Method and system for high frequency securities trading

The high-frequency securities trading method addresses latency and resource challenges by calculating optimal batch sizes for machine learning models, generating input data, and predicting future events, resulting in improved trading efficiency and accuracy.

JP7656879B2Active Publication Date: 2025-04-04REBELLIONS INC
3 Cites 0 Cited by

Patent Information

Application Number
JP2023027874
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2022-04-07
Filing Date
2023-02-27
Publication Date
2025-04-04
Estimated Expiration
2043-02-27

AI Technical Summary

Technical Problem

High-frequency securities trading using machine learning models faces challenges due to latency issues and the need for significant storage and processing resources, which existing processors often cannot support effectively.

Method used

A high-frequency securities trading method that calculates delay times for securities orders across multiple candidate batch sizes, selects an optimal batch size based on these calculations, generates input data using market data, and uses a machine learning model to predict future events, thereby minimizing latency and improving prediction accuracy.

Benefits of technology

This approach minimizes unnecessary operations in machine learning models, maintains or improves prediction accuracy, and ensures that prediction data is not delayed, allowing for timely and effective securities trading.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

To provide a method, a system, and a program for a highly frequent securities transaction which can minimize or prevent unnecessary operations of a machine learning model while improving or maintaining the accuracy of a predicted price, by determining a batch size for inputting of the machine learning model according to the frequency of stock data being currently received or collected.SOLUTION: A method includes: calculating a delay time to order securities for each of a plurality of candidate batch sizes and selecting a batch size on the basis of the calculated delay times from the candidate batch sizes; generating input data corresponding to the selected batch size by using market data for a target line; generating prediction data for a target line in a point of the future related to the batch size selected on the basis of the generated input data by using a machine learning model; and generating order data for the target line on the basis of the generated prediction data.SELECTED DRAWING: Figure 15
Need to check novelty before this filing date? Find Prior Art