Search

KR-102964600-B1 - The Method And System For Providing Personalized Auto-Trading Service Based On Conversational Interface

KR102964600B1KR 102964600 B1KR102964600 B1KR 102964600B1KR-102964600-B1

Abstract

The present invention relates to a method and system for providing a conversational interface-based personalized auto-trading service, wherein the method and system automatically generate a process module including verification of trading conditions and generation of trading commands from natural language text input by a user using an LLM, and load the generated process module to control the execution of automatic trading on actual financial assets.

Inventors

  • 황진희

Dates

Publication Date
20260512
Application Date
20251201

Claims (15)

  1. A method for providing a personalized auto-trading service performed by a computing system comprising one or more processors and one or more memories, wherein The above computing system includes a process generation module; and a process execution module; and The method of providing the above-mentioned personalized auto-trading service is, A process module generation step for generating a personalized process module comprising: a verification process that verifies pre-set trading conditions and a transaction process that generates a transaction order to automatically sell or buy financial assets according to the result of the verification process, wherein the verification process analyzes input information in the form of natural language text related to an investment strategy for financial assets received from a user using an internal or external LLM by means of a process generation module; and A process execution step comprising loading a process module generated by a process execution module to perform the verification process and performing the transaction process corresponding to the result of the verification process; The above process module is implemented as a LangGraph structure connected by multiple nodes and edges, wherein the verification process and the transaction process are each divided into one or more detailed process nodes, and The above process execution step is, A method for providing a personalized auto-trading service, comprising: a graph execution step that performs non-sequential flow control by means of a process execution module, wherein, depending on the execution result or state of an individual detailed process node on the Lang graph structure, proceeds to a node in a preset next order, returns to a node in a previous step, or skips a specific node.
  2. In claim 1, The above computing system further includes a conversational interface providing unit that provides a conversational interface for interaction with a user to a user terminal, and The above input information is obtained through the user's chat input via the above conversational interface, and The method of providing the above-mentioned personalized auto-trading service is, A method for providing a personalized auto-trading service, further comprising: a conversational interface providing step in which, by means of a conversational interface providing unit, a response to the user's input information is generated through the LLM and output to the conversational interface, and the structure of the process module generated in the process module generation step is visualized and provided to the user.
  3. In claim 1, The above process module creation step is, An information derivation step of deriving one or more confirmation information and one or more transaction information corresponding to the one or more confirmation information using the LLM based on input information including natural language text entered by a user; A verification process derivation step for deriving each verification process including code that can be executed in the computing system based on the above one or more verification information; and A method for providing a personalized auto-trading service, comprising: a transaction process derivation step of deriving each transaction process including code that can be executed in the computing system based on the above one or more transaction information.
  4. delete
  5. In claim 1, The above process module creation step is, A method for providing a personalized auto-trading service, further comprising: an editing interface provided to a user terminal that converts and displays one or more detailed processes included in each of the above-mentioned verification process and the above-mentioned trading process into a visual object in the form of a flowchart, and an editing step that modifies internal attribute information of the one or more detailed processes according to user input.
  6. In claim 1, The above verification process is, A detailed collection process for collecting financial data from one or more external data servers; and A method for providing a personalized auto-trading service, comprising: a detailed analysis process for analyzing whether analyzed financial data corresponds to the above trading conditions.
  7. In claim 6, The above verification process is, A method for providing a personalized auto-trading service, further comprising: a decision process that inputs a decision prompt including the results of the above-mentioned collection detailed process and the above-mentioned analysis detailed process into the above-mentioned LLM, and determines whether to execute a trade based on the inference result of the above-mentioned LLM.
  8. A method for providing a personalized auto-trading service performed by a computing system comprising one or more processors and one or more memories, wherein The above computing system includes a process generation module; and a process execution module; and The method of providing the above-mentioned personalized auto-trading service is, A process module generation step for generating a personalized process module comprising: a verification process that verifies pre-set trading conditions and a transaction process that generates a transaction order to automatically sell or buy financial assets according to the result of the verification process, wherein the verification process analyzes input information in the form of natural language text related to an investment strategy for financial assets received from a user using an internal or external LLM by means of a process generation module; and A process execution step comprising loading a process module generated by a process execution module to perform the verification process and performing the transaction process corresponding to the result of the verification process; The above transaction process is, Determine the type of financial asset to be sold or bought, and the trading volume (Quantity) of the financial asset to be sold or bought, and The method of providing the above-mentioned personalized auto-trading service is, A method for providing a personalized auto-trading service, further comprising: a trading volume determination step in which, by means of a process execution module, the trading volume is dynamically calculated according to the amount of the user's financial assets or other financial assets owned when the trading process is executed.
  9. A method for providing a personalized auto-trading service performed by a computing system comprising one or more processors and one or more memories, wherein The above computing system includes a process generation module; and a process execution module; and The method of providing the above-mentioned personalized auto-trading service is, A process module generation step for generating a personalized process module comprising: a verification process that verifies pre-set trading conditions and a transaction process that generates a transaction order to automatically sell or buy financial assets according to the result of the verification process, wherein the verification process analyzes input information in the form of natural language text related to an investment strategy for financial assets received from a user using an internal or external LLM by means of a process generation module; and A process execution step comprising loading a process module generated by a process execution module to perform the verification process and performing the transaction process corresponding to the result of the verification process; The above computing system further includes a verification module that checks the stability of the generated personalized process module, and A method for providing a personalized auto-trading service, further comprising a pre-verification step in which, by means of a verification module, structural verification confirming whether the process module is grammatically executable, unit verification testing the operation of each detailed process, and risk verification confirming whether the input value exceeds a trading limit or a loss limit.
  10. A method for providing a personalized auto-trading service performed by a computing system comprising one or more processors and one or more memories, wherein The above computing system includes a process generation module; and a process execution module; and The method of providing the above-mentioned personalized auto-trading service is, A process module generation step for generating a personalized process module comprising: a verification process that verifies pre-set trading conditions and a transaction process that generates a transaction order to automatically sell or buy financial assets according to the result of the verification process, wherein the verification process analyzes input information in the form of natural language text related to an investment strategy for financial assets received from a user using an internal or external LLM by means of a process generation module; and A process execution step comprising loading a process module generated by a process execution module to perform the verification process and performing the transaction process corresponding to the result of the verification process; The above computing system further includes a sandbox module that provides a virtualized simulated investment environment, and The method of providing the above-mentioned personalized auto-trading service is, A method for providing a personalized auto-trading service, further comprising: a simulation step in which the verification process receives actual financial data in real time and verifies the trading conditions through a sandbox module, wherein the trading by the trading process is simulated in the simulated investment environment.
  11. A method for providing a personalized auto-trading service performed by a computing system comprising one or more processors and one or more memories, wherein The above computing system includes a process generation module; and a process execution module; and The method of providing the above-mentioned personalized auto-trading service is, A process module generation step for generating a personalized process module comprising: a verification process that verifies pre-set trading conditions and a transaction process that generates a transaction order to automatically sell or buy financial assets according to the result of the verification process, wherein the verification process analyzes input information in the form of natural language text related to an investment strategy for financial assets received from a user using an internal or external LLM by means of a process generation module; and A process execution step comprising loading a process module generated by a process execution module to perform the verification process and performing the transaction process corresponding to the result of the verification process; The above transaction command is converted into API call signals of different specifications depending on the type of the financial asset and generated, and The method of providing the above-mentioned personalized auto-trading service is, A method for providing a personalized auto-trading service, further comprising: a heterogeneous asset trading step in which, by means of a process execution module, if the financial asset is a stock, a stock trading command conforming to the API specifications of an internal or external securities firm trading server is transmitted, and if the financial asset is a virtual currency, a virtual currency trading command conforming to the API specifications of an internal or external virtual currency exchange server is transmitted.
  12. In claim 11, The above computing system is, It further includes a secure storage unit that encrypts and stores access rights and transaction authentication keys for individual users' securities accounts or virtual currency wallets; and The above heterogeneous asset transaction step is, A method for providing a personalized auto-trading service, comprising: a server-side trading step in which, by means of a process execution module, the transaction authentication key of the user is called from the security storage unit when the transaction process is executed, and electronic signature and authentication for the transaction command are performed on the system server without direct intervention of the user terminal to execute a sell or buy.
  13. A computing system comprising one or more processors and one or more memories, and performing a method of providing a personalized auto-trading service, A process generation module that generates a personalized process module comprising a confirmation process that analyzes input information in the form of natural language text related to an investment strategy for financial assets received from a user using an internal or external LLM, confirms pre-set trading conditions, and generates a trading order that automatically sells or buys financial assets according to the result of the confirmation process; and A process execution module that loads a generated process module to perform the verification process and performs the transaction process in accordance with the result of the verification process; The above process module is implemented as a LangGraph structure connected by multiple nodes and edges, wherein the verification process and the transaction process are each divided into one or more detailed process nodes, and The above process execution module is, A computing system that performs a graph execution step, which performs non-sequential flow control in which, depending on the execution result or state of an individual detailed process node in the above Lang graph structure, proceed to a node in a preset next order, return to a node in a previous step, or skip a specific node.
  14. delete
  15. In claim 13, The above process generation module is, A computing system that provides an editing interface to a user terminal that converts and displays one or more detailed processes included in each of the above verification process and the above transaction process into a visual object in the form of a flowchart, and performs an editing step that modifies internal attribute information of the one or more detailed processes according to user input.

Description

The Method and System for Providing Personalized Auto-Trading Service Based on Conversational Interface The present invention relates to a method and system for providing a conversational interface-based personalized auto-trading service, wherein the method and system automatically generate a process module including verification of trading conditions and generation of trading commands from natural language text input by a user using an LLM, and load the generated process module to control the execution of automatic trading on actual financial assets. With the recent advancement of fintech and the widespread adoption of smartphones leading to a surge in individual investors' participation in the financial market, there is a growing demand for algorithmic trading based on systematic investment principles, moving beyond simple trading. In the past, stock trading was a manual process where users personally checked market prices and made trading decisions; however, this approach had limitations, such as the difficulty of responding in real-time to rapidly changing market conditions and the inability to execute consistent strategies due to psychological factors. To overcome this, automated trading programs have emerged that automatically execute orders when specific conditions are met; however, these programs present a very high barrier to entry for general individual investors, as they require proficiency in complex programming languages or intricate HTS formula managers. Conventional auto-trading systems include a stock trading platform server that supports an automatic trading bot using artificial intelligence and big data, as described in Korean Registered Patent No. 10-2453549. The stock trading platform server that supports an automatic trading bot using artificial intelligence and big data and the method thereof disclose a technology that collects news, social media data, chart information, etc., predicts fluctuations in stock prices using an artificial neural network, and automatically executes trading based on the results. However, while the above-mentioned conventional auto-trading system may be useful for predicting market direction through big data analysis, it has limitations in strategizing and reflecting the user's specific and individual investment intentions into the system. In other words, conventional auto-trading systems mainly take the form of a black box that relies entirely on prediction results produced by AI, and thus have the disadvantage that it is difficult for the user to understand the operating principles of the strategy or to control it finely according to their own intentions. Therefore, there is a need to develop a new type of auto-trading system that allows even users without programming knowledge to easily generate their own investment strategies using natural language. FIG. 1 schematically illustrates the execution steps of a method for providing an interactive interface-based personalized auto-trading service according to an embodiment of the present invention and the components of a computing system. FIG. 2 schematically illustrates detailed components of a computing system that performs a method for providing an interactive interface-based personalized auto-trading service according to an embodiment of the present invention. FIG. 3 illustrates an exemplary interactive interface provided to a user terminal by an interactive interface providing unit according to an embodiment of the present invention. FIG. 4 schematically illustrates a process of deriving one or more confirmation information and one or more transaction information by an internal or external LLM according to an embodiment of the present invention. FIG. 5 illustrates an exemplary process module implemented with a Langgraph structure according to one embodiment of the present invention. FIG. 6 exemplarily illustrates a process of modifying internal attribute information of a detailed process by means of an editing interface according to an embodiment of the present invention. FIG. 7 illustrates, exemplarily, the process of performing an editing step according to another embodiment of the present invention. FIG. 8 schematically illustrates the process of determining whether a transaction is executed by a verification process according to one embodiment of the present invention. FIG. 9 schematically illustrates the detailed execution steps of a method for providing an interactive interface-based personalized auto-trading service according to an embodiment of the present invention. FIG. 10 schematically illustrates the process of performing a heterogeneous asset transaction step and a server-side transaction step according to one embodiment of the present invention. FIG. 11 illustrates, in an exemplary manner, the internal configuration of a computing device according to one embodiment of the present invention. Hereinafter, various embodiments and/or aspects are disclosed with reference to the drawings. For illust