Financial product data processing method and device and electronic equipment

Abstract

The invention discloses a financial product data processing method and device and electronic equipment, and relates to the field of financial science and technology or other related technical fields, and the method comprises the steps: recognizing entities in financial product data, and obtaining L entities; based on the L entities and the entity information of each entity, an initial knowledge graph corresponding to the financial product data is generated, and edges in the initial knowledge graph are used for representing an affiliation relationship between two entities in the L entities; a correlation matrix corresponding to the L entities is obtained, one correlation coefficient in the correlation matrix is used for representing correlation between two product entities in the L entities, and each product entity corresponds to one financial product; and based on the correlation matrix, updating the initial knowledge graph to obtain a target knowledge graph corresponding to the financial product data. The technical problem that the effect of the generated knowledge graph is poor when the knowledge graph corresponding to the financial product is created based on the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of financial technology or other related technical fields, and more specifically, to a method, apparatus, and electronic device for processing financial product data. Background Technology

[0002] In the financial industry, the construction of knowledge graphs for financial products has become an important means to improve asset management efficiency and the scientific nature of investment decisions. Current technologies primarily rely on structured databases and historical statistical analysis methods to integrate and analyze financial product information. However, existing technologies mainly focus on the static attributes of financial products, such as product type, issuing institution, and risk level, lacking refined modeling and quantification of the dynamic complementary relationships between financial products. Furthermore, existing solutions typically only reflect the data patterns of financial products at historical moments, resulting in a lag in knowledge graph updates. In summary, knowledge graphs created based on existing technologies have the following shortcomings:

[0003] (1) Simple relationship definition: Current knowledge graphs rely heavily on predefined rules and attribute relationships, which are difficult to cover the complex and ever-changing functional interactions and deep logic between financial products, especially the complementary relationships of financial products that only manifest under specific market conditions.

[0004] (2) Data silo phenomenon: The information source for existing methods to generate knowledge graphs is structured data within financial institutions. They have limited processing capabilities for open-source unstructured text data from other platforms and cannot fully integrate the transaction information contained in unstructured text data.

[0005] (3) Lagging update mechanism: The knowledge graph has a long update cycle, making it difficult to reflect the market changes of financial products in a timely manner. Especially when the market changes suddenly, the relationship information in the knowledge graph may be outdated, which may lead to deviations in the data information on which users make investment decisions.

[0006] In summary, creating knowledge graphs corresponding to financial products based on existing technologies suffers from the technical problem of producing poor-quality knowledge graphs.

[0007] There is currently no effective solution to the above problems. Summary of the Invention

[0008] This application provides a method, apparatus, and electronic device for processing financial product data, to at least solve the technical problem that the knowledge graphs generated based on existing technologies for creating knowledge graphs corresponding to financial products have poor quality.

[0009] According to one aspect of this application, a method for processing financial product data is provided, comprising: identifying entities in the financial product data to obtain L entities, where L is a positive integer, the financial product data including at least standardized transaction data and text data with contextual relevance to the financial product, and the L entities including at least a financial product, a product type, and a product issuing institution; generating an initial knowledge graph corresponding to the financial product data based on the L entities and the entity information of each entity, wherein the edges in the initial knowledge graph are used to represent the subordinate relationship between two entities among the L entities; obtaining a relevance matrix corresponding to the L entities, wherein a relevance coefficient in the relevance matrix is ​​used to represent the relevance between two product entities among the L entities, and each product entity corresponds to a financial product; and updating the initial knowledge graph based on the relevance matrix to obtain a target knowledge graph corresponding to the financial product data.

[0010] Optionally, before identifying entities in the financial product data, the processing method for the financial product data further includes: searching for the original transaction data of the financial product from a preset database of a financial institution based on a preset interface, and collecting original text data with contextual relevance to the financial product from a preset platform based on a preset tool; performing data cleaning on the original transaction data and the original text data to obtain first transaction data and first text data, wherein the data cleaning is at least used to remove blank data and redundant data in the original data; performing standardization processing on the first transaction data and the second transaction data to obtain target transaction data and target text data, wherein the standardization processing is at least used to align the entity names and data formats of the same entities in heterogeneous data; and using the target transaction data and target text data as financial product data.

[0011] Optionally, identifying entities in the financial product data to obtain L entities includes: performing word segmentation on the financial product data to obtain M words, where M is a positive integer greater than or equal to L; determining keywords in the M words based on the semantics of each word to obtain L keywords; and generating L entities based on the L keywords.

[0012] Optionally, obtaining the correlation matrix corresponding to L entities includes: filtering the L entities to obtain N product entities, where N is a positive integer less than or equal to L; determining the attribute label of each product entity based on the entity information of each product entity in the N product entities and the preset rules in the preset rule base, where the preset rules are used to assess the transaction risk corresponding to the entity information, and the attribute label is used to characterize the risk level and return information of each product entity; and determining the correlation matrix based on the attribute labels corresponding to any two product entities in the N product entities.

[0013] Optionally, a correlation matrix is ​​determined based on the attribute tags corresponding to any two product entities among the N product entities, including: determining the correlation coefficient between the attribute tags corresponding to any two product entities among the N product entities based on a sliding window mechanism; and generating a correlation matrix based on all the correlation coefficients corresponding to the N product entities.

[0014] Optionally, after obtaining the correlation matrices corresponding to L entities, the processing method for financial product data further includes: performing vector transformation on each correlation coefficient in the correlation matrix to obtain P correlation coefficient vectors corresponding to the correlation matrix, where P is a positive integer; classifying each correlation coefficient vector in the P correlation coefficient vectors to obtain the coefficient type of each correlation coefficient vector, where the coefficient type is one of the following: risk hedging type, used to represent that when a user simultaneously purchases two financial products corresponding to the correlation coefficient, transaction risk hedging can be achieved; return enhancement type, used to represent that when a user simultaneously purchases two financial products corresponding to the correlation coefficient, the transaction return rate can be increased without increasing transaction risk; liquidity complementarity type, used to represent that the market liquidity of the two financial products corresponding to the correlation coefficient is complementary.

[0015] Optionally, based on the relevance matrix, the initial knowledge graph is updated to obtain the target knowledge graph corresponding to the financial product data, including: if the i-th relevance coefficient in the relevance matrix is ​​greater than or equal to a preset coefficient threshold, the i-th relevance coefficient is taken as the target coefficient; based on the target coefficient and the coefficient type corresponding to the target coefficient, the coefficient edge corresponding to the target coefficient is determined, wherein the coefficient edge is used to connect the two product entities corresponding to the target coefficient; all coefficient edges corresponding to the target coefficients in the relevance matrix are added to the initial knowledge graph to obtain the target knowledge graph.

[0016] According to another aspect of this application, a processing apparatus for financial product data is also provided, comprising: an entity recognition unit for recognizing entities in the financial product data to obtain L entities, where L is a positive integer, the financial product data including at least standardized transaction data and text data with contextual relevance to the financial product, and the L entities including at least a financial product, a product type, and a product issuing institution; a knowledge graph generation unit for generating an initial knowledge graph corresponding to the financial product data based on the L entities and the entity information of each entity, wherein the edges in the initial knowledge graph are used to represent the subordinate relationship between two entities among the L entities; a correlation matrix acquisition unit for acquiring a correlation matrix corresponding to the L entities, wherein a correlation coefficient in the correlation matrix is ​​used to represent the correlation between two product entities among the L entities, and each product entity corresponds to a financial product; and a knowledge graph updating unit for updating the initial knowledge graph based on the correlation matrix to obtain a target knowledge graph corresponding to the financial product data.

[0017] According to another aspect of this application, a computer program product is also provided, which stores a computer program, wherein, when the computer program is running, it controls the computer program product to execute the financial product data processing method described above.

[0018] According to another aspect of this application, an electronic device is also provided, wherein the electronic device includes one or more processors and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the financial product data processing method described above.

[0019] In this application, entities in the financial product data are first identified, resulting in L entities, where L is a positive integer. The financial product data includes at least standardized transaction data and text data with contextual relevance to the financial products. The L entities include at least the financial product, product type, and product issuing institution. Then, based on the L entities and their respective entity information, an initial knowledge graph corresponding to the financial product data is generated. Edges in the initial knowledge graph represent the hierarchical relationship between two entities among the L entities. Next, a relevance matrix is ​​obtained for the L entities. A relevance coefficient in the relevance matrix represents the relevance between two product entities among the L entities. Each product entity corresponds to a financial product. Subsequently, the initial knowledge graph is updated based on the relevance matrix to obtain the target knowledge graph corresponding to the financial product data.

[0020] As described above, this application first acquires multi-source heterogeneous data (i.e., financial product data), and creates an initial relationship graph based on this data. The multi-source heterogeneous data includes at least standardized transaction data and text data with contextual relevance to financial products, thus breaking down data silos in existing technologies and increasing the richness of data sources for the knowledge graph. Next, this application creates a relevance matrix corresponding to the financial entities, and then updates the initial knowledge graph based on the relevance matrix to obtain the target knowledge graph. This achieves the goal of associating the market relevance information of the financial products represented by the relevance matrix with the initial knowledge graph, enhancing the depth and breadth of the relationships between entities in the knowledge graph, providing more accurate data support for users' subsequent financial decisions, and thus solving the technical problem of poor performance in knowledge graphs generated based on existing technologies for creating knowledge graphs corresponding to financial products. Attached Figure Description

[0021] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0022] Figure 1 This is a hardware structure block diagram of an optional computer terminal (or mobile device) for implementing a method for processing financial product data, according to an embodiment of this application.

[0023] Figure 2 This is a flowchart of an optional method for processing financial product data according to an embodiment of this application;

[0024] Figure 3 This is a schematic diagram of an optional financial product data processing apparatus according to an embodiment of this application;

[0025] Figure 4 This is a structural block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0026] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0027] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0028] It should also be noted that all relevant information (including but not limited to information used for display and analysis) and data (including but not limited to original transaction data and original text data corresponding to financial products) involved in this application are information and data authorized by the user or fully authorized by all parties. For example, this system has an interface with relevant users or institutions. Before obtaining relevant information, it needs to send an acquisition request to the aforementioned user or institution through the interface, and obtain the relevant information after receiving consent from the aforementioned user or institution.

[0029] Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of relevant information and data involved in this application all comply with the relevant laws, regulations, and standards of the relevant regions, and necessary confidentiality measures have been taken. This application does not violate public order and good morals. In addition, this application provides a corresponding operation entry point for users to choose to agree to or refuse authorization. If the user chooses to refuse authorization, the corresponding expert decision-making process will be initiated.

[0030] The present invention will now be described in detail with reference to various embodiments.

[0031] Example 1

[0032] According to an embodiment of this application, an embodiment of a method for processing financial product data is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0033] The methods and embodiments provided in this application can be executed on mobile terminals, computer terminals, or similar computing devices. Figure 1 This is a hardware structure block diagram of an optional computer terminal (or mobile device) for implementing a method for processing financial product data, according to an embodiment of this application. Figure 1As shown, the computer terminal 10 (or mobile device) may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0034] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10 (or mobile device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0035] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the financial product data processing method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the aforementioned financial product data processing method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0036] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0037] The display can be configured as a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0038] Under the aforementioned operating environment, this application provides a financial product data processing system (hereinafter referred to as the financial product data processing system) for executing the financial product data processing method of this application. Figure 2 This is a flowchart of an optional financial product data processing method according to an embodiment of this application, such as... Figure 2 As shown, the method includes the following steps:

[0039] Step S201: Identify entities in the financial product data to obtain L entities, where L is a positive integer. The financial product data includes at least standardized transaction data and text data with contextual relevance to the financial product. The L entities include at least the financial product, the product type, and the product issuing institution.

[0040] Optionally, the financial product data processed by the processing system is heterogeneous data, including at least transaction data corresponding to financial products obtained internally by financial institutions, as well as open-source text data with contextual relevance to financial products obtained from a pre-set platform, such as product announcements, regulatory documents, and market research reports.

[0041] Optionally, the processing system can use natural language processing technology to automatically extract entity information from structured transaction data and unstructured text data, thereby constructing an entity library. This provides a node foundation for the subsequent construction of the knowledge graph. This step not only improves processing efficiency but also ensures the comprehensiveness and accuracy of the acquired entity information, enabling the knowledge graph to cover a wider range of entity information, increasing the richness of the data on which entity construction is based, and thus improving the display effect of the subsequently generated knowledge graph.

[0042] Step S202: Based on the L entities and the entity information of each entity, generate an initial knowledge graph corresponding to the financial product data, wherein the edges in the initial knowledge graph are used to represent the subordinate relationship between two entities among the L entities.

[0043] Optionally, the edges in the initial knowledge graph are determined based on the subordinate relationship between financial products and product types / product issuers. For example, financial product A1 is issued through product issuer 001, and financial product B2 belongs to stock products identified as type B.

[0044] Optionally, by generating an initial knowledge graph, the processing system can intuitively display the hierarchical structure and basic relationships between entities, thereby providing a data foundation for the processing system to initially understand the financial product network. This step provides a static background for subsequent dynamic relationship analysis between product entities.

[0045] Step S203: Obtain the correlation matrix corresponding to L entities. In the correlation matrix, a correlation coefficient is used to characterize the correlation between two product entities among the L entities. Each product entity corresponds to a financial product.

[0046] Optionally, the correlation matrix can be determined by calculating the correlation coefficient between the attribute labels of any two product entities among the L entities. The matrix elements in the correlation matrix are used to measure the dynamic complementary effect between product entities. In other words, the correlation coefficient quantifies the similarity of the price / yield change trends of two product entities within a specific time window, such as trading risk hedging, market liquidity complementarity, and product yield enhancement.

[0047] Optionally, by acquiring the correlation matrix, the processing system can capture the dynamic relationship changes between products under different market conditions, thereby providing a quantitative basis for the dynamic relationship strength for subsequent updates to the knowledge graph. This enables the final constructed target knowledge graph to dynamically reflect the market trading conditions corresponding to financial products, thereby enhancing the target knowledge graph's responsiveness to real-time trading events and providing users with more accurate data support for portfolio optimization based on the target knowledge graph.

[0048] Step S204: Based on the relevance matrix, update the initial knowledge graph to obtain the target knowledge graph corresponding to the financial product data.

[0049] Optionally, based on the initial knowledge graph, the processing system achieves dynamic evolution of the knowledge graph by quantitatively analyzing the correlation between product entities in real time. This generates a target knowledge graph that can promptly reflect the impact of market changes on product relationships. The target knowledge graph not only retains static attribute relationships but also reflects dynamic complementary effects (e.g., enhanced returns, risk hedging), thereby improving the accuracy and timeliness of the generated target knowledge graph. This provides users with real-time and comprehensive relational data support for subsequent financial decisions, helping users achieve more refined risk management and return optimization, and ultimately improving the user experience.

[0050] As described above, this application first acquires multi-source heterogeneous data (i.e., financial product data), and creates an initial relationship graph based on this data. The multi-source heterogeneous data includes at least standardized transaction data and text data with contextual relevance to financial products, thus breaking down data silos in existing technologies and increasing the richness of data sources for the knowledge graph. Next, this application creates a relevance matrix corresponding to the financial entities, and then updates the initial knowledge graph based on the relevance matrix to obtain the target knowledge graph. This achieves the goal of associating the market relevance information of the financial products represented by the relevance matrix with the initial knowledge graph, enhancing the depth and breadth of the relationships between entities in the knowledge graph, providing more accurate data support for users' subsequent financial decisions, and thus solving the technical problem of poor performance in knowledge graphs generated based on existing technologies for creating knowledge graphs corresponding to financial products.

[0051] In one optional embodiment, before identifying entities in the financial product data, the processing system first searches for the original transaction data of the financial product in a preset database of a financial institution based on a preset interface, and collects the original text data with contextual relevance to the financial product from a preset platform based on a preset tool. Then, the processing system performs data cleaning on the original transaction data and the original text data to obtain first transaction data and first text data. The data cleaning is at least used to remove blank data and redundant data in the original data. Then, the processing system performs standardization processing on the first transaction data and the second transaction data to obtain target transaction data and target text data. The standardization processing is at least used to align the entity names and data formats of the same entities in heterogeneous data. Subsequently, the processing system uses the target transaction data and target text data as financial product data.

[0052] Optionally, the processing system collects raw transaction data from a preset database by calling a preset interface; at the same time, the processing system parses open-source files or HTML (HyperText Markup Language) web pages in a preset platform using preset tools to obtain raw text data that has a contextual relationship with the financial product.

[0053] Optionally, the processing system collects financial product data. This step enables the system to acquire more comprehensive data resources, including real-time transaction data and rich market text information. Data collection through pre-defined interfaces and tools improves the efficiency and automation of data acquisition, reduces errors introduced by manual operation by technical personnel, and increases the system's processing speed for both structured and unstructured data.

[0054] Optionally, by cleaning the original transaction data and original text data, the processing system can eliminate blank and redundant information in the original data, thereby reducing the amount of data that the processing system needs to process subsequently, improving the processing efficiency of the data and the speed of generating knowledge graphs. At the same time, by removing blank and redundant data, the processing system can improve the quality of the data that needs to be processed subsequently, thereby improving the effectiveness of subsequent data analysis and preventing erroneous information from interfering with system decision-making.

[0055] Optionally, the processing system standardizes the first transaction data and the second text data obtained from data cleaning, making the entity names and data formats of financial entities from different data sources consistent. This eliminates format differences between different data sources and aligns entities with the same name in heterogeneous data. It integrates financial product data from multiple channels, avoids analysis errors caused by inconsistent data formats, and improves the efficiency and accuracy of knowledge graph construction.

[0056] In summary, through the steps described in the above embodiments, the processing system can not only quickly acquire and process heterogeneous data from different information sources, but also ensure the consistency of financial product data processing, thereby providing a more accurate data foundation for the subsequent construction and maintenance of a comprehensive knowledge graph of financial products.

[0057] In one optional embodiment, the processing system first performs word segmentation on the financial product data to obtain M words, where M is a positive integer greater than or equal to L. Then, based on the semantics of each word in the M words, the processing system determines the keywords in the M words to obtain L keywords. Finally, based on the L keywords, the processing system generates L entities.

[0058] Optionally, by performing word segmentation on financial product data, the processing system can decompose continuous text data into basic independent units with semantic meaning, which facilitates the next step of semantic understanding and keyword determination. Through word segmentation, the processing system can better understand the text content, thereby providing a clear lexical foundation for subsequent entity recognition and relation extraction, and thus improving the accuracy of information extracted from unstructured text.

[0059] Optionally, the processing system can employ a pre-defined semantic analysis method, such as term frequency-inverse document frequency statistics, to determine keywords among M words. The determination of keywords enables the processing system to focus on entity information that is truly valuable for building the knowledge graph. Through semantic analysis, the processing system can identify words that can effectively represent the characteristics of financial products, such as "stocks," "fixed income," and "issuer," thereby improving the efficiency and quality of subsequent knowledge graph construction, reducing the noise data that the processing system needs to process, and making the final knowledge graph more refined and accurate.

[0060] Optionally, the processing system transforms keywords into entities, realizing the conversion from text data to structured knowledge. This helps the processing system to understand the role and positioning of financial products in the market and provides a data foundation for further analysis of the relationships between entities. The generation of different types of entities increases the richness and practicality of the knowledge graph.

[0061] In summary, through word segmentation, keyword identification, and entity generation, the processing system is able to extract semantically clear entity information from complex financial product data, providing high-quality data support for the subsequent construction of knowledge graphs.

[0062] In one optional embodiment, the processing system first filters L entities to obtain N product entities, where N is a positive integer less than or equal to L. Then, based on the entity information of each product entity among the N product entities and the preset rules in the preset rule base, the processing system determines the attribute tags of each product entity. The preset rules are used to assess the transaction risk corresponding to the entity information, and the attribute tags are used to characterize the risk level and return information of each product entity. Then, the processing system determines the correlation matrix based on the attribute tags corresponding to any two product entities among the N product entities.

[0063] Optionally, the processing system can dynamically extract the correlation relationships between N product entities based on graph neural networks.

[0064] Optionally, the processing system removes entities that are not financial products (such as product issuers and product types) by screening product entities, thereby reducing the amount of data that needs to be analyzed for correlation and improving the speed of generating the correlation matrix.

[0065] Optionally, the processing system, based on the entity information of N product entities, assesses the transaction risk of the financial product corresponding to each entity through preset rules in a preset rule base, thereby assigning attribute tags to the entities. The attribute tags are used to characterize the risk level and return information of each product entity in the financial market transaction process. The determination of attribute tags makes the entity information in the knowledge graph richer and more insightful. By converting risk level and return information into attribute tags for product entities, the processing system improves the information richness and accuracy of the subsequently generated target knowledge graph.

[0066] Optionally, the processing system constructs a correlation matrix based on the attribute labels corresponding to any two product entities among the N product entities, which can reflect the dynamic complementary effect between products. Each correlation coefficient in the correlation matrix is ​​used to quantify the dynamic correlation effect of the financial products corresponding to the two product entities in the market transaction process, such as transaction risk hedging, market liquidity complementarity, and product return enhancement.

[0067] Optionally, the processing system can identify financial product portfolios that can achieve preset dynamic correlation effects under specific market conditions by dynamically determining the correlation coefficients between product entities. This step realizes the leap from static attribute description to dynamic relationship analysis, enhances the timeliness and financial decision support capabilities of the final generated target knowledge graph, and thus improves the accuracy of users' financial decisions based on the target knowledge graph.

[0068] In one alternative embodiment, the processing system first determines the correlation coefficient between the attribute tags of any two product entities among the N product entities based on a sliding window mechanism. Then, the processing system generates a correlation matrix based on all the correlation coefficients corresponding to the N product entities.

[0069] Optionally, the processing system determines the correlation coefficient between product entities based on a sliding window mechanism. The sliding window mechanism is a dynamic data analysis method in which the processing system sets a fixed-size time window, and the time window scrolls continuously over time to capture the changing trend of product attributes within each window.

[0070] Optionally, the processing system can capture changes and correlations in product entity attributes in the short term through a sliding window mechanism, thereby making the correlation coefficient closer to the actual market situation and providing dynamic and real-time data support for the subsequent generated correlation matrix, thus enhancing the timeliness and responsiveness of the knowledge graph.

[0071] Optionally, the processing system collects the correlation coefficients between all attribute tags corresponding to any two product entities among the N product entities, and obtains a correlation matrix. The correlation matrix is ​​a two-dimensional array, where each matrix element represents the correlation coefficient between two product entities corresponding to the matrix row and matrix column.

[0072] Optionally, the processing system generates a correlation matrix, which comprehensively reflects the dynamic interaction between product entities over time, thereby capturing the impact of changes in market conditions on the relationships between products. This step achieves a quantitative description of the dynamic complementary effects of financial products, providing a detailed and accurate data foundation for subsequent map updates. The correlation matrix not only reveals the direct relationships between products but also, through further analysis, discovers indirect or complex complementary relationships. This is particularly important for designing portfolio strategies that can adapt to market fluctuations, effectively improving the flexibility of asset allocation and risk control capabilities.

[0073] In summary, the processing system uses a sliding window mechanism to determine the correlation coefficients between the attribute tags of product entities, thereby constructing a correlation matrix. This enables real-time capture and quantification of dynamic relationships between financial products, overcoming the limitations of static graphs in adapting to market changes and enhancing the practical value and decision support capabilities of the final knowledge graph.

[0074] In one optional embodiment, the processing system first performs vector transformation on each correlation coefficient in the correlation matrix to obtain P correlation coefficient vectors corresponding to the correlation matrix, where P is a positive integer. Then, the processing system classifies each of the P correlation coefficient vectors to obtain the coefficient type of each correlation coefficient vector.

[0075] Optionally, the coefficient type is one of the following:

[0076] Risk hedging type is used to represent that users can hedge transaction risks by purchasing two financial products with the same correlation coefficient at the same time.

[0077] Enhanced returns type is used to indicate that when a user purchases two financial products with the same correlation coefficient, the transaction return rate can be increased without increasing the transaction risk.

[0078] Liquidity complementarity type is used to characterize the complementary relationship between the market liquidity of two financial products corresponding to the correlation coefficient.

[0079] Optionally, after generating the correlation matrix, the processing system can determine the coefficient type corresponding to each correlation coefficient based on a pre-built multi-label classifier, such as risk hedging type, yield enhancement type, and liquidity complementarity type.

[0080] Optionally, the processing system transforms and classifies each correlation coefficient into a categorization information that directly guides financial decision-making by performing vector transformation and classification on each correlation coefficient. For example, a correlation coefficient vector for risk hedging indicates that two financial products can hedge each other's risks under different market conditions, while a correlation coefficient vector for return enhancement indicates that the combination of two financial products can improve the expected rate of return while keeping the risk level constant. This step not only improves the intelligence level of the knowledge graph but also directly serves the optimization of financial decision-making and enhances the risk management capabilities in the financial transaction process.

[0081] In one optional embodiment, if the i-th correlation coefficient in the correlation matrix is ​​greater than or equal to a preset coefficient threshold, the processing system takes the i-th correlation coefficient as the target coefficient. Then, based on the target coefficient and the coefficient type corresponding to the target coefficient, the processing system determines the coefficient edge corresponding to the target coefficient, wherein the coefficient edge is used to connect the two product entities corresponding to the target coefficient. Then, the processing system adds all the coefficient edges corresponding to the target coefficients in the correlation matrix to the initial knowledge graph to obtain the target knowledge graph.

[0082] Optionally, the processing system uses the correlation coefficient as the target coefficient only if the correlation coefficient is greater than or equal to a preset coefficient threshold. The determination of the target coefficient is based on the quantitative screening of the coefficient strength in the correlation matrix. This step ensures that only those financial product portfolios that truly have market correlation are highlighted. By setting a threshold, the processing system can filter out financial product portfolios with weak or no market correlation, thereby improving the accuracy of the subsequently generated coefficient edges.

[0083] Optionally, the processing system constructs coefficient edges corresponding to the target coefficients and adds these edges to the initial knowledge graph. This allows the graph to not only show which products have complementary relationships but also to clarify the nature and strength of these relationships. Through the identification and addition of coefficient edges, the final generated target knowledge graph not only displays the basic information of product entities but also reflects their interactions under different market conditions. This achieves a comprehensive capture of the complex relationships between financial products. Based on the target knowledge graph, users can intuitively see which products can effectively hedge risks, enhance returns, or provide liquidity support. This enables them to make more rational and evidence-based asset allocation decisions, thereby improving the user experience.

[0084] Optionally, the coefficient edge finally constructed by the processing system can be represented in the form of a weighted triple. For example, if there is a risk hedging relationship between financial product A1 and financial product B1, and the correlation coefficient between the two is 0.85, the coefficient edge can be represented as <product A1, hedging coefficient 0.85, product B1>.

[0085] In one optional embodiment, the processing system adopts a hierarchical data processing architecture, consisting of a multi-source data acquisition module, a dynamic knowledge extraction module, a complementary relationship quantification module, and a knowledge graph visualization module. The modules communicate loosely through API interfaces and message queues. Data streams are transmitted and processed in the order of acquisition, extraction, quantification, and visualization. The functions of different modules are as follows:

[0086] (1) Multi-source data acquisition module:

[0087] Input: Structured data (i.e., raw transaction data) + unstructured data (i.e., raw text data).

[0088] Operation: Perform data cleaning and standardization on the input raw data.

[0089] Output: Standardized financial product data, stored in a pre-defined distributed database.

[0090] (2) Dynamic knowledge extraction module:

[0091] Input: Financial product data in a standardized format.

[0092] operate:

[0093] 1) Identify entities in the text, such as financial products, product issuers, and types of financial products;

[0094] 2) Extract the relationships between product entities:

[0095] Matching static attribute relationships based on preset rules (e.g., "Product A belongs to the bond category");

[0096] Dynamically extract the relationships between products based on graph neural networks (GNNs) (e.g., "there is a risk hedging relationship between product B1 and product C1").

[0097] 3) Determine the type of association (risk hedging / return enhancement / liquidity complementarity) through a multi-label classifier.

[0098] Output: The relationships between products and the types of relationships.

[0099] (3) Complementary Relationship Quantification Module:

[0100] Input: Attribute tags corresponding to product combinations that have a relationship.

[0101] Operation: Dynamic weight calculation, based on a sliding window mechanism to calculate the correlation coefficient between products.

[0102] Output: Weighted knowledge graph triples <Product A1, hedging coefficient 0.85, Product B1>.

[0103] (4) Knowledge Graph Visualization Module:

[0104] Front-end: Employs dynamic rendering technology to generate a dynamically displayable knowledge graph;

[0105] Backend: A graph database is used to store the relationships between product entities;

[0106] Interaction logic: Users input the investment risk level of financial products into the processing system, and the processing system returns an optimized portfolio path, such as "60% bonds + 40% stocks"; the correlation heatmap is updated in real time, and the color depth represents the correlation strength.

[0107] In summary, existing financial product knowledge graphs only statically store basic product attributes (such as yield and maturity), failing to dynamically quantify real-time correlations and complementary effects between products (e.g., dynamic changes in stock-bond hedging relationships), resulting in a lack of real-time data support for portfolio optimization. This processing system addresses this shortcoming by constructing a dynamic relational knowledge graph, achieving the following technical effects:

[0108] (1) Breaking through the limitations of static data and realizing the dynamic quantification of complementary effects: The processing system introduces text data to construct a knowledge graph for the first time, adopts a multi-source heterogeneous data fusion mechanism, and combines entity recognition and relationship classification in NLP relationship extraction technology. It can dynamically identify complementary relationships such as risk hedging and return enhancement between products. By upgrading the traditional static attribute graph to a real-time relational graph, the processing system can provide a real-time data engine for the combination optimization of financial products.

[0109] (2) Solving the technical blind spot of "dynamic relationship modeling": The processing system fills the gap in the existing knowledge graph modeling of cross-category product relationship, and reveals the implicit complementary logic that traditional methods cannot capture (for example, inferring the hedging relationship between gold and inflation bonds through policy announcements), thereby improving the robustness of financial product portfolio strategies generated based on the target knowledge graph.

[0110] (3) Improve the real-time performance and completeness of the knowledge graph: The processing system adopts streaming data processing + incremental knowledge fusion technology, solves product name ambiguity through concept disambiguation, and completes implicit relationships by link prediction, thus constructing a dynamic knowledge graph with high completeness and low redundancy, reducing financial decision-making bias caused by data fragmentation.

[0111] (4) Enable interactive insights into product relationships. The processing system uses a visualization engine to map multi-dimensional complementary relationships (product-risk-time) into a dynamic network topology, supporting penetrating queries (such as clicking on a fund to display all its associated hedging instruments), thereby helping users to more intuitively understand the internal relationship logic of financial product portfolios, reducing the transaction threshold of financial products, and improving user experience.

[0112] As described above, this application first acquires multi-source heterogeneous data (i.e., financial product data), and creates an initial relationship graph based on this data. The multi-source heterogeneous data includes at least standardized transaction data and text data with contextual relevance to financial products, thus breaking down data silos in existing technologies and increasing the richness of data sources for the knowledge graph. Next, this application creates a relevance matrix corresponding to the financial entities, and then updates the initial knowledge graph based on the relevance matrix to obtain the target knowledge graph. This achieves the goal of associating the market relevance information of the financial products represented by the relevance matrix with the initial knowledge graph, enhancing the depth and breadth of the relationships between entities in the knowledge graph, providing more accurate data support for users' subsequent financial decisions, and thus solving the technical problem of poor performance in knowledge graphs generated based on existing technologies for creating knowledge graphs corresponding to financial products.

[0113] Example 2

[0114] This application embodiment can also provide a financial product data processing device. It should be noted that the financial product data processing device of this application embodiment can be used to execute the financial product data processing method provided in this application embodiment. The following is a description of the financial product data processing device provided in this application embodiment.

[0115] According to an embodiment of this application, an apparatus for implementing the above-described financial product data processing method is also provided. Figure 3 This is a schematic diagram of an optional financial product data processing apparatus according to an embodiment of this application, such as... Figure 3 As shown, the device includes: an entity recognition unit 301, a knowledge graph generation unit 302, a relevance matrix acquisition unit 303, and a knowledge graph update unit 304.

[0116] Optionally, the entity recognition unit 301 is used to identify entities in the financial product data to obtain L entities, where L is a positive integer. The financial product data includes at least standardized transaction data and text data with contextual relevance to the financial products. The L entities include at least financial products, product types, and product issuing institutions. The knowledge graph generation unit 302 is used to generate an initial knowledge graph corresponding to the financial product data based on the L entities and the entity information of each entity. The edges in the initial knowledge graph are used to represent the subordinate relationship between two entities among the L entities. The relevance matrix acquisition unit 303 is used to acquire the relevance matrix corresponding to the L entities. A relevance coefficient in the relevance matrix is ​​used to represent the relevance between two product entities among the L entities. Each product entity corresponds to a financial product. The knowledge graph update unit 304 updates the initial knowledge graph based on the relevance matrix to obtain the target knowledge graph corresponding to the financial product data.

[0117] In one optional embodiment, the financial product data processing apparatus further includes: a raw transaction data search unit, a data cleaning unit, a standardization processing unit, and a financial product data determination unit.

[0118] Optionally, the original transaction data lookup unit is used to look up the original transaction data of financial products from the preset database of financial institutions based on a preset interface, and to collect the original text data with contextual relevance to the financial products from a preset platform based on a preset tool; the data cleaning unit is used to clean the original transaction data and the original text data to obtain the first transaction data and the first text data, wherein the data cleaning is at least used to remove blank data and redundant data in the original data; the standardization processing unit is used to standardize the first transaction data and the second transaction data to obtain the target transaction data and the target text data, wherein the standardization processing is at least used to align the entity names and data formats of the same entities in the heterogeneous data; and the financial product data determination unit is used to use the target transaction data and the target text data as the financial product data.

[0119] In one optional embodiment, the entity recognition unit 301 includes: a word segmentation processing subunit, a keyword extraction subunit, and an entity generation subunit.

[0120] Optionally, the word segmentation subunit is used to segment the financial product data to obtain M words, where M is a positive integer greater than or equal to L; the keyword extraction subunit is used to determine the keywords in the M words based on the semantics of each word in the M words, to obtain L keywords; and the entity generation subunit is used to generate L entities based on the L keywords.

[0121] In one optional embodiment, the relevance matrix acquisition unit 303 includes: a product entity filtering subunit, an attribute tag determination subunit, and a relevance matrix determination subunit.

[0122] Optionally, the product entity filtering subunit is used to filter L entities to obtain N product entities, where N is a positive integer less than or equal to L; the attribute label determination subunit is used to determine the attribute label of each product entity based on the entity information of each product entity in the N product entities and the preset rules in the preset rule base, where the preset rules are used to assess the transaction risk corresponding to the entity information, and the attribute label is used to characterize the risk level and return information of each product entity; the correlation matrix determination subunit is used to determine the correlation matrix based on the attribute labels corresponding to any two product entities in the N product entities.

[0123] In one optional embodiment, the correlation matrix determination subunit includes a correlation coefficient determination module and a correlation matrix generation module.

[0124] Optionally, the correlation coefficient determination module is used to determine the correlation coefficient between the attribute tags of any two product entities among the N product entities based on a sliding window mechanism; the correlation matrix generation module is used to generate a correlation matrix based on all the correlation coefficients corresponding to the N product entities.

[0125] In one alternative embodiment, the financial product data processing apparatus further includes a vector transformation unit and a coefficient classification unit.

[0126] Optionally, a vector transformation unit is used to perform vector transformation on each correlation coefficient in the correlation matrix to obtain P correlation coefficient vectors corresponding to the correlation matrix, where P is a positive integer; a coefficient classification unit is used to classify each of the P correlation coefficient vectors to obtain the coefficient type of each correlation coefficient vector, where the coefficient type is one of the following: risk hedging type, used to represent that when a user buys two financial products corresponding to the correlation coefficient at the same time, hedging of transaction risk can be achieved; return enhancement type, used to represent that when a user buys two financial products corresponding to the correlation coefficient at the same time, the transaction return rate can be increased without increasing transaction risk; liquidity complementarity type, used to represent that the market liquidity of the two financial products corresponding to the correlation coefficient is complementary.

[0127] In one optional embodiment, the knowledge graph update unit 304 includes: a target coefficient determination subunit, a coefficient edge determination subunit, and a coefficient edge addition subunit.

[0128] Optionally, the target coefficient determination subunit is used to take the i-th correlation coefficient as the target coefficient if the i-th correlation coefficient in the correlation matrix is ​​greater than or equal to a preset coefficient threshold; the coefficient edge determination subunit is used to determine the coefficient edge corresponding to the target coefficient based on the target coefficient and the coefficient type corresponding to the target coefficient, wherein the coefficient edge is used to connect the two product entities corresponding to the target coefficient; the coefficient edge adding subunit is used to add all the coefficient edges corresponding to the target coefficients in the correlation matrix to the initial knowledge graph to obtain the target knowledge graph.

[0129] As described above, this application first acquires multi-source heterogeneous data (i.e., financial product data), and creates an initial relationship graph based on this data. The multi-source heterogeneous data includes at least standardized transaction data and text data with contextual relevance to financial products, thus breaking down data silos in existing technologies and increasing the richness of data sources for the knowledge graph. Next, this application creates a relevance matrix corresponding to the financial entities, and then updates the initial knowledge graph based on the relevance matrix to obtain the target knowledge graph. This achieves the goal of associating the market relevance information of the financial products represented by the relevance matrix with the initial knowledge graph, enhancing the depth and breadth of the relationships between entities in the knowledge graph, providing more accurate data support for users' subsequent financial decisions, and thus solving the technical problem of poor performance in knowledge graphs generated based on existing technologies for creating knowledge graphs corresponding to financial products.

[0130] It should be noted that the entity recognition unit 301, knowledge graph generation unit 302, relevance matrix acquisition unit 303, and knowledge graph update unit 304 mentioned above correspond to steps S201 to S204 in the method embodiment. The instances and application scenarios implemented by the above units and corresponding steps are the same, but are not limited to the content disclosed in the above embodiments. It should be noted that the above modules or units can be hardware or software components stored in memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n). The above modules can also be part of the device and run in the computer terminal 10 provided in the embodiment.

[0131] Example 3

[0132] Embodiments of this application can also provide an electronic device. Figure 4 This is a structural block diagram of an electronic device according to an embodiment of this application, such as... Figure 4 As shown, the electronic device includes: one or more ( Figure 4 (Only one is shown) Processor 402, memory 404, memory controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0133] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and devices in the embodiments of this application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby realizing the above-mentioned method for processing financial product data.

[0134] The memory may include high-speed random access memory (RAM), and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, which can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks (LANs), mobile communication networks, and combinations thereof.

[0135] The processor can access information and applications stored in memory via a transmission device to perform the following steps: Identify entities in the financial product data to obtain L entities, where L is a positive integer. The financial product data includes at least standardized transaction data and text data with contextual relevance to the financial products. The L entities include at least a financial product, a product type, and a product issuing institution. Based on the L entities and their entity information, generate an initial knowledge graph corresponding to the financial product data, where edges in the initial knowledge graph represent the dependency relationship between two entities among the L entities. Obtain a relevance matrix corresponding to the L entities, where a relevance coefficient in the relevance matrix represents the relevance between two product entities among the L entities, and each product entity corresponds to a financial product. Update the initial knowledge graph based on the relevance matrix to obtain the target knowledge graph corresponding to the financial product data.

[0136] The processor can invoke information and applications stored in memory via a transmission device to perform the following steps: Based on a preset interface, retrieve the original transaction data of financial products from a preset database of a financial institution, and based on a preset tool, collect original text data with contextual relevance to the financial products from a preset platform; perform data cleaning on the original transaction data and original text data to obtain first transaction data and first text data, wherein data cleaning is at least used to remove blank and redundant data from the original data; perform standardization processing on the first transaction data and second transaction data to obtain target transaction data and target text data, wherein standardization processing is at least used to align the entity names and data formats of the same entities in heterogeneous data; and use the target transaction data and target text data as financial product data.

[0137] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: perform word segmentation on the financial product data to obtain M words, where M is a positive integer greater than or equal to L; determine the keywords in the M words based on the semantics of each word to obtain L keywords; generate L entities based on the L keywords.

[0138] The processor can access information and applications stored in the memory via a transmission device to perform the following steps: filtering L entities to obtain N product entities, where N is a positive integer less than or equal to L; determining the attribute tags for each product entity based on its entity information and preset rules in a preset rule base, where the preset rules are used to assess the transaction risk corresponding to the entity information, and the attribute tags are used to characterize the risk level and return information of each product entity; and determining a correlation matrix based on the attribute tags corresponding to any two product entities among the N product entities.

[0139] The processor can access the information and application programs stored in the memory via the transmission device to perform the following steps: based on the sliding window mechanism, determine the correlation coefficient between the attribute tags of any two product entities among the N product entities; based on all the correlation coefficients corresponding to the N product entities, generate a correlation matrix.

[0140] The processor can access information and applications stored in the memory via a transmission device to execute the following steps: Perform vector transformation on each correlation coefficient in the correlation matrix to obtain P correlation coefficient vectors corresponding to the correlation matrix, where P is a positive integer; classify each of the P correlation coefficient vectors to obtain the coefficient type of each correlation coefficient vector, where the coefficient type is one of the following: risk hedging type, used to represent that when a user simultaneously purchases two financial products corresponding to the correlation coefficient, transaction risk can be hedging; return enhancement type, used to represent that when a user simultaneously purchases two financial products corresponding to the correlation coefficient, the transaction return rate can be increased without increasing transaction risk; liquidity complementarity type, used to represent that the market liquidity of the two financial products corresponding to the correlation coefficient is complementary.

[0141] Optionally, based on the relevance matrix, the initial knowledge graph is updated to obtain the target knowledge graph corresponding to the financial product data, including: if the i-th relevance coefficient in the relevance matrix is ​​greater than or equal to a preset coefficient threshold, the i-th relevance coefficient is taken as the target coefficient; based on the target coefficient and the coefficient type corresponding to the target coefficient, the coefficient edge corresponding to the target coefficient is determined, wherein the coefficient edge is used to connect the two product entities corresponding to the target coefficient; all coefficient edges corresponding to the target coefficients in the relevance matrix are added to the initial knowledge graph to obtain the target knowledge graph.

[0142] This application provides a solution for processing financial product data. First, it acquires multi-source heterogeneous data (i.e., financial product data), and creates an initial relationship graph based on this data. The multi-source heterogeneous data includes at least standardized transaction data and text data with contextual relevance to financial products, thus breaking down data silos in existing technologies and increasing the richness of data sources for the knowledge graph. Next, it creates a relevance matrix corresponding to the financial entities, and then updates the initial knowledge graph based on the relevance matrix to obtain the target knowledge graph. This achieves the goal of associating the market relevance information of the financial products represented by the relevance matrix with the initial knowledge graph, enhancing the depth and breadth of the relationships between entities in the knowledge graph. This provides more accurate data support for users' subsequent financial decisions, thereby solving the technical problem of poor performance in knowledge graphs generated based on existing technologies for creating financial product knowledge graphs.

[0143] Those skilled in the art will understand that Figure 4 The structure shown is for illustrative purposes only. Electronic devices can also be smartphones, tablets, PDAs, mobile internet devices, PADs, and other terminal devices. Figure 4 This does not limit the structure of the aforementioned electronic device. For example, electronic devices may also include components that are more... Figure 4 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 4 The different configurations shown.

[0144] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0145] Example 4

[0146] Embodiments of this application may also provide a storage medium.

[0147] Optionally, in this embodiment of the application, the storage medium can be used to store the program code executed by the financial product data processing method provided in the above method embodiment.

[0148] Optionally, in this embodiment, the storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0149] This application also provides a computer program product, which, when executed on a data processing device, is suitable for performing the steps of a method for processing financial product data.

[0150] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0151] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0152] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0153] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0154] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0155] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0156] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method of processing financial product data, characterized by, The method comprises the following steps: identifying entities in financial product data to obtain L entities, wherein L is a positive integer, the financial product data at least includes standardized transaction data and text data contextually associated with the financial product, and the L entities at least include a financial product, a product type, and a product issuing institution; generating an initial knowledge graph corresponding to the financial product data based on the L entities and entity information of each entity, wherein edges in the initial knowledge graph represent a dependency relationship between two entities in the L entities; obtaining a correlation matrix corresponding to the L entities, wherein a correlation coefficient in the correlation matrix represents the correlation between two product entities in the L entities, and each product entity corresponds to a financial product; updating the initial knowledge graph based on the correlation matrix to obtain a target knowledge graph corresponding to the financial product data.

2. The method of processing financial product data according to claim 1, wherein, Before identifying entities in the financial product data, the processing method of the financial product data further comprises: based on a preset interface, searching for original transaction data of the financial product from a preset database of a financial institution, and based on a preset tool, collecting original text data contextually associated with the financial product from a preset platform; performing data cleaning on the original transaction data and the original text data to obtain first transaction data and first text data, wherein the data cleaning at least removes blank data and redundant data in the original data; performing standardization processing on the first transaction data and the second transaction data to obtain the target transaction data and the target text data, wherein the standardization processing at least aligns entity names and data formats of the same entity in heterogeneous data; taking the target transaction data and the target text data as the financial product data.

3. The method of processing financial product data according to claim 1, wherein, Identifying entities in the financial product data to obtain L entities comprises: performing word segmentation processing on the financial product data to obtain M words, wherein M is a positive integer greater than or equal to L; determining key words in the M words based on the semantics of each word in the M words to obtain L key words; generating the L entities based on the L key words.

4. The method of processing financial product data according to claim 1, wherein, Obtaining a correlation matrix corresponding to the L entities comprises: screening the L entities to obtain N product entities, wherein N is a positive integer less than or equal to L; determining an attribute label of each product entity in the N product entities based on entity information of the product entity and a preset rule in a preset rule library, wherein the preset rule is used to evaluate transaction risk corresponding to the entity information, and the attribute label is used to represent a risk level and return information of the product entity; determining the correlation matrix based on the attribute labels corresponding to any two product entities in the N product entities.

5. The method of processing financial product data according to claim 4, wherein, Determining the correlation matrix based on the attribute labels corresponding to any two product entities in the N product entities comprises: Determine a correlation coefficient between attribute labels corresponding to any two product entities in the N product entities based on a sliding window mechanism; Generate the correlation matrix based on all correlation coefficients corresponding to the N product entities.

6. The method of processing financial product data according to claim 1, wherein, After obtaining the correlation matrix corresponding to the L entities, the processing method of the financial product data further includes: Vector transform each correlation coefficient in the correlation matrix to obtain P correlation coefficient vectors corresponding to the correlation matrix, where P is a positive integer; Classify each correlation coefficient vector in the P correlation coefficient vectors to obtain a coefficient type of each correlation coefficient vector, where the coefficient type is one of the following: A risk hedging type for representing that a user simultaneously purchases two financial products corresponding to the correlation coefficient, and can achieve transaction risk hedging; A yield enhancement type for representing that a user simultaneously purchases two financial products corresponding to the correlation coefficient, and can improve transaction yield without enhancing transaction risk; A liquidity complement type for representing that market liquidity of two financial products corresponding to the correlation coefficient is in a complementary relationship.

7. The method of processing financial product data according to claim 1, wherein, Update the initial knowledge graph based on the correlation matrix to obtain a target knowledge graph corresponding to the financial product data, including: In a case where an i-th correlation coefficient in the correlation matrix is greater than or equal to a preset coefficient threshold, the i-th correlation coefficient is taken as a target coefficient; Determine a coefficient edge corresponding to the target coefficient based on the target coefficient and a coefficient type corresponding to the target coefficient, where the coefficient edge is used to connect two product entities corresponding to the target coefficient; Add the coefficient edges corresponding to all target coefficients in the correlation matrix to the initial knowledge graph to obtain the target knowledge graph.

8. An apparatus for processing financial product data, characterized by Comprise: An entity recognition unit configured to recognize entities in financial product data to obtain L entities, where L is a positive integer, the financial product data at least includes standardized transaction data and text data having a context association with a financial product, and the L entities at least include a financial product, a product type, and a product issuing institution; A knowledge graph generation unit configured to generate an initial knowledge graph corresponding to the financial product data based on the L entities and entity information of each entity, where an edge in the initial knowledge graph is used to represent a dependency relationship between two entities in the L entities; A correlation matrix acquisition unit configured to obtain a correlation matrix corresponding to the L entities, where a correlation coefficient in the correlation matrix is used to represent a correlation between two product entities in the L entities, and each product entity corresponds to a financial product; A knowledge graph updating unit configured to update the initial knowledge graph based on the correlation matrix to obtain a target knowledge graph corresponding to the financial product data.

9. A computer program product, characterised in that, The computer program product includes a computer program, wherein the computer program product performs the processing method of the financial product data according to any one of claims 1 to 7 when the computer program runs.

10. An electronic device, comprising: The computer program product includes one or more processors and a memory, and the memory is used to store one or more programs, wherein the one or more programs make the one or more processors implement the processing method of the financial product data according to any one of claims 1 to 7 when the one or more programs are executed by the one or more processors.

Images