volume and profitability. To overcome these issues, in this paper, a trading system based on Long Short-Term Memory neural network (TS-LSTM) has been proposed to increase the implementation of online trading and the total amount of day-to-day purchases on the TS platform. Long Short-Term Memory efficiently analyzes multiple financial information to open, closing many trading activities in a minimal time taken to improve the effective management of the TS. Long Short-Term Memory is used to obtain insightful economic characteristics that distinguish intrinsic attributes of the real economy for an adaptive trading system based on the time-series existence of the financial sector. A neural network controls the evolutionary resolution based on the prediction for the share price development, and the estimation of stock market performance has been further enhanced. The experimental results show that the suggested system has been validated by Kaggle bitcoin datasets and improves the accuracy ratio of 97.22% to predict stock valuation.

【Keywords】Trading system (TS); Long Short-Term Memory (LSTM); Trading activities; Stock market; Share price

【发表时间】

【收录时间】2022-01-01

【文献类型】

【主题类别】

--