Futures model training and transaction implementation method based on multi-scale self-attention

Abstract

The invention discloses a futures model training and transaction implementation method based on multi-scale self-attention, and the method comprises the steps: in the futures high-frequency data set construction stage, collecting five types of high-frequency data of a futures main contract, carrying out the preprocessing of the data, and constructing a label through the future price change; in thedeep feature extraction layer training stage, constructing a deep neural network based on multi-scale self-attention, using the constructed label training network, and storing model parameters; in the transaction model training stage, constructing a transaction model, using features output by the deep feature extraction layer, and training the transaction model by using a method of maximizing theSharp ratio; and in the stage of outputting a transaction decision by using the deep feature extraction layer and the transaction model, outputting a transaction action by using the features extracted by the deep feature extraction layer and the trained transaction model. According to the method, the multi-scale characteristics of the financial time sequence and the correlation between differenttime sequences are considered from the perspective of the model, and the prediction accuracy of the futures transaction data is improved.

Description

technical field [0001] The invention relates to a futures model training and transaction realization method based on multi-scale self-attention, and belongs to the technical field of model quantification. Background technique [0002] As an investment methodology, quantitative investment has many advantages such as discipline, system, timeliness and quantification. With the development of computer technology and the advent of the era of big data, more and more people choose to use quantitative investment methods based on computer technology to replace the previous subjective investment methods, and various quantitative investment funds have also been established one after another. In recent years, with the development of machine learning and artificial intelligence technology, more and more people are trying to use such technology to construct intelligent quantitative investment strategies. [0003] At present, there are many quantitative investment research works based on ...

AI Technical Summary

Problems solved by technology

[0004] However, these works have not considered the multi-scale characteristics of financial time series data and the correlation information between different time series data from the model.

Images