Construction of Financial Market Transaction Signal Recognition Data Mining Prediction Model Based on Deep Learning

An online communication system supporting stock trading and stock market supervision is constructed. The application of HAR-RV in high frequency signal processing and ACD model based on ultra-high frequency signal are discussed. The corresponding theoretical analysis and parameter estimation methods are also given. Secondly, this paper uses the high frequency data of the Shanghai Composite Index of isochron more than 1 minute to conduct a basic statistical test on the high frequency and high frequency signals of the Chinese stock market. The results show that these signals have a strong “intraday effect’ and a strong “peak and thick tail” characteristic. The HAR-BACD-V model based on ultra-high frequency data is established and the Chinese stock market is taken as the research object to carry out related research. The research conclusion of this paper also confirms that there is heterogeneity in the stock market, and there are differences in the transaction frequency. It is found that the number of micro-factors in the stock market has a significant negative effect on the stock trading time, and empirically tests the micro-structure hypothesis of the stock market. At the same time, the trading volume has a significant positive effect on the return rate and volatility of the stock market, indicating that the message transmission in the Chinese stock market basically follows the continuous information arrival hypothesis.