KR-20260066681-A - General-Purpose Event-Driven Stock Recommendation System

Abstract

The present invention relates to a general-purpose event-responsive stock recommendation system (100) that recommends stock items by analyzing external data collected through a communication network. The system (100) comprises: a data collection unit (110) that collects target events in real time that cause social, industrial, environmental, and geopolitical changes from web search, social media, news feeds, and public data; an issue analysis unit (120) that identifies key issues from the collected events and derives alternative technologies or solutions to resolve the issues through a natural language processing engine (121) and a large language model (122); a company mapping unit (130) that extracts companies possessing source technologies, product portfolios, and patents related to the derived solutions from a company database (131) and maps them multidimensionally through technical similarity, patent quality indicators, supply chain analysis, and technology maturity evaluation; a stock recommendation unit (150) that generates a final investment recommendation list by analyzing the financial soundness, market data, and momentum of the mapped companies; and a reinforcement learning module (160) that self-improves the model through feedback of the recommendation results. It includes an event extinction prediction module (170) that provides a position liquidation signal by predicting the timing of event extinction. The present invention provides a significant effect of early identification of potential beneficiary companies that the market has not yet reflected, through a causal inference chain extending from the occurrence of an event to the derivation of a Technical Solution and the mapping of patent-holding companies.

Inventors

• 안범주

Assignees

• 안범주

Dates

Publication Date

20260512

Application Date

20260415

Claims (1)

1. A general-purpose event-responsive stock recommendation system comprising: a data collection unit that collects target events causing social, industrial, environmental, or geopolitical fluctuations in real time from external channels including web search, social media, news feeds, and public data; an issue analysis unit that identifies key issues including causes, scale of damage, and ripple effects from the collected target events, and derives alternative technologies or solutions necessary to resolve the key issues or minimize damage through a natural language processing engine; a company mapping unit that extracts and maps a list of companies possessing source technologies, product portfolios, or related patents associated with the derived technical solutions from a company database; and a stock recommendation unit that generates a final investment recommendation list by analyzing the market data and financial soundness of the mapped companies.

Description

General-Purpose Event-Driven Stock Recommendation System The present invention relates to a system for recommending stock items by analyzing external data collected through a communication network, and more specifically, to a general-purpose event-responsive stock recommendation system that collects target events causing social, industrial, environmental, or geopolitical changes in real time, derives technical solutions necessary to resolve the said events through a natural language processing engine, identifies companies possessing source technologies or patents related to said solutions, and generates a final investment recommendation list. In modern financial markets, the performance of stock investments is heavily dependent on the quality and processing speed of information. Traditional stock recommendation systems have performed analysis based on quantitative indicators such as historical stock price data, financial statements, and trading volume; however, this approach has limitations in that it fails to adequately reflect the structural and event-driven factors that actually move the market. With the recent advancement of Natural Language Processing (NLP) technology, attempts are being made to predict stock price fluctuations by analyzing unstructured text data such as news articles, social media, and corporate disclosures. However, existing systems focus only on analyzing the sentiment of events or their impact on companies directly associated with them; they fail to provide the functionality to infer what technical solutions are needed to address problems arising from the events and to systematically search for companies possessing the technical capabilities to provide those solutions. For example, when a large-scale wildfire occurs in a specific region, existing systems are limited to predicting stock price declines for companies in the affected area or analyzing losses for insurance-related firms; they fail to provide investors with causal inferences indicating that companies possessing new material technologies or remote control drone technologies necessary to prevent the spread of wildfires will benefit in the medium to long term. Furthermore, while thematic investment systems employ a method of searching for relevant companies based on predefined investment themes, these themes are statically set and are not dynamically derived from real-time external events; additionally, there is no structure in place to extract candidate companies by directly linking the technical capabilities required to resolve events with patent portfolios. Therefore, there is a need for an integrated system that automatically derives the technical solutions required to resolve external events, systematically identifies companies possessing such technological capabilities at the level of source patents or product portfolios, and links them to investment recommendations. FIG. 1 is a block diagram showing the overall configuration of a general-purpose event-responsive stock recommendation system (100) according to one embodiment of the present invention. FIG. 2 is a flowchart showing the event collection, classification, and reliability filtering flow of the data collection unit (110). Figure 3 is a flowchart showing the process of deriving a Technical Solution of the Issue Analysis Department (120). Figure 4 is a tree-type conceptual diagram showing the structure for predicting first, second, and third-order derivative event chains and calculating composite benefit weights. FIG. 5 is a conceptual diagram showing the structure for calculating IPC/CPC technology similarity and patent monopoly concentration of the corporate mapping unit (130). FIG. 6 is a flowchart showing the flow of the buy/sell signal switching judgment flow by the destructive technology detector (137). FIG. 7 is a conceptual diagram showing the upstream and downstream automatic expansion and damping scoring structure of a supply chain expander (140). FIG. 8 is a conceptual diagram showing the structure for generating a recommendation list by investment period by a technology maturity evaluator (141). FIG. 9 is a flowchart showing the feedback loop and model update flow of the reinforcement learning module (160). FIG. 10 is a flowchart showing the position liquidation signal generation flow of the event extinction prediction module (170). Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings. In describing the present invention, specific descriptions of related known functions or configurations are omitted if it is determined that such detailed descriptions may unnecessarily obscure the essence of the present invention. Referring to FIG. 1, a general-purpose event-responsive stock recommendation system (100) according to one embodiment of the present invention includes a data collection unit (110), an issue analysis unit (120), a company mapping unit (130), a stock recommendation un