Atlas index construction method, device, equipment, medium and program product

By constructing a product graph and calculating the association strength value, products with strong associations are grouped into the same set, which solves the problem of high query latency in existing technologies and enables efficient product query and recommendation.

CN123115477APending Publication Date: 2026-11-17CHINA CONSTRUCTION BANK +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611602219.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-16
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In existing technologies, graph databases store index information according to product identifiers without considering the relationships between products, resulting in high query latency and difficulty in quickly locating financial products that match user needs.

Method used

Construct a product graph, determine associated nodes based on the association type of the edges, build product chains and calculate the association strength value, divide product sets, and build query indexes based on product sets, storing index data of the same product set in the same storage shard.

Benefits of technology

It improves query efficiency, reduces network communication overhead and cross-slice index lookup latency, enhances the matching degree between query results and user needs, and facilitates product recommendations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123115477A_ABST
    Figure CN123115477A_ABST
Patent Text Reader

Abstract

The present disclosure provides a graph index construction method, device, equipment, medium and program product, relating to the technical field of big data, the technical field of knowledge graph, and the technical field of finance. The graph index construction method comprises: obtaining product graph and information of historical products held by a user; determining a plurality of associated nodes associated with the historical products from a plurality of nodes based on an associated type corresponding to an edge, and constructing a product chain for each associated product; calculating an associated strength value of the product chain based on the associated type between a plurality of products included in the product chain; determining a plurality of product sets according to the associated strength value of each product chain; and constructing a plurality of query indexes corresponding to the plurality of product sets based on the feature information of the products, wherein the data in the same query index is stored in the same storage shard.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This disclosure relates to the fields of big data technology, knowledge graph technology, and financial technology, and more specifically to a method, apparatus, device, medium, and program product for constructing a knowledge graph index. Background Technology

[0002] With the continuous development of the financial market, the number and types of financial products on the market are constantly increasing. When users choose products or allocate assets, they often face a large number of options and find it difficult to quickly locate products that match their needs. To improve the efficiency of product search and recommendation, a graph structure can be introduced to organize product data. With the help of a product graph, the relationships between products can be expressed, thereby leveraging the correlation characteristics between products during product search and recommendation.

[0003] In related technologies, graph databases typically store index information based on product identifiers, without considering the relationships between products. When querying a product from the graph, the search can only be performed product by product, resulting in high query latency. Summary of the Invention

[0004] In view of the above problems, this disclosure provides a method, apparatus, device, medium and program product for constructing a spectral index.

[0005] According to the first aspect of this disclosure, a method for constructing a map index is provided, comprising:

[0006] Obtain product graphs and information on historical products held by users. In the product graph, multiple nodes represent multiple products, the edges between nodes represent the association types between products, and the attributes of nodes represent the characteristic information of products.

[0007] Based on the association type corresponding to the edge, multiple association nodes associated with historical products are determined from multiple nodes, and a product chain is constructed for each associated product, where the association node represents the associated product associated with the historical product;

[0008] The association strength value of the product chain is calculated based on the association type between multiple related products included in the product chain.

[0009] Multiple product sets are determined based on the correlation strength values ​​of each product chain.

[0010] Based on the feature information of related products, construct multiple sets of query indexes corresponding to multiple product sets.

[0011] According to embodiments of this disclosure, it further includes:

[0012] Store multiple sets of user query indexes in the first storage shard;

[0013] Based on the full range of products corresponding to the product map, construct a full query index corresponding to all products;

[0014] The full query index is stored in the second storage shard.

[0015] According to embodiments of this disclosure, it further includes:

[0016] Obtain product update information and identify at least one product to be updated based on the product update information;

[0017] Determine at least one set of products to be updated, corresponding to at least one product to be updated;

[0018] Update the query index corresponding to at least one set of products to be updated, based on the product update information.

[0019] According to embodiments of this disclosure, it further includes:

[0020] Obtain user transaction data;

[0021] Based on user transaction data and information about the user's historical products, determine the expected product characteristic values;

[0022] Based on the expected product feature value and multiple sets of query indexes, at least one candidate product matching the expected product feature value is obtained from multiple products included in the product map.

[0023] The recommended products are determined based on the type of association between at least one candidate product and the user's historical products.

[0024] According to embodiments of this disclosure, the association type includes one of substitute products, complementary products, and mutually exclusive products;

[0025] Based on the association type corresponding to the edge, the multiple associated nodes related to historical products are determined from multiple nodes, including:

[0026] The first type of node, representing historical products, and the second type of node, whose association type with the first type of node belongs to substitute products and / or complementary products, are identified as multiple associated nodes.

[0027] According to embodiments of this disclosure, the product to be recommended is determined based on the association type between at least one candidate product and the user's historical products, including:

[0028] Remove candidate products whose association type is mutually exclusive or that are not associated with historical products from at least one candidate product;

[0029] Candidate products whose association type is "substitute product" are prioritized as first priority.

[0030] Candidate products with a complementary product association type are prioritized as the second priority, where the first priority is higher than the second priority;

[0031] Based on the candidate products corresponding to the first and second priorities, the products to be recommended are determined.

[0032] According to embodiments of this disclosure, calculating the association strength value of a product chain based on the association type among multiple related products included in the product chain includes:

[0033] Based on the association type among multiple related products included in the product chain, determine the contribution value corresponding to at least one edge in the product chain;

[0034] The association strength value of the product chain is calculated based on the contribution value corresponding to at least one edge in the product chain.

[0035] According to embodiments of this disclosure, calculating the association strength value of a product chain based on the contribution value corresponding to at least one edge in the product chain includes:

[0036] Based on the target product of the product chain, determine the attenuation coefficient corresponding to at least one edge according to the number of hops between each edge and the node of the target product.

[0037] The correlation strength value of the product chain is calculated based on the contribution value corresponding to at least one edge and the attenuation coefficient corresponding to at least one edge.

[0038] According to embodiments of this disclosure, constructing multiple sets of query indexes corresponding to multiple product sets based on product feature information includes:

[0039] Based on the product's feature information, construct multiple sets of feature value ranges;

[0040] Establish mapping relationships between multiple sets of feature value ranges and multiple products to generate a query index.

[0041] A second aspect of this disclosure provides a map indexing construction apparatus, comprising:

[0042] The graph acquisition module is used to acquire product graphs and information on historical products held by users. Multiple nodes in the product graph represent multiple products, the edges between nodes represent the association types between products, and the attributes of the nodes represent the characteristic information of the products.

[0043] The product chain construction module is used to determine multiple associated nodes that are associated with historical products from multiple nodes based on the association type corresponding to the edge, and to construct a product chain for each associated product. Here, the associated node represents the associated product that is associated with the historical product.

[0044] The association strength calculation module is used to calculate the association strength value of a product chain based on the association type between multiple related products included in the product chain.

[0045] Multiple product sets are determined based on the correlation strength values ​​of each product chain.

[0046] The index building module is used to build multiple sets of query indexes corresponding to multiple product sets based on the feature information of related products.

[0047] A third aspect of this disclosure provides an electronic device comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method described above.

[0048] A fourth aspect of this disclosure also provides a computer-readable storage medium having a computer program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0049] The fifth aspect of this disclosure also provides a computer program product, including a computer program or instructions that, when executed by a processor, implement the steps of the above-described method.

[0050] According to embodiments of this disclosure, by constructing product chains and calculating association strength values, products with strong associations are grouped into the same product set. A query index is then built based on the dimensions of the product set, and the product index data for the same product set is stored in a single storage shard. Thus, the organization of the query index matches the association relationships between products. Product index data within the same product set is stored centrally. When querying similar products, the query can be completed within the same storage shard without cross-shard traversal, reducing network communication overhead and cross-shard index lookup latency, thereby improving query efficiency. Simultaneously, since the product set is determined based on the association strength value of the product chain, products with high association strength are grouped into the same product set, making the query results more reflective of the association characteristics between products. This improves the matching degree between the query results and user needs, allowing for the rapid retrieval of multiple strongly related products, facilitating subsequent product recommendations to users. Furthermore, when constructing the index, the user's product holdings information is taken into account. By building a chain of related products, products that are strongly related to the user's holdings are grouped into the same product set. The index information in the index group constructed in this way is information about the user's related products. In subsequent product recommendation scenarios, products related to the user can be quickly retrieved from the graph for analysis and recommendation, thereby improving retrieval efficiency. Attached Figure Description

[0051] The foregoing contents, as well as other objects, features, and advantages of this disclosure, will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0052] Figure 1 The illustrations depict application scenarios of the map index construction method, apparatus, device, medium, and program product according to embodiments of the present disclosure.

[0053] Figure 2 A flowchart illustrating a method for constructing a map index according to an embodiment of the present disclosure is shown schematically;

[0054] Figure 3 A schematic diagram of a product portfolio according to embodiments of the present disclosure is shown.

[0055] Figure 4 This diagram schematically illustrates a system architecture diagram to which the atlas index construction method can be applied according to embodiments of the present disclosure;

[0056] Figure 5 This schematic diagram illustrates a structural block diagram of a map index construction apparatus according to an embodiment of the present disclosure;

[0057] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing a map indexing construction method according to an embodiment of the present disclosure. Detailed Implementation

[0058] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0059] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0060] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0061] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0062] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0063] In the technical solution disclosed herein, the user information (including but not limited to user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, and displayed data) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with relevant laws, regulations, and standards, and necessary measures have been taken to ensure that they do not violate public order and good morals. Corresponding operation entry points are provided for users to choose to authorize or refuse.

[0064] In scenarios involving automated decision-making using personal information, the methods, devices, and systems provided in this disclosure all offer users corresponding entry points for choosing to agree to or reject the automated decision-making results. If the user chooses to reject, the process proceeds to the expert decision-making stage. Here, "automated decision-making" refers to the activity of automatically analyzing and evaluating an individual's behavioral habits, interests, or economic, health, and credit status through computer programs, and then making a decision. Here, "expert decision-making" refers to the activity of making decisions by personnel who specialize in a particular field, possess specialized experience, knowledge, and skills, and have reached a certain level of professional expertise.

[0065] Figure 1 The illustration schematically depicts application scenarios of the map index construction method, apparatus, device, medium, and program product according to embodiments of the present disclosure.

[0066] like Figure 1 As shown, application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0067] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0068] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0069] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0070] It should be noted that the map index construction method provided in this embodiment can generally be executed by server 105. Correspondingly, the map index construction apparatus provided in this embodiment can generally be located in server 105. The map index construction method provided in this embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the map index construction apparatus provided in this embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105.

[0071] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0072] The following will be based on Figure 1 The described scene, through Figures 2-4 The video processing method of the disclosed embodiments will be described in detail.

[0073] Figure 2 A flowchart illustrating a map index construction method according to an embodiment of the present disclosure is shown schematically.

[0074] like Figure 2 As shown, the map index construction method of this embodiment includes operations S201 to S205.

[0075] In operation S201, the product graph and information on the historical products held by the user are obtained. In the product graph, multiple nodes represent multiple products, the edges between nodes represent the association type between products, and the attributes of the nodes represent the characteristic information of the products.

[0076] In operation S202, based on the association type corresponding to the edge, multiple association nodes associated with the historical product are determined from multiple nodes, and a product chain is constructed for each associated product, where the association node represents the associated product associated with the historical product;

[0077] In operation S203, the association strength value of the product chain is calculated based on the association type between multiple related products included in the product chain.

[0078] In operation S204, multiple product sets are determined based on the correlation strength values ​​of each of the multiple product chains;

[0079] In operation S205, based on the feature information of related products, multiple sets of query indexes are constructed corresponding to multiple product sets.

[0080] The embodiments disclosed herein can be applied to scenarios such as product recommendation, asset allocation, and portfolio management in the fintech field. With the continuous growth in the number and variety of products in the financial market, complex relationships such as substitution, complementarity, and mutual exclusion may exist between different products. When faced with a large number of products, users often find it difficult to quickly locate the product that matches their needs.

[0081] In operation S201, the product graph and information on the historical products held by the user are obtained. In the product graph, multiple nodes represent multiple products, the edges between nodes represent the association type between products, and the attributes of the nodes represent the characteristic information of the products.

[0082] Specifically, a product graph is a data structure that organizes product data using a graph structure. A product graph includes multiple nodes, each corresponding to a product. Edges between nodes represent the type of association between two products. Node attributes represent the characteristic information of that product.

[0083] In one example, the association type can include at least one of the following: substitute product, complementary product, and mutually exclusive product. A substitute product means that two products can be substituted for each other in terms of function or purpose, such as two equity funds with similar investment styles. A complementary product means that two products can complement each other in terms of function or purpose, such as using an equity fund and a bond fund together to diversify risk. A mutually exclusive product means that two products are not suitable for simultaneous use in terms of function or purpose, such as two products with conflicting investment strategies.

[0084] In one example, product characteristics may include quantitative metrics such as Sharpe ratio, maximum drawdown, beta coefficient, volatility, and annualized return. Product characteristics may also include basic information such as product name, product type, issuing institution, and establishment date.

[0085] For example, a product graph includes multiple product nodes, such as product nodes P1, P2, P3, P4, P5, P6, and P7. Edges between product nodes represent the relationship types between products. For instance, the edge between product node P1 and product node P3 represents a substitute product, the edge between product node P1 and product node P7 represents a complementary product, and the edge between product node P4 and product node P6 represents a mutually exclusive product. Each product node has corresponding attributes that record the product's characteristic information, such as Sharpe ratio, maximum drawdown, and beta coefficient.

[0086] In one example, the product graph can be pre-built and stored in a preset storage space. When the graph index construction method needs to be executed, the product graph is retrieved from the preset storage space. The construction method of the product graph may include: extracting basic and feature information of the products from data sources such as product recruitment brochures, market data, and historical net asset values ​​of the products to generate product nodes; and generating edges between product nodes based on the relationships between products. The construction method of the product graph can be selected according to the specific application scenario, and the embodiments of this disclosure do not limit it.

[0087] In one example, information about a user's historical holdings can be obtained from a user account database, a transaction history database, or a position registration system. This information may include the identifier of the product the user currently holds or has previously held, the product type, the percentage held, the holding duration, and buy / sell records.

[0088] In one example, retrieving information about the product portfolio and the user's historical product holdings could include: in response to receiving a financial planning request from a user, reading the product portfolio from a pre-set storage space and retrieving information about the user's historical product holdings from the user's account database. In another example, the retrieval operation could also be triggered periodically by the system, for example, by retrieving the product portfolio and information about the user's historical product holdings from the pre-set storage space and the user's account database at regular intervals each day.

[0089] In operation S202, based on the association type corresponding to the edge, multiple associated nodes that are associated with the historical product are determined from multiple nodes, and a product chain is constructed for each associated product, where the associated node represents the associated product associated with the historical product.

[0090] The association types between products include one of the following: substitute products, complementary products, and mutually exclusive products. Based on the association type corresponding to the edges, determining multiple associated nodes from multiple nodes that are related to historical products can include: product nodes that have a substitute or complementary relationship with historical products as associated nodes, while excluding product nodes that have a mutually exclusive relationship or no association with historical products. In this way, based on the association type corresponding to the edges, multiple associated nodes related to historical products are determined from multiple nodes, and a product chain is constructed for each associated product. Since associated nodes are product nodes that have a substitute and / or complementary relationship with historical products, products in the product chain have substitute and / or complementary relationships, but do not contain mutually exclusive relationships. Therefore, the product chain can accurately reflect the combinable or substitutable association characteristics with the historical products held by the user, providing a data foundation for subsequently calculating association strength values ​​and determining product sets based on the product chain, enabling the subsequently constructed query index to be organized according to the association characteristics related to the user.

[0091] After determining the associated nodes, product chains can be constructed for each product based on the association type corresponding to the edges. This includes: traversing the product graph starting from each associated product node; expanding along the edges with association types of substitute products and / or complementary products during the traversal to obtain the product chain; terminating the product chain when traversing the edges with association types of mutually exclusive products, so that at least one product chain is obtained for each product.

[0092] Figure 3 A schematic diagram of a product portfolio according to embodiments of the present disclosure is shown. The following is in conjunction with... Figure 3 The method for constructing a product chain for each product according to embodiments of this disclosure will be described.

[0093] like Figure 3The image shown is a portion of the product graph, displaying nodes for product 1, 2, 3, 4, 5, 6, and 7. The edges between product nodes represent the relationship types between them. Specifically, the edge between product 1 and product 2 corresponds to a mutually exclusive relationship; the edge between product 1 and product 3 corresponds to a substitute relationship; the edge between product 2 and product 5 corresponds to a substitute relationship; the edge between product 3 and product 4 corresponds to a substitute relationship; the edge between product 3 and product 6 corresponds to a mutually exclusive relationship; and the edge between product 3 and product 7 corresponds to a complementary relationship.

[0094] In one example, assume product 1 is a known associated product among the user's historical products. Starting with product 1, traverse the product graph. The edge between product 1 and product 2 corresponds to a mutually exclusive product association; traversing this edge terminates the product chain expansion in that direction, and no product chain is formed between product 1 and product 2. The edge between product 1 and product 3 corresponds to a substitute product association; extending along this edge to product 3 yields a path from product 1 to product 3. Continuing traversal with product 3 as the current node, the edge between product 3 and product 4 corresponds to a substitute product association; extending along this edge to product 4 yields a path from product 1 to product 3 to product 4. The edge between product 3 and product 7 corresponds to a complementary product association; extending along this edge to product 7 yields a path from product 1 to product 3 to product 7. The edge between product 3 and product 6 corresponds to a mutually exclusive product association; traversing this edge terminates the product chain expansion in that direction, and no further product chain is formed between product 3 and product 6. Thus, starting with product 1, we obtain the product chain from product 1 to product 3 to product 4, and the product chain from product 1 to product 3 to product 7.

[0095] In one example, assume product 3 is a known associated product with a user's historical products. Starting with product 3, traverse the product graph. The edge between product 3 and product 1 corresponds to a substitute product association; extending along this edge to product 1 yields a path from product 3 to product 1. Continuing traversal with product 1 as the current node, the edge between product 1 and product 2 corresponds to a mutually exclusive product association; traversing to this edge terminates the product chain extension in that direction. The edge between product 3 and product 4 corresponds to a substitute product association; extending along this edge to product 4 yields a path from product 3 to product 4. The edge between product 3 and product 7 corresponds to a complementary product association; extending along this edge to product 7 yields a path from product 3 to product 7. The edge between product 3 and product 6 corresponds to a mutually exclusive product association; traversing to this edge terminates the product chain extension in that direction. Thus, starting with product 3, we obtain product chains from product 3 to product 1, product chains from product 3 to product 4, and product chains from product 3 to product 7.

[0096] In one example, suppose product 2 is a known associated product with a user's historical products. Starting with product 2, traverse product graph 300. The edge between product 2 and product 1 corresponds to a mutually exclusive product association type; traversing this edge terminates the product chain extension in that direction. The edge between product 2 and product 5 corresponds to a substitute product association type; extending along this edge to product 5 yields a path from product 2 to product 5. Thus, starting with product 2, we obtain the product chain from product 2 to product 5.

[0097] In one example, the length of the product chain can be limited to a preset number of nodes during traversal, such as no more than 5 product nodes. During traversal, the same product node appears only once in the same product chain. When a visited product node is encountered, the product chain expansion in that direction stops.

[0098] In one example, for each associated product node, at least one product chain can be obtained. Product chains corresponding to different product nodes can overlap. For example, starting with product 1, a product chain from product 1 to product 3 to product 4 can be obtained, and starting with product 3, a product chain from product 3 to product 1 can be obtained. The two product chains contain the same product nodes, product 1 and product 3. However, the association strength values ​​of these two product chains may be different.

[0099] By traversing the product graph starting from each product node, the process expands along edges indicating substitutive and / or complementary product associations. The product chain terminates when an edge indicating mutually exclusive product associations is encountered. This ensures that the resulting product chain contains products with substitutive and / or complementary relationships, but no mutually exclusive relationships. Therefore, the product chain accurately reflects the composable or substitutable association characteristics between products, providing a data foundation for calculating association strength values ​​and determining product sets based on the product chain. This allows the subsequently constructed query index to be organized according to the association characteristics between products.

[0100] In operation S203, the association strength value of the product chain is calculated based on the association type between multiple related products included in the product chain.

[0101] Specifically, each edge in the product chain has a corresponding association type. Different association types contribute differently to the association strength value of the product chain. Based on the association type of each edge in the product chain, the contribution value of that edge is determined, and then the association strength value of the product chain is calculated based on the contribution values ​​of each edge in the product chain.

[0102] In one example, a corresponding contribution value can be preset for each association type. For instance, the contribution value for a substitute product can be greater than that for a complementary product, and the contribution value for a mutually exclusive product can be negative. In a specific example, the contribution value for a substitute product can be 0.5, the contribution value for a complementary product can be 0.3, and the contribution value for a mutually exclusive product can be -0.4. Furthermore, the association strength value of the entire product chain can be calculated cumulatively based on the contribution values ​​of multiple edges included in the product chain.

[0103] In operation S204, multiple product sets are determined based on the correlation strength values ​​of each of the multiple product chains.

[0104] Specifically, each associated product corresponds to at least one product chain. For each associated product, the product chain with the strongest association strength can be selected based on the association strength values ​​of these product chains and used as the product chain corresponding to each associated product.

[0105] Furthermore, the products in the product chain corresponding to each associated product are treated as a product set, resulting in multiple product sets. Then, based on the association strength values ​​of each product chain, the product chains are assigned to different product sets. A product set is a collection of product nodes, and product nodes within the same product set have a high association strength.

[0106] In operation S205, based on the feature information of related products, multiple sets of query indexes are constructed corresponding to multiple product sets. Data in the same set of query indexes are stored in the same storage shard.

[0107] Specifically, for each product set, the feature information of each product within that set is extracted. Based on this feature information, a corresponding query index is constructed for that product set. The query indexes for different product sets are independent of each other. Data in the same set of query indexes is stored in the same storage shard.

[0108] In one example, the query index may include a feature-value inverted index to record the mapping between feature value ranges and product identifiers. The query index may also include a feature range index to support product retrieval by feature value range. Furthermore, the query index may include an association edge index to record the mapping between product identifiers and associated product identifiers.

[0109] In one example, based on the feature information of related products, multiple sets of feature value ranges can be constructed, and mapping relationships between these ranges and multiple products can be established to generate a query index. For example, for the feature information of Sharpe ratio, feature value ranges "greater than 2.0", "1.5 to 2.0", "1.0 to 1.5", "0.5 to 1.0", and "less than 0.5" can be constructed, and a mapping relationship can be established between each feature value range and the product identifier that falls within that range.

[0110] In one example, product set A corresponds to query index A, and product set B corresponds to query index B. Data in query index A is stored in storage shard 0, and data in query index B is stored in storage shard 1. Index entries corresponding to product nodes within the same product set are arranged adjacently in physical storage.

[0111] In one example, a corresponding storage shard is allocated for each product set. Product nodes and their query indexes within the same product set are stored in the same storage shard. Therefore, queries targeting that product set can be completed within the same storage shard, reducing cross-shard query operations.

[0112] In related technologies, graph databases typically store index information based on product identifiers, without considering the relationships between products. When querying a product from the graph, it can only be retrieved one product at a time, resulting in high query latency. Furthermore, if the information of the product and its related products is distributed across different physical shards, the query requires traversing across shards, incurring network communication overhead and cross-shard index lookup latency.

[0113] This disclosure provides a graph index construction method. It involves constructing a product graph, building product chains along association types, calculating the association strength values ​​of the product chains, determining a product set, and building a query index based on the product set. This ensures that the organization of the query index matches the association relationships between products, thereby improving the efficiency and accuracy of product queries. This graph index construction method can be applied to financial planning systems, product recommendation systems, asset allocation systems, and other scenarios requiring index construction and querying of related products.

[0114] The above method obtains a product graph, constructs product chains based on the association types corresponding to edges, calculates association strength values ​​based on the association types between products in the product chains, determines product sets based on the association strength values, and constructs query indexes corresponding to the product sets based on product feature information. Data in the same set of query indexes is stored in the same storage shard. Therefore, the embodiments of this disclosure, by constructing product chains and calculating association strength values, group products with strong associations into the same product set, constructs query indexes based on the dimensions of the product sets, and stores the product index data of the same product set in a single storage shard.

[0115] Therefore, the organization of the query index matches the relationships between products. Product indexes within the same product set are stored centrally. When querying similar products, the query can be completed within the same storage shard, eliminating the need for cross-shard traversal, reducing network communication overhead and cross-shard index lookup latency, and improving query efficiency. Simultaneously, since product sets are determined based on the correlation strength of product chains, products with higher correlation strengths are grouped into the same product set. This makes the query results more reflective of the correlation characteristics between products, improving the matching degree between query results and user needs. Multiple highly correlated products can be quickly retrieved, facilitating subsequent product recommendations to users. When constructing the index, user-held product information is considered. By building related product chains, products strongly correlated with the user's held products are grouped into the same product set. The index information in this constructed index group is information about products related to the user. In subsequent product recommendation scenarios, products related to the user can be quickly retrieved from the graph for analysis and recommendation, improving retrieval efficiency.

[0116] Furthermore, determining multiple product sets based on the association strength values ​​of each product chain includes: grouping product chains with association strength values ​​greater than or equal to a preset threshold into the same product set; merging the product sets corresponding to the two product chains if two product chains contain the same product nodes; and grouping the two product chains into different product sets if two product chains do not contain the same product nodes.

[0117] According to embodiments of this disclosure, the association type includes one of substitute products, complementary products, and mutually exclusive products.

[0118] Based on the association type corresponding to the edge, multiple associated nodes that are associated with the historical product are identified from multiple nodes, including: the first type of node representing the historical product, and the second type of node whose association type with the first type of node belongs to the substitute product and / or complementary product, are identified as multiple associated nodes.

[0119] Specifically, nodes in the product graph can be divided into two categories: the first category represents historical products held by the user, and the second category represents products that are associated with the first category of nodes. Edges between nodes represent the association type between products, which can be one of the following: substitute products, complementary products, or mutually exclusive products. When identifying multiple associated nodes related to historical products from a pool of nodes, only the second category of nodes, whose association type is substitute products and / or complementary products, are identified as associated nodes, while the second category of nodes whose association type is mutually exclusive products are excluded.

[0120] In one example, consider a product graph including product nodes: Product 1, Product 2, Product 3, Product 4, Product 5, Product 6, and Product 7. The user's historical products include Product 1 and Product 3. The edge between Product 1 and Product 2 corresponds to a mutually exclusive product association, while the edge between Product 1 and Product 3 corresponds to a substitute product association. Since Product 1 is a historical product, its corresponding node is a first-type node. The association type between Product 2 and Product 1 is mutually exclusive, which does not meet the condition of being a substitute product and / or complementary product; therefore, Product 2 is not identified as an associated node. The association type between Product 3 and Product 1 is a substitute product association, which meets the condition of being a substitute product and / or complementary product; therefore, Product 3 is identified as an associated node.

[0121] In one example, if there are multiple edges between a first-type node and a second-type node, and the association types corresponding to these edges are different, the second-type node is determined as an associated node based on the association type with the highest priority. For example, if there are two edges between a first-type node and a second-type node, one edge corresponding to the association type of a substitute product, and the other edge corresponding to the association type of a mutually exclusive product, then the second-type node is determined to be an associated node because substitute products have a higher priority than mutually exclusive products, and substitute products fall within the scope of substitute products and / or complementary products. If one edge corresponds to the association type of a complementary product, and the other edge corresponds to the association type of a mutually exclusive product, then the second-type node is determined to be an associated node because complementary products fall within the scope of substitute products and / or complementary products. If multiple edges correspond to the association type of mutually exclusive products, then the second-type node is not determined to be an associated node.

[0122] Using the above method, the first type of nodes representing historical products, and the second type of nodes whose association type with the first type of nodes belongs to substitute products and / or complementary products, are identified as multiple associated nodes. Since mutually exclusive products are excluded from associated nodes, the subsequent product chain does not include products mutually exclusive with historical products, avoiding conflicts between recommendation results and the user's current holdings. Because substitute and complementary products are included in associated nodes, the subsequent product chain can cover products that are substitutable or combinable with historical products. This provides a data foundation for calculating association strength values ​​and determining product sets based on product chains, ensuring that the constructed query index accurately reflects the range of products associated with the user's historical products.

[0123] According to embodiments of this disclosure, the method further includes: storing multiple sets of user query indexes in a first storage shard; constructing a full query index corresponding to all products based on the full product map; and storing the full query index in a second storage shard.

[0124] It also includes: responding to query requests by first querying the user-level index, and then querying the global index when the number of results retrieved from the user-level index is lower than a preset threshold.

[0125] Specifically, the user query index is an index of relevant products for a user, built based on information about the user's historical products. The full query index is an index built based on all products corresponding to the product graph. The user query index and the full query index are stored in different storage shards. The first storage shard stores the user query index, and the second storage shard stores the full query index.

[0126] In one example, there can be multiple sets of user query indexes. For instance, product set A corresponds to query index A, and product set B corresponds to query index B. Both query index A and query index B are user query indexes and are stored in the first storage shard. The full query index is built based on all product nodes in the product graph and is stored in the second storage shard. The first and second storage shards can be different physical storage devices or different storage regions partitioned within the same physical storage device.

[0127] In one example, in response to a received query request, the user query index is queried first. Specifically, based on the expected product feature value in the query request, a target user query index matching the expected product feature value is determined. The target user query index is then searched to obtain candidate products matching the expected product feature value. If the number of candidate products is greater than or equal to a preset threshold, these candidate products are output as query results, and the full query index is not queried. If the number of candidate products is less than the preset threshold, the full query index is queried further.

[0128] In one example, the preset threshold can be set according to the query precision requirements. For example, the preset threshold can be set to 3. If the number of candidate products retrieved from the user query index is 3 or more, the query requirement is met, and there is no need to query the entire query index. If the number of candidate products retrieved from the user query index is 2, which is less than the preset threshold of 3, then the entire query index is queried to supplement the number of candidate products.

[0129] In one example, when querying the full query index, the query conditions can be determined based on the desired product feature values. Supplementary candidate products matching the desired product feature values ​​are then retrieved from the full query index. The candidate products obtained from the user query index are then merged with the supplementary candidate products obtained from the full query index to obtain the final candidate products.

[0130] In one example, the data in the user query index has a strong correlation with the user's historical products. Since the user query index is built specifically for the user's relevant products, the hit rate is high, and in most cases, it's not necessary to query the entire query index to obtain candidate products that meet the query requirements. Therefore, the query operation is preferentially completed within the first storage shard, reducing cross-shard query operations, lowering network communication overhead, and reducing query latency. When the number of retrieved results from the user query index is insufficient, the full query index in the second storage shard is then queried to supplement candidate products, ensuring the completeness of the query results.

[0131] By employing the above method, multiple sets of user query indexes are stored in the first storage shard, while the full query index is stored in the second storage shard. In response to query requests, the user query index is queried first. Only when the number of results retrieved from the user query index falls below a preset threshold is the full query index queried. Since the user query index is built specifically for the user's relevant products, the query hit rate is high, and most queries can be completed within the first storage shard, eliminating the need for cross-shard traversal. This reduces network communication overhead and cross-shard index lookup latency, improving query efficiency. Furthermore, when the user query index returns insufficient results, the full query index is queried to supplement candidate products, ensuring the completeness and accuracy of the query results.

[0132] According to embodiments of this disclosure, constructing multiple sets of query indexes corresponding to multiple product sets based on product feature information includes: constructing multiple sets of feature value ranges based on product feature information; establishing a mapping relationship between the multiple sets of feature value ranges and multiple products; and generating query indexes.

[0133] Specifically, product characteristic information can include quantitative indicators such as Sharpe ratio, maximum drawdown, beta coefficient, volatility, and annualized return. For each product set, the characteristic information of each product within that set is extracted. Based on this characteristic information, multiple sets of characteristic value intervals are constructed. Each set of characteristic value intervals corresponds to a characteristic dimension. A mapping relationship is established between the multiple sets of characteristic value intervals and multiple products, generating a query index.

[0134] In one example, let's consider the Sharpe ratio as a feature. Product set A includes product 1, product 3, and product 7. Product 1 has a Sharpe ratio of 1.20, product 3 has a Sharpe ratio of 1.15, and product 7 has a Sharpe ratio of 0.80. Based on the Sharpe ratio, we construct multiple sets of feature value ranges, such as "greater than 2.0", "1.5 to 2.0", "1.0 to 1.5", "0.5 to 1.0", and "less than 0.5". We then establish a mapping relationship between these feature value ranges and the products. Product 1's Sharpe ratio of 1.20 falls into the "1.0 to 1.5" range, product 3's Sharpe ratio of 1.15 falls into the "1.0 to 1.5" range, and product 7's Sharpe ratio of 0.80 falls into the "0.5 to 1.0" range. Therefore, we generate a query index corresponding to the Sharpe ratio, where the "1.0 to 1.5" range corresponds to product 1 and product 3, and the "0.5 to 1.0" range corresponds to product 7.

[0135] In one example, the method for dividing the feature value interval can be determined based on the distribution of feature information of products within the product set. For instance, if the Sharpe ratio of products within the product set is concentrated between 0.5 and 1.5, the feature value interval within this range can be appropriately refined, such as dividing it into "0.5 to 0.8", "0.8 to 1.0", "1.0 to 1.2", and "1.2 to 1.5" to improve query accuracy. The boundaries of the feature value interval can be set as left-closed and right-open intervals, left-open and right-closed intervals, or closed intervals, depending on the application scenario.

[0136] By employing the above method, multiple sets of feature value ranges are constructed based on the product's characteristic information. A mapping relationship is established between these ranges and multiple products, generating a query index. Therefore, during a query, products falling within a feature value range can be quickly located without comparing feature values ​​for each product individually, reducing the number of comparison operations required for the query and improving query efficiency.

[0137] According to embodiments of this disclosure, the above-mentioned graph index construction method further includes: receiving user transaction data originating from the data acquisition layer, and information on historical products held by the user; determining desired product feature values ​​based on the user transaction data and information on historical products held by the user; querying at least one candidate product matching the desired product feature value from multiple products included in the product graph based on the desired product feature value and multiple sets of query indexes; and determining the product to be recommended based on the association type between the at least one candidate product and the historical products held by the user.

[0138] Specifically, the data acquisition layer can be used to collect user transaction data and information about the historical products held by users. User transaction data may include user income and expenditure records, consumption records, transfer records, etc. Information about the historical products held by users may include the identifiers, product types, holding ratios, holding durations, buy and sell records, etc. of the products currently held or previously held by users. The data acquisition layer provides the collected user transaction data and information about the historical products held by users to the business decision-making layer or the intelligent computing layer.

[0139] In one example, user transaction data can be processed by a dynamic profiling engine in the data acquisition layer. The dynamic profiling engine updates the user feature vector incrementally based on the user transaction data, outputting a user state vector. The user state vector can include the user's income and expenditure characteristics, savings rate, consumption preferences, etc. Information about the user's historical holdings can be used to determine the user's historical portfolio characteristics, such as the type of holding products, the proportion of holdings, and the duration of holdings.

[0140] In one example, expected product characteristics are determined based on user transaction data and information about the user's historical product holdings. Expected product characteristics represent the features the user expects the recommended product to possess. These characteristics may include at least one of risk tolerance, expected returns, and liquidity requirements.

[0141] In one example, the user's income and expenditure characteristics are determined based on their transaction data. For instance, a user's monthly income is 35,000 yuan, monthly expenditure is 15,000 yuan, and savings rate is 57%. The user's historical holdings are determined based on information about their historical product holdings. For instance, the user currently holds Product 1 and Product 3, with holding ratios of 60% and 40% respectively, and a holding period of 12 months for each. Based on the income and expenditure characteristics and historical holdings, expected product characteristic values ​​are determined. For instance, based on a high savings rate and long holding period, the user's expected return is determined to be above average; based on a high proportion of equity products in their historical holdings, the user's risk tolerance is determined to be medium to high; and based on stable monthly expenditures, the user's liquidity needs are determined to be low.

[0142] In one example, based on desired product feature values ​​and multiple sets of query indexes, at least one candidate product matching the desired product feature values ​​is retrieved from multiple products included in the product graph. Specifically, query conditions are determined based on the desired product feature values. Query conditions may include feature value ranges and / or feature value intervals. Based on the query conditions, a target query index matching the query conditions is determined from the multiple sets of query indexes. A search is performed in the target query index to obtain at least one candidate product matching the desired product feature values.

[0143] In one example, the expected product characteristics include medium-high risk tolerance, medium-to-high expected return, and low liquidity requirement. Query criteria are determined based on these characteristics, such as a Sharpe ratio greater than 1.0, maximum drawdown less than 20%, and a beta coefficient between 0.8 and 1.2. Based on these criteria, target query indexes matching the query conditions are determined. For example, if the query conditions involve Sharpe ratio, maximum drawdown, and beta coefficient, then indexes containing these characteristics are searched among multiple sets of query indexes, such as query index A corresponding to product set A and query index B corresponding to product set B. The target query index is then used to retrieve products that meet the criteria of a Sharpe ratio greater than 1.0, maximum drawdown less than 20%, and a beta coefficient between 0.8 and 1.2, which are then considered as candidate products. For example, candidate products include products 4, 7, and 8.

[0144] According to embodiments of this disclosure, determining the product to be recommended based on the association type between at least one candidate product and the user's historical products includes: deleting candidate products with mutually exclusive association types from the at least one candidate product list; determining the priority of candidate products with alternative association types as a first priority; and determining the priority of candidate products with complementary association types as a second priority, wherein the first priority is higher than the second priority. The product to be recommended is then determined based on the candidate products corresponding to the first and second priorities.

[0145] Specifically, the association type between candidate products and a user's historical products can include at least one of the following: substitute products, complementary products, and mutually exclusive products. Different association types have different impacts on the recommendation results. Substitute products indicate that the candidate product and the user's historical products are interchangeable in function or use, and the user can replace their historical products with the candidate product. Complementary products indicate that the candidate product and the user's historical products complement each other in function or use, and the user can combine the candidate product with their historical products. Mutually exclusive products indicate that the candidate product and the user's historical products are not suitable for simultaneous use in terms of function or use; recommending mutually exclusive products may lead to conflicts in the user's portfolio.

[0146] In one example, candidate products with a mutually exclusive association type are removed from at least one candidate product list. Specifically, the association type between each candidate product and the user's historical products is determined. If a candidate product has a mutually exclusive association type with any of the user's historical products, then that candidate product is removed from at least one candidate product list. For example, candidate products include products 4, 7, and 8, and the user's historical products include products 1 and 3. If product 8 and product 3 have a mutually exclusive association type, then product 8 is removed from the candidate products list. After removal, the candidate products include products 4 and 7.

[0147] In one example, candidate products whose association type is "substitute product" are given the highest priority. Specifically, if a candidate product's association type with a user's historical products is "substitute product," then that candidate product is assigned the highest priority. For example, if product 4's association type with product 3 is "substitute product," then product 4 is assigned the highest priority.

[0148] In one example, candidate products with a complementary product association type are prioritized as the second priority. Specifically, if a candidate product has a complementary product association type with a user's historical products, then that candidate product is assigned the second priority. For example, if product 7 has a complementary product association type with product 3, then product 7 is assigned the second priority. The first priority is higher than the second priority.

[0149] In one example, the proposed recommended products are determined based on candidate products corresponding to first and second priorities. Specifically, candidate products assigned first priority and candidate products assigned second priority are selected as the proposed recommended products. When outputting the proposed recommended products, they can be arranged in priority order, with candidate products of first priority placed before candidate products of second priority. For example, if the proposed recommended products include product 4 and product 7, product 4 is placed before product 7.

[0150] In one example, if multiple candidate products are assigned first priority, they can be further sorted based on the association strength value of the product chains to which they belong. Candidate products in product chains with higher association strength values ​​are ranked higher. Similarly, if multiple candidate products are assigned second priority, they can also be further sorted based on their association strength values.

[0151] In one example, if a candidate product has different association types with multiple historical products held by the user, the priority of the candidate product can be determined according to the association type with the highest priority. For example, if the association type between the candidate product and the user's historical product 1 is a substitute product, and the association type between the candidate product and the user's historical product 3 is a complementary product, then the priority of the candidate product is determined according to the first priority corresponding to the substitute product.

[0152] By employing the above method, candidate products with mutually exclusive association types are removed from at least one candidate product list. Candidate products with substitute association types are prioritized as first priority, and candidate products with complementary association types are prioritized as second priority. The recommended product is then determined based on the candidate products corresponding to the first and second priorities. Because mutually exclusive products are removed, conflicts between the recommended results and the user's current holdings are avoided. Since substitute products are given higher priority than complementary products, the recommended results prioritize meeting the user's need to replace their current holdings, and secondly, their need to supplement their current portfolio. Therefore, the association characteristics between the recommended results and the user's current holdings are matched, improving the degree of matching between the recommended results and the user's needs.

[0153] Figure 4 The diagram illustrates a system architecture diagram to which the atlas index construction method can be applied according to an embodiment of the present disclosure.

[0154] like Figure 4 As shown, the system architecture includes a data acquisition layer, an intelligent computing layer, a business decision-making layer, and an execution monitoring layer. The data acquisition layer is connected to the intelligent computing layer, which in turn is connected to both the business decision-making layer and the execution monitoring layer. The execution monitoring layer is also connected to the business decision-making layer, forming a feedback loop.

[0155] In one example, the data acquisition layer collects product and user dynamic information and provides it to the intelligent computing layer. This product and user dynamic information may include market data, product net asset value data, updated product prospectus data, market event data, user transaction data, and information on users' historical product holdings. The data acquisition layer can normalize, time-slice, and incrementally window-identify the collected data before providing it to the intelligent computing layer.

[0156] In one example, the intelligent computing layer receives product and user dynamic information from the data acquisition layer and performs product queries and recommendations based on this information. The intelligent computing layer interacts with the business decision-making layer using graph data. The intelligent computing layer obtains graph data from the business decision-making layer, which includes feature information of product nodes in the product graph, the types of associations between product nodes, and query indexes. Based on the graph data and product and user dynamic information, the intelligent computing layer determines desired product feature values ​​and, based on these desired product feature values ​​and the query indexes, queries multiple products included in the product graph to obtain candidate products, further determining the products to be recommended.

[0157] In one example, the business decision layer is used to perform product graph creation and dynamic updates. The business decision layer obtains the product graph, constructs product chains for each product based on the association types corresponding to edges, calculates the association strength value of the product chains based on the association types between multiple products included in the product chains, determines multiple product sets based on the association strength values ​​of each product chain, and constructs multiple sets of query indexes corresponding to the multiple product sets based on the product feature information. Data in the same set of query indexes is stored in the same storage shard. The business decision layer is also used to receive product update information, determine the products to be updated based on the product update information, determine the sets of products to be updated corresponding to the products to be updated, and update the query indexes corresponding to the sets of products to be updated based on the product update information.

[0158] In one example, the intelligent computing layer provides recommended product information to the execution monitoring layer. This information includes the identifier of the product to be recommended, its features, and its priority. The execution monitoring layer performs corresponding operations based on the recommended product information and sends the results back to the business decision-making layer through a feedback loop. The feedback data from the execution monitoring layer may include the execution status of the recommended products and user feedback. The business decision-making layer then dynamically updates the product map based on this feedback data.

[0159] In one example, the data acquisition layer collects user transaction data and information about the user's historical products, and provides this information to the intelligent computing layer. The intelligent computing layer determines desired product feature values ​​based on these features and, using the query index obtained from the business decision layer, retrieves candidate products from multiple products included in the product graph. Based on the association type between the candidate products and the user's historical products, it determines the products to be recommended and provides this information to the execution monitoring layer. The execution monitoring layer performs corresponding operations based on the recommended product information and sends the execution results back to the business decision layer through a feedback loop.

[0160] Through the above system architecture, a closed loop of data interaction and feedback is formed between the data acquisition layer, intelligent computing layer, business decision-making layer, and execution monitoring layer. The data acquisition layer is responsible for collecting and preprocessing dynamic product and user information; the intelligent computing layer is responsible for executing product queries and recommendations based on graph data and dynamic information; the business decision-making layer is responsible for creating and dynamically updating the product graph; and the execution monitoring layer is responsible for executing and providing feedback on recommended products. Because the business decision-making layer stores query indexes for the same product set in the same storage shard, the intelligent computing layer can complete product queries within the same storage shard when performing queries based on the query index, reducing cross-shard query operations. Since the execution monitoring layer sends execution results back to the business decision-making layer through the feedback loop, the business decision-making layer can dynamically update the product graph based on the feedback data, ensuring that the product graph and query indexes can promptly reflect changes in product information and user feedback.

[0161] According to embodiments of this disclosure, the method further includes: receiving product update information from the data acquisition layer, and determining at least one product to be updated based on the product update information; determining at least one set of products to be updated corresponding to the at least one product to be updated; and updating the query index corresponding to the at least one set of products to be updated based on the product update information.

[0162] Specifically, the data acquisition layer can be used to collect product-related updated data, such as market data, product net asset value data, updated product prospectus data, and market event data. The data acquisition layer then provides the collected data to the business decision-making layer as product update information. This product update information may include the identifier of the changed product, the characteristics of the change, and the time of the change.

[0163] In one example, product update information may include a product identifier and the changed characteristics of the product. Based on the product identifier in the product update information, at least one product to be updated is identified. For example, if the product update information indicates that the Sharpe ratio of product 1 has changed from 1.20 to 0.80, then product 1 is identified as the product to be updated. As another example, if the product update information indicates that the maximum drawdown of product 3 has changed from 18% to 22%, then product 3 is identified as the product to be updated.

[0164] In one example, the product set to which the product to be updated belongs is determined based on its product identifier. Since the query index is built along the dimension of product sets, and each product set corresponds to a set of query indexes, determining the set of products to be updated for the product to be updated determines the scope of the query index that needs to be updated.

[0165] In one example, the product to be updated is product 1. Based on the product set determined in the aforementioned operation S204, product 1 belongs to product set A. Product set A corresponds to query index A, and the data in query index A is stored in storage shard 0. Therefore, the product set to be updated is determined to be product set A, the query index to be updated is query index A, and the storage shard to be updated is storage shard 0.

[0166] In one example, the product update information indicates that multiple products have changed, such as both product 1 and product 3. If product 1 belongs to product set A and product 3 belongs to product set A, then the product set to be updated is determined to be product set A. If product 1 belongs to product set A and product 3 belongs to product set B, then the product set to be updated includes both product set A and product set B.

[0167] In one example, the query index corresponding to the set of products to be updated is updated based on product update information. Specifically, based on the changed feature information of the products to be updated in the product update information, the index entries for those products are updated in the query index corresponding to the set of products to be updated.

[0168] In one example, let's take product 1 as the product to be updated and product set A as the set of products to be updated. The Sharpe ratio of product 1 changes from 1.20 to 0.80. In related technologies, the Sharpe ratio of product 1 of 1.20 falls within the characteristic value range of "1.0 to 1.5", while the changed Sharpe ratio of 0.80 falls within the characteristic value range of "0.5 to 1.0". When updating query index A, the identifier of product 1 is deleted from the product identifier list corresponding to the characteristic value range of "1.0 to 1.5", and the identifier of product 1 is inserted into the product identifier list corresponding to the characteristic value range of "0.5 to 1.0".

[0169] In one example, when query index A is updated, the index entries for other products in query index A remain unchanged. For instance, product set A also includes products 3 and 7. Since the feature information for products 3 and 7 has not changed, the index entries for products 3 and 7 in query index A remain unchanged. Simultaneously, the query indexes corresponding to other product sets, such as query index B corresponding to product set B, remain unchanged. Only query index A is updated in storage shard 0; query index B in storage shard 1 is not updated.

[0170] In one example, when the number of changes indicated by product update information exceeds a preset threshold, the current time window can be marked as a high-increment window. Under a high-increment window, in addition to updating the query index corresponding to the current set of products to be updated, the query index corresponding to adjacent time windows can also be prefetched into memory to improve the response speed of subsequent queries.

[0171] In one example, multiple index versions can be maintained within each storage shard, organized by time window. When product update information is received, the target time window corresponding to the update information is determined. Within the index versions corresponding to the target time window, the index entries for the product to be updated are then updated. Index versions corresponding to historical time windows remain unchanged and can be used for cross-time queries or rollbacks.

[0172] The above method receives product update information from the data acquisition layer, identifies at least one product to be updated based on this information, determines at least one set of products to be updated corresponding to this product, and updates the query index corresponding to this set of products based on the product update information. Since the query index is grouped and stored according to the dimension of product sets, and the query indexes for the same product set are stored in the same storage shard, when product information changes and index information needs to be updated, only the query index corresponding to the product set to which the product to be updated belongs needs to be updated, without updating the entire query index. Therefore, the scope of index updates is narrowed from the entire query index to the query index corresponding to the product set to be updated, reducing the amount of data involved in index updates and lowering the computational resources and time overhead required for index updates. Simultaneously, since the query indexes corresponding to different product sets are stored independently, updating the index of one product set does not affect the query operations of other product sets, improving the stability of the query service during the index update process.

[0173] According to an embodiment of this disclosure, in one implementation, calculating the association strength value of a product chain based on the association type among multiple associated products included in the product chain includes: determining the contribution value corresponding to at least one edge in the product chain based on the association type among multiple associated products included in the product chain; and calculating the association strength value of the product chain based on the contribution value corresponding to at least one edge in the product chain.

[0174] Specifically, each edge in the product chain has a corresponding association type. Different association types contribute differently to the association strength value of the product chain. A corresponding contribution value can be preset for each association type, and the contribution value of the edge is determined based on the association type of the edge in the product chain.

[0175] In one example, the correspondence between association type and contribution value can be set as follows: the contribution value for a substitute product is 0.5, the contribution value for a complementary product is 0.3, and the contribution value for a mutually exclusive product is -0.4. Since the product chain expands along the edges of substitute and / or complementary products during construction and terminates when it reaches the edge of a mutually exclusive product, the edges in the product chain are typically substitute or complementary products of the associated products. Taking the product chain from product 1 to product 3 to product 4 as an example, the association type corresponding to the edge between product 1 and product 3 is substitute, with a contribution value of 0.5; the association type corresponding to the edge between product 3 and product 4 is substitute, with a contribution value of 0.5.

[0176] In one implementation, calculating the association strength value of a product chain based on the contribution value corresponding to at least one edge in the product chain includes: determining the attenuation coefficient corresponding to at least one edge based on the target product targeted by the product chain and the number of hops between each edge and the node of the target product; and calculating the association strength value of the product chain based on the contribution value corresponding to at least one edge and the attenuation coefficient corresponding to at least one edge.

[0177] Specifically, the target product in a product chain is the starting product of that chain. Each edge in the product chain has a hop count between it and the node representing the target product. The hop count indicates the distance between that edge and the target product. The farther an edge is from the target product, the smaller its attenuation coefficient. The attenuation coefficient corresponding to that edge is determined based on the hop count between it and the target product.

[0178] In one example, the attenuation factor can be determined as follows: the attenuation factor is equal to the preset attenuation factor raised to the power of the hop count. The preset attenuation factor is a value greater than 0 and less than 1, such as 0.9. An edge directly connected to the target product has 1 hop count and an attenuation factor of 0.9 to the power of 1, which is 0.9. An edge separated from the target product by one edge has 2 hop counts and an attenuation factor of 0.9 to the power of 2, which is 0.81. And so on, the farther the edge is from the target product, the smaller the attenuation factor.

[0179] In one example, the association strength value of the product chain is calculated based on the contribution value and attenuation coefficient of at least one edge. Specifically, the contribution value of each edge is multiplied by its attenuation coefficient to obtain the attenuation contribution value of that edge. The association strength value of the product chain is obtained by summing the attenuation contribution values ​​of all edges in the product chain.

[0180] In one example, taking the product chain from product 1 to product 3 to product 4 as an example, the target product is product 1. The edge between product 1 and product 3 is the first edge, with a hop count of 1, a contribution value of 0.5, and a decay coefficient of 0.9^1 (0.9). The decay contribution value is the product of 0.5 and 0.9, which is 0.45. The edge between product 3 and product 4 is the second edge, with a hop count of 2, a contribution value of 0.5, and a decay coefficient of 0.9^2 (0.81). The decay contribution value is the product of 0.5 and 0.81, which is 0.405. The association strength of the product chain from product 1 to product 3 to product 4 is the sum of 0.45 and 0.405, which is 0.855.

[0181] In one implementation, calculating the association strength value of a product chain based on the contribution value and attenuation coefficient of at least one edge includes: calculating the total strength value of the product chain based on the contribution value and attenuation coefficient of at least one edge; and normalizing the total strength value based on the number of nodes in the product chain to obtain the association strength value of the product chain.

[0182] Specifically, the total strength value is the sum of the decay contributions of all edges in the product chain. Normalization involves dividing the total strength value by the number of nodes in the product chain to obtain the normalized association strength value. Normalization makes the association strength values ​​comparable between product chains of different lengths.

[0183] In one example, taking the product chain from product 1 to product 3 to product 4 as an example, the number of nodes in the product chain is 3. The total strength value is the sum of 0.45 and 0.405, which is 0.855. Normalizing the total strength value, we get the association strength value as 0.855 divided by 3, which is 0.285.

[0184] In one example, taking a product chain from product 1 to product 3 to product 7 as an example, the number of nodes in the product chain is 3. The edge between product 1 and product 3 corresponds to the association type of substitute product, with a contribution value of 0.5, a hop count of 1, a decay coefficient of 0.9, and a decay contribution value of 0.45. The edge between product 3 and product 7 corresponds to the association type of complementary product, with a contribution value of 0.3, a hop count of 2, a decay coefficient of 0.81, and a decay contribution value of 0.243. The total strength value is the sum of 0.45 and 0.243, which is 0.693. Normalizing the total strength value, we get the association strength value as 0.693 divided by 3, which is 0.231.

[0185] The above method determines the contribution value of each edge based on the association type between multiple products in the product chain, determines the attenuation coefficient based on the number of hops between the edge and the target product, calculates the association strength value of the product chain based on the contribution value and the attenuation coefficient, and normalizes the total strength value based on the number of nodes in the product chain. Therefore, the association strength value of the product chain reflects the tightness of the association between products in the chain, and the attenuation coefficient shows that the farther the edge is from the target product, the smaller its impact on the association strength value. Normalization ensures the comparability of association strength values ​​between product chains of different lengths. The product set is determined based on the association strength value, ensuring that closely related products are grouped into the same product set, providing a basis for subsequently constructing a query index corresponding to the product set.

[0186] Based on the above-described map indexing construction method, this disclosure also provides a map indexing construction apparatus. The following will be combined with... Figure 5 The device is described in detail.

[0187] Figure 5 A schematic block diagram of a map index construction apparatus according to an embodiment of the present disclosure is shown.

[0188] like Figure 5 As shown, the map index construction device 500 of this embodiment includes a map acquisition module 501, a product chain construction module 502, a correlation strength calculation module 503, a product set determination module 504, and an index construction module 505.

[0189] The graph acquisition module 501 is used to acquire product graphs and information on historical products held by users. In the product graph, multiple nodes represent multiple products, the edges between nodes represent the association type between products, and the attributes of the nodes represent the feature information of the products.

[0190] Product chain construction module 502 is used to determine multiple associated nodes associated with historical products from the multiple nodes, and construct a product chain for each product, wherein the associated node represents an associated product associated with a historical product;

[0191] The association strength calculation module 503 is used to calculate the association strength value of the product chain based on the association type between multiple associated products included in the product chain;

[0192] Product set determination module 504 is used to determine multiple product sets based on the association strength values ​​of the multiple product chains;

[0193] The index building module 505 is used to build multiple sets of query indexes corresponding to the multiple product sets based on the feature information of the associated products, wherein the data in the same set of query indexes are stored in the same storage shard.

[0194] According to embodiments of this disclosure, the apparatus further includes: an update acquisition module, configured to acquire product update information and determine at least one product to be updated based on the product update information; a product set to be updated determination module, configured to determine at least one set of products to be updated corresponding to the at least one product to be updated; and an update module, configured to update a query index corresponding to the at least one set of products to be updated based on the product update information.

[0195] According to embodiments of this disclosure, the apparatus further includes: a data acquisition module for acquiring user transaction data and information on historical products held by the user; an expected value determination module for determining an expected product feature value based on the user transaction data and the information on historical products held by the user; a query module for querying at least one candidate product matching the expected product feature value from multiple products included in the product map based on the expected product feature value and the multiple sets of query indexes; and a product determination module for determining a product to be recommended based on the association type between the at least one candidate product and the historical products held by the user.

[0196] According to embodiments of this disclosure, the association type includes one of substitute products, complementary products, and mutually exclusive products;

[0197] The product chain construction module 502 includes an associated node determination unit, which is used to determine a first type of node representing the historical product and a second type of node whose association type with the first type of node belongs to the substitute product and / or complementary product as the multiple associated nodes.

[0198] According to embodiments of this disclosure, the product determination module includes: a deletion unit, configured to delete candidate products whose association type is mutually exclusive or not associated with the historical product from the at least one candidate product; a first determination unit, configured to determine the priority of candidate products whose association type is substitute product as a first priority; a second determination unit, configured to determine the priority of candidate products whose association type is complementary product as a second priority, wherein the first priority is higher than the second priority; and to determine the product to be recommended based on the candidate products corresponding to the first priority and the second priority.

[0199] According to an embodiment of this disclosure, the association strength calculation module 503 includes: a contribution value calculation unit, used to determine a contribution value corresponding to at least one edge in the product chain based on the association type between the multiple products included in the product chain; and an association strength calculation unit, used to calculate the association strength value of the product chain based on the contribution value corresponding to at least one edge in the product chain.

[0200] According to an embodiment of this disclosure, the association strength calculation unit includes: an attenuation coefficient calculation subunit, used to determine the attenuation coefficient corresponding to the at least one edge based on the target product targeted by the product chain and the number of hops between each of the at least one edge and the node of the target product; and an association strength calculation subunit, used to calculate the association strength value of the product chain based on the contribution value corresponding to the at least one edge and the attenuation coefficient corresponding to the at least one edge.

[0201] According to an embodiment of this disclosure, the index building module 505 includes: a numerical range determination unit, used to build multiple sets of feature numerical ranges based on the feature information of the product; and an index building unit, used to establish a mapping relationship between the multiple sets of feature numerical ranges and the multiple products, and generate the query index.

[0202] According to embodiments of this disclosure, any multiple modules among the spectrum acquisition module 501, product chain construction module 502, correlation strength calculation module 503, product set determination module 504, and index construction module 505 can be merged into one module, or any one of these modules can be split into multiple modules. Alternatively, at least some of the functions of one or more of these modules can be combined with at least some of the functions of other modules and implemented in one module. According to embodiments of this disclosure, at least one of the spectrum acquisition module 501, product chain construction module 502, correlation strength calculation module 503, product set determination module 504, and index construction module 505 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, at least one of the following modules can be implemented, at least partially, as a computer program module: the map acquisition module 501, the product chain construction module 502, the association strength calculation module 503, the product set determination module 504, and the index construction module 505. When the computer program module is run, it can perform the corresponding functions.

[0203] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing a map indexing construction method according to an embodiment of the present disclosure.

[0204] like Figure 6As shown, an electronic device 600 according to an embodiment of this disclosure includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage portion 608 into a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this disclosure.

[0205] RAM 603 stores various programs and data required for the operation of electronic device 600. Processor 601, ROM 602, and RAM 603 are interconnected via bus 604. Processor 601 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 602 and / or RAM 603. It should be noted that the programs may also be stored in one or more memories other than ROM 602 and RAM 603. Processor 601 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in said one or more memories.

[0206] According to embodiments of this disclosure, the electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to a bus 604. The electronic device 600 may also include one or more of the following components connected to the input / output (I / O) interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 610 as needed so that computer programs read from it can be installed into the storage section 608 as needed.

[0207] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.

[0208] According to embodiments of this disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this disclosure, the computer-readable storage medium may include ROM 602 and / or RAM 603 and / or one or more memories other than ROM 602 and RAM 603 described above.

[0209] Embodiments of this disclosure also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to cause the computer system to implement the methods provided in the embodiments of this disclosure.

[0210] When the computer program is executed by the processor 601, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0211] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and downloaded and installed via the communication section 609, and / or installed from the removable medium 611. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0212] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, it performs the functions defined in the system of this disclosure embodiment. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0213] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can execute entirely on a user's computing device, partially on a user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0214] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0215] Those skilled in the art will understand that the features described in the various embodiments of this disclosure can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this disclosure. In particular, the features described in the various embodiments of this disclosure can be combined and / or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.

[0216] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A method for constructing a map index, comprising: Obtain a product graph and information on historical products held by the user, wherein multiple nodes in the product graph represent multiple products, edges between nodes represent the association types between products, and attributes of nodes represent the characteristic information of products; based on the association types corresponding to the edges, determine multiple associated nodes associated with historical products from the multiple nodes, and construct a product chain for each associated product, wherein the associated node represents an associated product associated with a historical product; Based on the association types among the multiple related products included in the product chain, the association strength value of the product chain is calculated; Multiple product sets are determined based on the correlation strength values ​​of each of the aforementioned product chains; Based on the feature information of the associated products, multiple sets of user query indexes corresponding to the multiple product sets are constructed.

2. The method according to claim 1, further comprising: Obtain product update information and determine at least one product to be updated based on the product update information; Determine at least one set of products to be updated corresponding to the at least one product to be updated; Based on the product update information, update the query index corresponding to the at least one set of products to be updated.

3. The method according to claim 1, wherein, The association type includes one of the following: substitute products, complementary products, and mutually exclusive products; Based on the association type corresponding to the edge, the multiple associated nodes determined from the multiple nodes that are associated with historical products include: The first type of node representing the historical product, and the second type of node whose association type with the first type of node belongs to the substitute product and / or complementary product, are identified as the plurality of associated nodes.

4. The method according to claim 1, further comprising: Obtain user transaction data; Based on the user's transaction data and the information on the user's historical products, the desired product characteristic value is determined; Based on the expected product feature value and the multiple sets of query indexes, at least one candidate product matching the expected product feature value is obtained from the multiple products included in the product map; The product to be recommended is determined based on the association type between the at least one candidate product and the user's historical products.

5. The method according to claim 4, wherein, Based on the association type between the at least one candidate product and the user's historical products, the products to be recommended include: Remove candidate products whose association type is mutually exclusive or that are not associated with the historical products from the at least one candidate product; Candidate products whose association type is "substitute product" are prioritized as first priority. The priority of candidate products whose association type is complementary products is determined to be the second priority, wherein the first priority is higher than the second priority; Based on the candidate products corresponding to the first priority and the second priority, the products to be recommended are determined.

6. The method according to claim 1, wherein, Based on the association types among the multiple related products included in the product chain, the association strength value of the product chain is calculated as follows: Based on the association type among the multiple related products included in the product chain, determine the contribution value corresponding to at least one edge in the product chain; The association strength value of the product chain is calculated based on the contribution value corresponding to at least one edge in the product chain.

7. The method according to claim 6, wherein, The association strength value of the product chain is calculated based on the contribution value corresponding to at least one edge in the product chain, including: Based on the target product of the product chain, the attenuation coefficient corresponding to the at least one edge is determined according to the number of hops between each edge and the node of the target product. The correlation strength value of the product chain is calculated based on the contribution value corresponding to the at least one edge and the attenuation coefficient corresponding to the at least one edge.

8. The method according to claim 1, wherein, Based on the feature information of the products, constructing multiple sets of query indexes corresponding to the multiple product sets includes: Based on the feature information of the product, multiple sets of feature value ranges are constructed; Establish mapping relationships between multiple sets of the aforementioned feature value ranges and multiple of the aforementioned products to generate the query index.

9. A map index construction apparatus, comprising: The graph acquisition module is used to acquire product graphs and information on historical products held by users. In the product graph, multiple nodes represent multiple products, the edges between nodes represent the association type between products, and the attributes of the nodes represent the feature information of the products. The product chain construction module is used to determine multiple associated nodes associated with historical products from the multiple nodes based on the association type corresponding to the edge, and construct a product chain for each product, wherein the associated node represents an associated product associated with a historical product; The association strength calculation module is used to calculate the association strength value of the product chain based on the association type between multiple associated products included in the product chain; The product set determination module is used to determine multiple product sets based on the association strength values ​​of the multiple product chains. The index building module is used to build multiple sets of query indexes corresponding to the multiple product sets based on the feature information of the associated products, wherein the data in the same set of query indexes are stored in the same storage shard.

10. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 8.

12. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 8.