Analysis and processing method for quantitative transaction based on neural network

Abstract

The invention discloses a method for analyzing and processing a quantitative transaction based on a neural network in the field of quantitative transactions. The method mainly comprises the followingsteps: providing a data set A for a data model basis and training the data set A in a corresponding neural network data model; based on parallel training of a plurality of computing nodes, training ina neural network data model, and sequentially determining a feature data set B, a feature data set C, an algorithm data set D, an application strategy data set E, result data F, an efficiency data set G and an application strategy data set; and returning to the step S10 for training, and continuously optimizing strategies and parameters. Data features and feature recognition modes are fine in logic, more objective and sound; the selection and configuration of algorithm indexes are timely and diverse, and the problem of indication passivation is solved; strategy selection and configuration aretimely and perfect and adapt to the continuous change of trends; parameters are flexible and accurate and adapt to continuously changing data environments; execution results are summarized in time and the strategies are corrected so as to adapt to continuously changing market environments.

Description

technical field [0001] The invention relates to the technical field of quantitative trading, in particular to an analysis and processing method of quantitative trading based on a neural network. Background technique [0002] Quantitative trading refers to the use of advanced mathematical models to replace human subjective judgments, and the use of computer technology to select a variety of "high probability" events that can bring excess returns from huge historical data to formulate strategies, which greatly reduces investor sentiment. Avoid irrational investment decisions when the market is extremely frenzied or pessimistic. [0003] Quantitative trading is based on a large amount of data and algorithmic models, through analysis and backtesting to find out the appropriate trading strategy and position configuration. The following main problems often existed in conventional quantitative transactions in the past: [0004] (1) Due to the singleness of the identification meth...

AI Technical Summary

Problems solved by technology

[0004] (1) Due to the singleness of the identification method of data characteristics, the grasp of objective data characteristics is not accurate;

[0005] (2) Due to the singleness of the algorithm, the matching degree is prone to be unstable in real trading;

[0006] (3) Since there is no strategy learning mechanism, it is easy to passivate the indicators;

[0007] (4) Due to the inflexibility of parameters, outdated parameter values ​​cannot be found in time during the firm offer process

Images