Resource pool classification method and device, computer equipment and storage medium

Abstract

The embodiment of the invention provides a resource pool classification method and device, computer equipment and a storage medium, belongs to the technical field of classification, and is suitable for the field of financial science and technology. The method comprises the following steps: acquiring target investment data of a target resource pool; wherein the target investment data comprises investor attribute information, investment product information and an investment relationship category; performing graph network construction according to the investor attribute information, the investment product information and the investment relationship category to obtain an investment relationship graph network; performing feature extraction on the target investment data according to the investment relation graph network to obtain target investment features of the target resource pool; screening out a target cluster from preset reference clusters according to the target investment characteristics and preset reference investment characteristics; wherein the reference investment features are derived from a reference cluster; and classifying the target resource pool according to the target cluster. According to the embodiment of the invention, the resource pool classification manpower can be reduced, and the classification accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of classification technology, applicable to the field of financial technology, and in particular to a resource pool classification method and device, computer equipment and storage medium. Background Art

[0002] In the financial field, various types of resource pools are set up to store different types of financial resources. Before storing financial resources, it is necessary to classify the resource pools according to their characteristics to make the division of financial resources more accurate. In related technologies, the classification of resource pools is mainly done manually. Financial service personnel automatically collect investment data of resource pools in the financial field and manually classify resource pools based on the investment data. However, manual analysis of a large amount of data is required, which results in a large amount of manpower consumed in the classification of resource pools and low classification accuracy. Therefore, how to reduce the manpower required for resource pool classification and improve classification accuracy has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to propose a resource pool classification method and apparatus, computer equipment and storage medium, aiming to reduce the manpower of resource pool classification and improve the accuracy of classification.

[0004] To achieve the above object, a first aspect of an embodiment of the present application proposes a resource pool classification method, the method comprising:

[0005] Acquire target investment data of the target resource pool; wherein the target investment data includes: investor attribute information, investment product information and investment relationship category, the investor attribute information is the attribute information of the investment object, and the investment relationship category represents the relationship category between the investment object and the investment product;

[0006] Constructing a graph network according to the investor attribute information, the investment product information and the investment relationship category to obtain an investment relationship graph network;

[0007] Extracting features of the target investment data according to the investment relationship graph network to obtain target investment features of the target resource pool;

[0008] Filtering a target cluster from a preset reference cluster according to the target investment feature and a preset reference investment feature; wherein the reference investment feature is derived from the reference cluster;

[0009] The target resource pool is classified according to the target cluster.

[0010] In some embodiments, constructing a graph network according to the investor attribute information, the investment product information and the investment relationship category to obtain an investment relationship graph network includes:

[0011] Performing node conversion on the investor attribute information and the investment product information to obtain candidate investment nodes;

[0012] The candidate investment nodes are connected according to the investment relationship categories to obtain the investment relationship graph network.

[0013] In some embodiments, extracting features from the target investment data according to the investment relationship graph network to obtain target investment features of the target resource pool includes:

[0014] Select any one of the candidate investment nodes from the investment relationship graph network as a reference investment node;

[0015] Perform node traversal in the investment relationship graph network according to the preset node traversal indication information and the reference investment node to obtain a target node list; wherein the target node list includes at least one target investment node and a node relationship chain between the target investment nodes;

[0016] Feature extraction is performed on the target investment data according to the target investment node and the node relationship chain to obtain the target investment feature of the target resource pool.

[0017] In some embodiments, the node traversal indication information includes: the number of traversed nodes and the type of traversed nodes;

[0018] The step of performing node traversal in the investment relationship graph network according to the preset node traversal indication information and the reference investment node to obtain a target node list includes:

[0019] Extracting associated investment nodes from the investment relationship graph network according to the reference investment nodes and the traversal node categories;

[0020] Acquire the link between the reference investment node and the associated investment node according to the investment relationship graph network to obtain the node relationship chain;

[0021] Obtaining the number of associated investment nodes to obtain the number of associated nodes;

[0022] The associated investment nodes, the reference investment nodes and the node relationship chain are concatenated according to the associated node quantity and the traversed node quantity to obtain the target node list.

[0023] In some embodiments, screening out a target cluster from preset reference clusters according to the target investment feature and preset reference investment feature includes:

[0024] Calculating similarity between the target investment feature and the reference investment feature to obtain target feature similarity;

[0025] The target cluster is selected from the reference clusters according to the target feature similarity.

[0026] In some embodiments, before selecting a target cluster from a preset reference cluster according to the target investment feature and the preset reference investment feature, the method further includes:

[0027] Constructing the reference cluster specifically includes:

[0028] Obtain the investment characteristics of the historical resource pool to obtain the historical investment characteristics;

[0029] Selecting a preliminary investment characteristic from any one of the historical investment characteristics;

[0030] Constructing clusters according to the preliminary investment characteristics to obtain preliminary clusters;

[0031] Calculating similarity between the preliminary investment features and the historical investment features to obtain preliminary feature similarity;

[0032] The preliminary cluster is updated according to the preliminary feature similarity and the historical investment feature to obtain the reference cluster.

[0033] In some embodiments, updating the preliminary cluster according to the preliminary feature similarity and the historical investment feature to obtain the reference cluster includes:

[0034] screening selected investment features from the historical investment features based on the preliminary feature similarity;

[0035] updating the preliminary cluster according to the selected investment characteristics to obtain a candidate cluster;

[0036] Acquire investment features of the candidate cluster to obtain at least one candidate investment feature;

[0037] Calculating the feature mean of at least one of the candidate investment features to obtain an updated investment feature; wherein the updated investment feature is the investment feature corresponding to the center of the candidate cluster;

[0038] The candidate cluster is updated according to the updated investment feature and the historical investment feature to obtain the reference cluster.

[0039] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application proposes a resource pool classification device, the device comprising:

[0040] A data acquisition module is used to acquire target investment data of a target resource pool; wherein the target investment data includes: investor attribute information, investment product information and investment relationship category, wherein the investor attribute information is the attribute information of the investment object, and the investment relationship category represents the relationship category between the investment object and the investment product;

[0041] A graph network construction module, used to construct a graph network according to the investor attribute information, the investment product information and the investment relationship category to obtain an investment relationship graph network;

[0042] A feature extraction module, used to extract features from the target investment data according to the investment relationship graph network to obtain target investment features of the target resource pool;

[0043] A cluster screening module, used to screen out a target cluster from a preset reference cluster according to the target investment characteristics and preset reference investment characteristics; wherein the reference investment characteristics are derived from the reference cluster;

[0044] A classification module is used to classify the target resource pool according to the target cluster.

[0045] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes a computer device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

[0046] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0047] The resource pool classification method and device, computer equipment and storage medium proposed in the present application construct an investment relationship graph network of the target resource pool in advance according to the investor attribute information, investment product information and investment relationship category of the target resource pool, and then determine the investment characteristics of the target resource pool based on the investment relationship graph network to obtain more accurate and comprehensive target investment characteristics. Therefore, when classifying the target resource pool, the target cluster is found based on the target investment characteristics and the reference investment characteristics, so that the search for the target cluster is accurate, so the classification operation of the target resource pool completed through the target cluster is also more accurate. Therefore, by automating the construction of the investment relationship graph network and the classification operation, not only the manpower of the target resource pool classification is saved, but also the accuracy of the target resource pool classification can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a system architecture diagram of the resource pool classification method provided in the embodiment of the present application;

[0049] Figure 2 is a flow chart of a resource pool classification method provided in an embodiment of the present application;

[0050] Figure 3 yes Figure 2 Flow chart of step S202 in FIG.

[0051] Figure 4 It is a schematic diagram of an investment relationship graph network in the resource pool classification method provided in an embodiment of the present application;

[0052] Figure 5 yes Figure 2 Flow chart of step S203 in FIG.

[0053] Figure 6 It is a schematic diagram of the traversal process of the investment relationship graph network in the resource pool classification method provided in the embodiment of the present application;

[0054] Figure 7 yes Figure 5 Flow chart of step S502 in FIG.

[0055] Figure 8 is a flow chart of a resource pool classification method provided by another embodiment of the present application;

[0056] Fig. 9 It is an illustration of dividing historical investment characteristics into multiple reference clusters in the resource pool classification method provided in the embodiment of the present application;

[0057] Fig.10 yes Figure 8 Flow chart of step S805 in FIG.

[0058] Fig.11 yes Figure 2 Flow chart of step S204 in FIG.

[0059] Fig.12 is a structural diagram of a resource pool classification device provided in an embodiment of the present application;

[0060] Fig.13 It is a schematic diagram of the hardware structure of the computer device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0063] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0064] First, some nouns involved in this application are analyzed:

[0065] Artificial intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. AI is a branch of computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems. AI can simulate the information process of human consciousness and thinking. AI is also a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0066] Unsupervised Learning: It is a type of machine learning algorithm that does not require labeled training data, but automatically discovers patterns, structures, and relationships from unlabeled data. The purpose of unsupervised learning is to cluster data, reduce dimensionality, discover implicit structures and relationships, etc.

[0067] Reachability Model: It is a mathematical model that describes the reachability relationship between nodes in a network. The reachability model can be used to describe the topology of a network, calculate network reachability, detect network attacks, and optimize network reliability.

[0068] Graph Neural Network (GNN): is a deep learning algorithm designed to process graph structure data. Graph Neural Network combines the topological structure information of the graph structure with the feature information of nodes or edges to learn to represent nodes or edges in the graph.

[0069] Cluster: Clustering attempts to divide samples in the data into several usually disjoint subsets, each of which is called a "cluster". Through such division, each cluster may correspond to some potential concepts (categories).

[0070] K-means algorithm: K-means algorithm originated from a vector quantization method in signal processing, and is now more popular in the field of data mining as a clustering analysis method. The purpose of K-means clustering is to divide n points into k clusters so that each point belongs to the cluster corresponding to the nearest mean, which is used as the clustering standard.

[0071] In the financial field, there will be a resource pool for storing various financial resources, and the resource pool is also called a fundraising fund group, which is used to collect various financial resources. In order to facilitate the collection of financial resources, different resource pools will be classified in advance. In the related technology, the classification operation of the resource pool is performed by financial service personnel who automatically collect the financial investment data of the resource pool, and then determine the investment characteristics of the resource pool based on the investment data, and then complete the classification of the resource pool according to the investment characteristics. However, the classification operation of the entire resource pool is completed manually, which not only requires a lot of manpower, but also affects the accuracy of the resource pool classification.

[0072] Based on this, the embodiment of the present application provides a resource pool classification method and device, computer equipment and storage medium, which aims to collect the investment attribute information, investment product information and investment relationship categories of the target resource pool in terms of investment, establish an investment relationship graph network that characterizes the investment aspects of the target resource pool according to the investor attribute information, investment product information and investment relationship categories, and determine more accurate investment characteristics of the target resource pool through the investment relationship graph network. Therefore, the target resource pool is clustered through more accurate investment characteristics and existing reference investment characteristics, and the automatic classification operation of the target resource pool is realized, which not only saves manpower for the classification of the target resource pool, but also makes the classification of the target resource pool more accurate.

[0073] The resource pool classification method and apparatus, computer equipment and storage medium provided in the embodiments of the present application are specifically described through the following embodiments. First, the resource pool classification method in the embodiments of the present application is described.

[0074] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0075] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0076] The resource pool classification method provided in the embodiment of the present application relates to the field of classification technology and is applicable to the field of financial technology. The resource pool classification method provided in the embodiment of the present application can be applied to the terminal, can also be applied to the server side, and can also be software running in the terminal or the server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or it can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the resource pool classification method, etc., but is not limited to the above forms.

[0077] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer computer devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0078] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0079] Please refer to Figure 1 , Figure 1 It is a system architecture diagram of the resource pool classification system provided in an embodiment of the present application, including: a terminal 110, the Internet 120, a gateway 130 and a server 140, etc.

[0080] The terminal 110 is a device for collecting target investment data of the target resource pool, and these data can be input from the terminal 110 in the form of a code string. The terminal 110 includes various forms such as desktop computers, laptop computers, PDAs (personal digital assistants), mobile phones, vehicle-mounted terminals, and dedicated terminals. In addition, it can be a single device or a combination of multiple devices. For example, multiple desktop computers are interconnected through a local area network, share a display, etc. to work together, and together constitute a terminal 110. The terminal 110 can communicate with the Internet 120 in a wired or wireless manner to exchange data.

[0081] The server 140 is a computer system that completes the classification operation of the target resource pool based on the target investment data uploaded by the terminal 110. Compared with the ordinary terminal 110, the server 140 has higher requirements in terms of stability, security, performance, etc. The server 140 can be a high-performance computer in the network platform, a cluster of multiple high-performance computers, a part of a high-performance computer (such as a virtual machine), a combination of parts of multiple high-performance computers (such as virtual machines), etc. In some embodiments, the server 140 includes a data acquisition module, a graph network construction module, a feature extraction module, a cluster screening module and a classification module. It should be noted that the data acquisition module is used to obtain the target investment data of the target resource pool from the terminal 110, the graph network construction module is used to construct the investment relationship graph network of the target resource pool according to the target investment data, the feature extraction module is used to extract the target investment features of the target investment data according to the investment relationship graph network, the cluster screening module is used to determine the target cluster of the target resource pool according to the target investment features, and the classification module is used to determine the category of the target resource pool according to the target cluster. It should be understood that the data acquisition module, graph network construction module, feature extraction module, cluster screening module and classification module can be integrated in the same server 140, or can be deployed in different servers 140 respectively, and are not limited to the specific embodiments listed above.

[0082] The gateway 130 is also called an internetwork connector or a protocol converter. The gateway 130 realizes network interconnection at the transport layer and is a computer system or device that acts as a converter. The gateway 130 is a translator between two systems that use different communication protocols, data formats or languages, or even have completely different architectures. At the same time, the gateway 130 can also provide filtering and security functions. The message sent by the terminal 110 to the server 140 must be sent to the corresponding server 140 through the gateway 130. The message sent by the server 140 to the terminal 110 must also be sent to the corresponding terminal 110 through the gateway 130.

[0083] Figure 2 is an optional flowchart of the resource pool classification method provided in an embodiment of the present application. Figure 2 The method may include but is not limited to steps S201 to S205.

[0084] Step S201, obtaining target investment data of a target resource pool; wherein the target investment data includes: investor attribute information, investment product information and investment relationship category, the investor attribute information is the attribute information of the investment object, and the investment relationship category represents the relationship category between the investment object and the investment product;

[0085] Step S202, constructing a graph network according to the investor attribute information, investment product information and investment relationship categories to obtain an investment relationship graph network;

[0086] Step S203, extracting features from the target investment data according to the investment relationship graph network to obtain target investment features of the target resource pool;

[0087] Step S204, selecting a target cluster from preset reference clusters according to the target investment characteristics and preset reference investment characteristics; wherein the reference investment characteristics are derived from the reference cluster;

[0088] Step S205: classify the target resource pool according to the target cluster.

[0089] In the steps S201 to S205 shown in the embodiment of the present application, by acquiring the target investment data of the target resource pool, and the target investment data includes the investor attribute information, investment product information and investment relationship category, the investment relationship graph network of the target resource pool is first constructed according to the investor attribute information, investment product information and investment relationship category to clarify the relationship of the target resource pool in terms of investment. Therefore, when classifying the target resource pool, the target investment characteristics of the target resource pool will be extracted based on the investment relationship graph network to describe the target investment characteristics of the target resource pool in terms of investment in a more comprehensive and accurate manner, and the target cluster where the target resource pool is located will be determined according to the target investment characteristics, and the classification operation of the target resource pool will be completed through the target cluster, which not only realizes the automatic classification operation of the target resource pool and saves manpower, but also the entire classification operation will be completed based on comprehensive and accurate investment characteristics, and the accuracy of the classification of the target resource pool will be improved.

[0090] In step S101 of some embodiments, the target resource pool is a database for storing resources. If in the financial field, the target resource pool stores financial resources. If in the computer field, the target resource pool can store computer resources. The target investment data is the resource allocation relationship data of the target resource pool, and the resource allocation relationship data includes the resource input object, the relationship data between the resource input object and the resource. The relationship of the target resource pool when making resource allocation can be determined through the resource allocation relationship data, and different target resource pools have different resource allocation relationship data. Among them, the target investment data includes: investor attribute information, investment product information and investment relationship category. The investor attribute information represents the attribute information of the investment object existing in the target resource pool, and the investor attribute information includes: object category, object age, object gender and object identity information, etc. This embodiment does not make specific restrictions on the investor attribute information. The investment product information is the information of the investment product invested by the investment object. In the financial field, the investment product information includes: product category, product nature, product risk information and product purchase group category, etc. The investment relationship category represents the relationship category between the investment object and the investment product, and the investment relationship category includes: management relationship category and investment relationship category. The management relationship category represents the existence of a management relationship between the investment object and the investment product, and the investment relationship category represents the existence of an investment relationship between the investment object and the investment product.

[0091] In one application scenario, the target resource pool is an investment fund group, and the investment fund group is used to collect different categories of financial resources, complete the management of financial resources, and achieve the growth of financial resources. The investor attribute information of the investment fund group is specifically the attribute information of the investment users participating in the investment fund group, the investment product information is the financial products purchased by the investment users, and the financial products can be bonds, funds, and equity, etc. The investment relationship category represents the relationship between the investment user and the financial product, which can be a management relationship or an investment relationship, etc. Therefore, in the field of financial investment, by obtaining the attribute information of the investment users of the investment fund group in the process of collecting financial resources and the relationship between the investment users and the financial products, the investment characteristics of the investment fund group can be clearly understood, and the classification operation of the investment fund group can be accurately completed.

[0092] See also Figure 3 In some embodiments, step S202 may include but is not limited to steps S301 to S302:

[0093] Step S301, converting the investor attribute information and investment product information into nodes to obtain candidate investment nodes;

[0094] Step S302, connecting the candidate investment nodes according to the investment relationship categories to obtain an investment relationship graph network.

[0095] In step S301 of some embodiments, the candidate investment nodes are obtained by converting the investor attribute information and the investment product information into nodes, defining the candidate investment node corresponding to the investor attribute information as the first investment node, defining the candidate investment node corresponding to the investment product information as the second investment node, and the first investment node and the second investment node are represented in different forms. Specifically, the investor attribute information includes object categories, and the forms represented by the first investment nodes corresponding to different object categories are also different, and the object categories can be individuals and institutions. The investment product information also includes product categories, and the forms of the second investment nodes corresponding to different product categories are also different. Therefore, by constructing candidate investment nodes of different forms to represent investment objects and investment products, the characteristics of the investment objects and investment products in the target resource pool can be more clearly understood.

[0096] Specifically, Figure 4 As shown, Figure 4 The schematic diagram of the investment relationship graph network is shown. The first investment node corresponding to an individual is represented by a square node, the first investment node corresponding to an institution is represented by a cylindrical node, the product categories include equity, bonds and funds, the second investment node corresponding to equity is represented by a triangular node, the second investment node corresponding to bonds is represented by a star node, and the second investment node corresponding to funds is represented by a circular node. Therefore, through the investment relationship graph network, it is possible to clearly understand the categories of investment objects involved in the target resource pool and the categories of investment products, so as to determine the characteristics of the target resource pool in terms of investment.

[0097] In step S302 of some embodiments, the investment relationship category represents the relationship category between the investment object and the investment product, and different investment relationship categories use different connection chains, and then two candidate investment nodes with a relationship are connected through the connection chain to construct an investment relationship graph network. It should be noted that before connecting the candidate investment nodes, an adjacency matrix is ​​first constructed. The adjacency matrix records the connection relationship and connection category between any two candidate investment nodes. The connection relationship represents whether there is an investment relationship or a management relationship, and 0 and 1 are used to represent whether there is a connection between any two candidate investment nodes. Then, multiple candidate investment nodes are connected according to the adjacency matrix to obtain an investment relationship graph network.

[0098] Specifically, Figure 4 As shown, the investment relationship category of this embodiment includes a management relationship category and an investment relationship category. The connection chain corresponding to the management relationship category is a dotted line, and the connection chain corresponding to the investment relationship category is a solid line. Two candidate investment nodes with an investment relationship are connected by a solid line, and two candidate investment nodes with a management relationship are connected by a dotted line to form an investment relationship graph network. Therefore, the relationship between the investment objects and investment products in the target resource pool can be clearly seen through the investment relationship graph network.

[0099] In steps S301 to S302 shown in this embodiment, the investor attribute information and investment product information are represented in the form of nodes, and the nodes are connected according to the investment relationship categories to construct an investment relationship graph network that can represent the target resource pool, so that the investment relationship of the target resource pool is clearer and the investment characteristics of the target resource pool are easily mined.

[0100] It should be noted that after completing the investment relationship graph network, it is necessary to use the investment relationship graph network to complete the investment feature mining of the target resource pool. This embodiment uses the touch chain model and the investment relationship graph network to mine the target investment characteristics that can accurately characterize the target resource pool in terms of investment from the target investment data.

[0101] See also Figure 5 In some embodiments, step S203 may include but is not limited to steps S501 to S503:

[0102] Step S501, selecting any candidate investment node from the investment relationship graph network as a reference investment node;

[0103] Step S502, performing node traversal in the investment relationship graph network according to preset node traversal indication information and reference investment nodes to obtain a target node list; wherein the target node list includes at least one target investment node and a node relationship chain between target investment nodes;

[0104] Step S503: extracting features from the target investment data according to the target investment node and the node relationship chain to obtain target investment features of the target resource pool.

[0105] In step S501 of some embodiments, as disclosed above, the investment relationship graph network includes multiple candidate investment nodes, so before traversing the investment relationship graph network through the touch chain model, it is necessary to first select any candidate investment node as a reference investment node, and the reference investment node is used as the initial node when starting the traversal. It should be noted that after traversing the investment relationship graph network with the reference investment node as the initial node and traversing to the last candidate investment node, it is necessary to replace another candidate investment node as the reference investment node, and continue to traverse the investment relationship graph network with the new reference investment node, thereby repeating the cycle to complete the path traversal of all candidate investment nodes.

[0106] In step S502 of some embodiments, the node traversal indication information represents the operation instructions of the node traversal process, that is, the reference standard for determining the investment relationship graph network to be traversed. The node relationship chain is equivalent to the above-mentioned connection chain, which represents the relationship between any two candidate investment nodes. It should be noted that there is at least one target node list, and each reference investment node corresponds to a target node list. Among them, the target node list records the target investment nodes that can be reached by each reference investment node, and also records the node relationship chain between the reference investment node and the target investment node, and the node relationship chain between any two target investment nodes. Specifically, before starting the traversal, an empty touch chain list will be constructed first, and the node traversal process will be performed in the investment relationship graph network with the reference investment node and the node traversal indication information. The candidate investment nodes reached by the reference investment node will be continuously stored in the touch chain list as the target investment node until the last reachable candidate investment node is traversed to obtain the target node list.

[0107] In step S503 of some embodiments, feature extraction is performed on the target investment data according to the target investment node and the node relationship chain, that is, feature extraction is performed on the target investment data according to the target node list of each reference investment node, then the candidate investment features corresponding to each reference investment node are extracted, and the candidate investment features not only include their own investment features, but also include investment features with associated relationships, and then all candidate investment features are combined to obtain the target investment features of the target resource pool, so the target investment features are richer and more comprehensive, and the classification operation of the target resource pool based on the rich and comprehensive investment features will be more accurate. Specifically, the node features of the target investment data are extracted according to the target investment node, and the node features can be the regional features, industry features, and category features of the investment object or investment product. The relationship features of the target investment data are extracted according to the node relationship chain, and the relationship features represent the distance between investment objects and investment objects, the distance between investment objects and investment products, and the distance between investment products. This embodiment does not make specific restrictions on node features and relationship features. Finally, the node features and relationship features are spliced ​​into target investment features to obtain comprehensive and rich investment features.

[0108] Specifically, Figure 6 As shown, Figure 6A schematic diagram of the traversal process of the investment relationship graph network is shown. If the node traversal starts with node A, and the nodes that node A can reach are recorded as node C, node E, node F, and node H, the path information of the touch chain of the traversal process of node A is recorded and stored in the touch chain list to obtain the target node list B1. Then continue to traverse the nodes in the investment relationship graph network with node B to obtain the target node list B2. Thus, repeated traversal will obtain multiple target node lists, so that according to multiple target node lists, richer investment features of the target investment data can be mined, which are not only the independent features of the investment objects and investment products, but also the features of the investment objects or investment products associated with each investment object or investment product, as a more comprehensive and rich investment feature of the investment object or investment product.

[0109] For example, if applied to an equity fundraising scenario, the candidate investment nodes correspond to the funding party, the fundraising product or the investment target. According to the equity relationship, starting from a certain reference investment node (such as the funding party), the candidate investment node corresponding to the fundraising product is found through the node relationship chain, and then the candidate investment node corresponding to the investment target is found through the investment information of the fundraising product. Then, starting from the reference investment node corresponding to the funding party mentioned above, the candidate investment node corresponding to the second invested fundraising product is selected, and the candidate investment node corresponding to the investment target is also found. The cycle is repeated in sequence to complete the depth-first search and breadth-first search to obtain multiple target node lists.

[0110] In step S501 to step S503 shown in this embodiment, by selecting candidate investment nodes as reference investment nodes one by one, and completing the node traversal of the investment relationship graph network in a loop with the reference investment nodes and the node traversal indication information, a target node list corresponding to each reference investment node is obtained. Then, based on the target investment nodes and the node relationship chains in the multiple target node lists of the multiple reference investment nodes, the target investment features in the target investment data are extracted to obtain more comprehensive and rich investment features, so that the target resource pool can perform classification operations based on the comprehensive and rich target investment features more accurately.

[0111] In some embodiments, the node traversal indication information includes: the number of traversed nodes and the traversed node categories. It should be noted that the number of traversed nodes is used to indicate that the investment relationship graph network traversal process needs to traverse several categories of candidate investment nodes and node relationship chains, and the traversed node categories are the categories of candidate investment nodes that need to be traversed. For example, if the number of traversed nodes is 4, and the traversed node categories are: investment object category, investment product category, investment target and investment link, then each time a traversal operation is performed with reference to the investment node, the four traversal node categories need to be traversed.

[0112] See also Figure 7In some embodiments, step S502 may include but is not limited to steps S701 to S704:

[0113] Step S701, extracting associated investment nodes from the investment relationship graph network according to the reference investment nodes and the traversal node categories;

[0114] Step S702, obtaining a link connecting a reference investment node to an associated investment node according to the investment relationship graph network, and obtaining a node relationship chain;

[0115] Step S703, obtaining the number of associated investment nodes to obtain the number of associated nodes;

[0116] Step S704: according to the number of associated nodes and the number of traversed nodes, the associated investment nodes, the reference investment nodes and the node relationship chain are concatenated to obtain a target node list.

[0117] In step S701 of some embodiments, as disclosed above, the traversal node category gives the category of the candidate investment node that needs to be traversed by the reference investment node, so starting from the reference investment node, the category of the adjacent candidate investment node is obtained to obtain the candidate node category, and the candidate investment node whose candidate node category is the traversal node category is selected as the associated investment node, thereby repeatedly retrieving and obtaining nodes to output at least one associated investment node. It should be noted that when performing node traversal with the reference investment node, if there are multiple traversal node categories, first retrieve the associated investment node according to the second ranked traversal node category, and then continue to retrieve the candidate investment node corresponding to the third ranked traversal node category with the associated investment node, and repeatedly retrieve the candidate investment node of the last traversal node category to obtain at least one associated investment node. Specifically, if the traversal node categories are: investment object category, investment product category and investment target, first start from the reference investment node corresponding to the investment object, retrieve the candidate investment node corresponding to the investment product as the associated investment node, and then continue to retrieve the candidate investment node corresponding to the investment target with the associated investment node corresponding to the investment product to obtain two associated investment nodes.

[0118] In step S702 of some embodiments, as disclosed above, the associated investment nodes of each reference investment node have been retrieved, and the link from the reference investment node to each associated investment node can be determined as a node relationship chain through the investment relationship graph network to determine the relationship characteristics between the investment object and the investment product.

[0119] In steps S703 to S704 of some embodiments, when performing node traversal with the reference investment node, each node traversal operation will record the number of associated investment nodes touched by the reference investment association to obtain the number of associated nodes, and determine whether the traversal of the reference investment node is completed by the number of associated nodes. If the number of associated nodes reaches the number of traversed nodes, stop performing node traversal with the reference investment node, store at least one associated investment node and node relationship chain retrieved in the touched chain list, and then continue to start the retrieval of another link with the reference investment node until the traversal of the entire investment relationship graph network is completed to output the target node list.

[0120] In the step S701 to step S704 shown in the present embodiment, when the reference investment node is used for node traversal, the associated investment nodes of the reference investment node are first retrieved from the investment relationship graph network according to the traversal node category, and then the node relationship chain between the reference investment node and the associated investment node is determined. At the same time, after the number of associated investment nodes is calculated to reach the number of traversed nodes, the retrieved associated investment nodes, reference investment nodes and node relationship chains are spliced ​​into a target node list, so that the node traversal operation is accurate and simple, and a target node list with rich content can be obtained, so that the investment features extracted according to the target node list are also richer.

[0121] In some embodiments, before step S204, the resource pool classification method may further include constructing a reference cluster. It should be noted that the reference cluster is a resource pool cluster formed by clustering the investment characteristics of historical resource pools, and specifically, multiple historical resource pools are classified by clustering.

[0122] See also Figure 8 In some embodiments, building a reference cluster may include, but is not limited to, steps S801 to S805:

[0123] Step S801, obtaining the investment characteristics of the historical resource pool to obtain the historical investment characteristics;

[0124] Step S802, selecting a preliminary investment feature from any historical investment feature;

[0125] Step S803, constructing clusters according to the preliminary investment characteristics to obtain preliminary clusters;

[0126] Step S804, calculating the similarity between the selected investment features and the historical investment features to obtain a preliminary feature similarity;

[0127] Step S805, updating the preliminary cluster according to the preliminary feature similarity and the historical investment features to obtain a reference cluster.

[0128] In step S801 of some embodiments, a historical relationship network diagram of the historical resource pool is first constructed, and the historical investment data of the historical resource pool is extracted through the historical relationship network diagram to obtain historical investment features. It should be noted that the specific extraction process of historical investment features is similar to the extraction process of target investment features, which will not be repeated here.

[0129] In steps S802 to S803 of some embodiments, any one of the historical investment features is selected as a preliminary investment feature, and the preliminary feature is used as the feature corresponding to the cluster center to construct a preliminary cluster. It should be noted that similarity calculation can be performed on any two historical investment features to obtain an initial similarity, and multiple historical investment features whose initial similarity is less than a preset similarity are respectively constructed into multiple preliminary clusters.

[0130] In step S804 of some embodiments, as disclosed above, each preliminary cluster contains only one preliminary investment feature, so the preliminary investment feature is the investment feature corresponding to the cluster center of the preliminary cluster. The preliminary feature similarity is obtained by calculating the similarity between the preliminary investment feature and the historical investment feature. Specifically, the preliminary investment feature can be vectorized to obtain a preliminary investment vector, and then the historical investment feature can be vectorized to obtain a historical investment vector, and then the vector similarity between the preliminary investment vector and the historical investment vector is calculated to obtain the preliminary feature similarity. It should be noted that the vector similarity calculation can be completed using a vector similarity function, and the vector similarity function can include any of the following: cosine similarity function, Euclidean distance function, Manhattan distance function, Pearson correlation coefficient function and Hamming distance function. This embodiment does not impose specific restrictions on the vector similarity function.

[0131] For example, if the initial investment vector is P1(x1, y1) and the historical investment vector is P2(x2, y2), this embodiment uses the Euclidean distance function to determine the initial feature similarity as

[0132] In step S805 of some embodiments, the historical investment features are divided into preliminary clusters according to the preliminary feature similarity, so as to update the preliminary clusters to obtain reference clusters. It should be noted that the updating of the preliminary clusters to the reference clusters is to continuously and cyclically update the cluster centers of the preliminary clusters and the division of the historical investment features until the clustering of all the historical investment features is completed.

[0133] Please refer to Fig. 9 , Fig. 9 A schematic diagram showing the division of multiple historical investment features into multiple reference clusters is shown. Fig. 9 It can be seen that after the historical investment characteristics are represented by vectors, clustering processing is gradually performed to achieve the construction of multiple reference clusters. The whole process is equivalent to an unsupervised learning process. There is no need to label the historical investment characteristics in advance to complete the construction of reference clusters and realize automatic category division.

[0134] In step S801 to step S805 shown in this embodiment, preliminary investment features that can construct preliminary clusters are selected from the historical investment features, and then the feature similarity between each preliminary investment feature and the historical investment feature is calculated, and the remaining historical investment features are divided into corresponding preliminary clusters according to the feature similarity, so as to realize the updating of preliminary clusters and form reference clusters. Therefore, the classification of resource pools can be automatically completed without manual category labeling in advance, making the construction operation of reference clusters simple.

[0135] See also Fig.10 In some embodiments, step S805 includes but is not limited to steps S1001 to S1005:

[0136] Step S1001, selecting selected investment features from historical investment features according to preliminary feature similarity;

[0137] Step S1002, updating the preliminary clusters according to the selected investment features to obtain candidate clusters;

[0138] Step S1003, obtaining investment features of the candidate cluster to obtain at least one candidate investment feature;

[0139] Step S1004, calculating the feature mean of at least one candidate investment feature to obtain an updated investment feature; wherein the updated investment feature is the investment feature corresponding to the center of the candidate cluster;

[0140] Step S1005: updating the candidate cluster according to the updated investment characteristics and the historical investment characteristics to obtain a reference cluster.

[0141] In steps S1001 to S1002 of some embodiments, for each preliminary cluster, the historical investment feature corresponding to the maximum preliminary feature similarity is used as the selected investment feature, and the selected investment feature is added to the preliminary cluster to update the cluster center of the preliminary cluster.

[0142] In some embodiments, in steps S1003 to S1004, after the preliminary cluster is updated to the candidate cluster, the investment feature corresponding to the cluster center of the candidate cluster is no longer the preliminary investment feature, and the updated investment feature corresponding to the cluster center of the candidate cluster needs to be re-determined. By obtaining all candidate investment features in the candidate cluster, all candidate investment features are vectorized to obtain at least one candidate investment vector, and then the at least one candidate investment vector is averaged to obtain an updated investment vector, and the investment feature corresponding to the updated investment vector is used as the updated investment feature. Among them, this embodiment uses the K-means algorithm to average at least one candidate investment vector to obtain the updated investment vector. Therefore, the cluster center is updated by replacing the preliminary investment features in the preliminary cluster.

[0143] In step S1005 of some embodiments, similarity calculation is continued according to the updated investment features and the remaining historical investment features to obtain updated similarity, and the remaining historical investment features are placed in the candidate cluster according to the updated similarity to achieve the update of the candidate cluster. It should be noted that the update of the candidate cluster is the same as the update operation of the preliminary cluster, which is mainly through the continuous update of the cluster center, and then the correlation between the investment feature corresponding to the updated cluster center and the remaining historical investment features is calculated, and then the historical investment features are continuously divided until the last historical investment feature is divided. It should be further noted that after the last historical investment feature is classified, it is still necessary to calculate the distance between the cluster center of each cluster and the corresponding node of the adjacent investment feature, and continuously calculate the similarity and cluster update until the cluster center of each reference cluster no longer changes, so as to achieve the construction of the reference cluster.

[0144] See also Fig.11 In some embodiments, step S204 may include but is not limited to steps S1101 to S1102:

[0145] Step S1101, calculating the similarity between the target investment feature and the reference investment feature to obtain the target feature similarity;

[0146] Step S1102: Filter out a target cluster from the reference clusters according to target feature similarity.

[0147] In step S1101 of some embodiments, the target investment feature is first vectorized to obtain a target investment vector, the reference investment feature is vectorized to obtain a reference investment vector, and the target feature similarity is obtained by calculating the vector similarity between the target investment vector and the reference investment vector. It should be noted that the vector similarity between the target investment vector and the reference investment vector is completed using a vector similarity function, and the vector similarity function of this embodiment is the same as the aforementioned vector similarity function, which will not be repeated here.

[0148] In step S1102 of some embodiments, the reference cluster corresponding to the highest target feature similarity is used as the target cluster to determine the resource pool cluster closest to the target resource pool.

[0149] In step S1101 to step S1102 shown in this embodiment, the target cluster corresponding to the target resource pool is determined by feature similarity, so that the target cluster screening operation is simple, and thus the target resource pool classification operation is also simple.

[0150] In step S205 of some embodiments, the target cluster is also a resource pool cluster, and resource pool category information is set on the target cluster. The resource pool category information of the target resource pool is obtained to determine the resource pool category information of the target resource pool, making the target resource pool category determination operation simple.

[0151] The embodiment of the present application takes the financial scenario as an example to obtain the equity relationship data of an investment fund group, and the equity relationship data includes: fund party attribute information, fundraising product information, fundraising target information and relationship category. The fund party attribute information, fundraising product information and fundraising target information are first converted into nodes in a fundraising graph network, and then the nodes in the fundraising graph network are connected according to the relationship category to construct a fundraising relationship graph network representing the investment fund group.

[0152] When it is necessary to determine the category of the investment fund group, it is necessary to determine whether the investment fund group is a stable category, a low-risk category, a medium-risk category or a high-risk category. It is necessary to traverse the nodes and links of the fundraising relationship graph network through the touch chain model. Specifically, starting from a certain node (for example, the node corresponding to the fund party attribute information), find the node corresponding to the fundraising product information through the investment link, and then find the node corresponding to the fundraising target information through the node corresponding to the fundraising product information, and then circulate from the node corresponding to the fund party attribute information mentioned above, find the node corresponding to the second invested fundraising product information, and then find the node corresponding to the fundraising target information, and repeat the cycle in sequence, so as to complete the depth-first search and breadth-first search. It should be noted that when the nodes and links of each node (for example, the node corresponding to the fund party attribute information) are traversed, the length of the node relationship chain traversed each time will be calculated. When the length of the node relationship chain reaches the predefined touch chain length, the search is stopped, and the retrieved nodes and node relationship chains are input into the touch chain list, and then other paths are searched until the entire fundraising relationship graph network is traversed to obtain the target node list.

[0153] The node features are obtained by extracting the features of the equity relationship data according to the nodes in the target node list, and then the relationship features are obtained by extracting the features of the equity relationship data according to the node relationship chain in the target node list. The node features include the features of multiple funding parties, as well as the features of the fundraising products and the fundraising targets associated with the funding parties. The relationship features include the relationship features between multiple funding parties and the associated fundraising products and fundraising targets. Therefore, by combining the node features and the relationship features into the target investment features, a more comprehensive and rich investment feature of the investment fund group can be obtained.

[0154] When performing classification operations on investment fund groups, the investment characteristics corresponding to the cluster center of each reference fund cluster are first obtained as reference investment characteristics, and the target characteristic similarity is obtained by calculating the similarity between the target investment characteristics and the reference investment characteristics, and then the reference fund cluster corresponding to the highest target characteristic similarity is selected as the target fund cluster. By obtaining the fundraising category information corresponding to the target fund cluster, the target category information of the investment fund group is determined according to the fundraising category information, and the classification operation of the investment fund group is realized.

[0155] In summary, the classification operation of investment fund groups is automated, reducing manpower. In addition, in the classification process of investment fund groups, the fundraising relationship graph network of the investment fund groups will be constructed first, and then comprehensive and rich investment features will be extracted through the fundraising relationship graph network, so the classification of investment fund groups based on comprehensive and rich investment features is more accurate.

[0156] See also Fig.12 The embodiment of the present application further provides a resource pool classification device, which can implement the above resource pool classification method, and the device includes:

[0157] The data acquisition module 1201 is used to acquire target investment data of the target resource pool; wherein the target investment data includes: investor attribute information, investment product information and investment relationship category, the investor attribute information is the attribute information of the investment object, and the investment relationship category represents the relationship category between the investment object and the investment product;

[0158] A graph network construction module 1202 is used to construct a graph network according to the investor attribute information, investment product information and investment relationship categories to obtain an investment relationship graph network;

[0159] The feature extraction module 1203 is used to extract features of the target investment data according to the investment relationship graph network to obtain target investment features of the target resource pool;

[0160] The cluster screening module 1204 is used to screen out a target cluster from a preset reference cluster according to the target investment characteristics and the preset reference investment characteristics; wherein the reference investment characteristics are derived from the reference cluster;

[0161] The classification module 1205 is used to classify the target resource pool according to the target cluster.

[0162] The specific implementation of the resource pool classification device is basically the same as the specific implementation of the above-mentioned resource pool classification method, and will not be repeated here.

[0163] The embodiment of the present application also provides a computer device, the computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the resource pool classification method when executing the computer program. The computer device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0164] See also Fig.13 , Fig.13 The hardware structure of a computer device according to another embodiment is shown, and the computer device includes:

[0165] The processor 1301 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0166] The memory 1302 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1302 can store an operating system and other applications. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 1302, and the processor 1301 calls and executes the resource pool classification method of the embodiment of this application;

[0167] Input / output interface 1303, used to implement information input and output;

[0168] The communication interface 1304 is used to realize the communication interaction between the device and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);

[0169] A bus 1305 that transmits information between the various components of the device (e.g., the processor 1301, the memory 1302, the input / output interface 1303, and the communication interface 1304);

[0170] The processor 1301 , the memory 1302 , the input / output interface 1303 and the communication interface 1304 are connected to each other in communication within the device via a bus 1305 .

[0171] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned resource pool classification method is implemented.

[0172] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0173] The resource pool classification method and device, computer equipment and storage medium provided in the embodiments of the present application construct an investment relationship graph network of the target resource pool in advance according to the investor attribute information, investment product information and investment relationship category of the target resource pool, and then determine the investment characteristics of the target resource pool in combination with the investment relationship graph network to obtain more accurate and comprehensive target investment characteristics. Therefore, when classifying the target resource pool, the corresponding target cluster is found based on the target investment characteristics and the reference investment characteristics, so that the search operation of the target cluster is accurate, so the classification operation of the target resource pool completed through the target cluster is also more accurate. Therefore, by automating the construction of the investment relationship graph network and the classification operation, not only the manpower of the target resource pool classification is saved, but also the accuracy of the target resource pool classification can be improved.

[0174] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0175] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0176] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0177] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0178] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0179] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0180] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0181] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0182] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0183] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0184] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A resource pool classification method, characterized in that: The method comprises: Acquire target investment data of the target resource pool; wherein the target investment data includes: investor attribute information, investment product information and investment relationship category, the investor attribute information is the attribute information of the investment object, and the investment relationship category represents the relationship category between the investment object and the investment product; Constructing a graph network according to the investor attribute information, the investment product information and the investment relationship category to obtain an investment relationship graph network; Extracting features of the target investment data according to the investment relationship graph network to obtain target investment features of the target resource pool; Filtering a target cluster from a preset reference cluster according to the target investment feature and a preset reference investment feature; wherein the reference investment feature is derived from the reference cluster; The target resource pool is classified according to the target cluster.

2. The method according to claim 1, characterized in that The step of constructing a graph network according to the investor attribute information, the investment product information and the investment relationship category to obtain an investment relationship graph network includes: Performing node conversion on the investor attribute information and the investment product information to obtain candidate investment nodes; The candidate investment nodes are connected according to the investment relationship categories to obtain the investment relationship graph network.

3. The method according to claim 2, characterized in that The step of extracting features from the target investment data according to the investment relationship graph network to obtain target investment features of the target resource pool includes: Select any one of the candidate investment nodes from the investment relationship graph network as a reference investment node; Perform node traversal in the investment relationship graph network according to the preset node traversal indication information and the reference investment node to obtain a target node list; wherein the target node list includes at least one target investment node and a node relationship chain between the target investment nodes; Feature extraction is performed on the target investment data according to the target investment node and the node relationship chain to obtain the target investment feature of the target resource pool.

4. The method according to claim 3, characterized in that The node traversal indication information includes: the number of traversed nodes and the type of traversed nodes; The step of performing node traversal in the investment relationship graph network according to the preset node traversal indication information and the reference investment node to obtain a target node list includes: Extracting associated investment nodes from the investment relationship graph network according to the reference investment nodes and the traversal node categories; Acquire the link between the reference investment node and the associated investment node according to the investment relationship graph network to obtain the node relationship chain; Obtaining the number of associated investment nodes to obtain the number of associated nodes; The associated investment nodes, the reference investment nodes and the node relationship chain are concatenated according to the associated node quantity and the traversed node quantity to obtain the target node list.

5. The method according to any one of claims 1 to 4, characterized in that: The step of selecting a target cluster from preset reference clusters according to the target investment characteristics and preset reference investment characteristics includes: Calculating similarity between the target investment feature and the reference investment feature to obtain target feature similarity; The target cluster is selected from the reference clusters according to the target feature similarity.

6. The method according to any one of claims 1 to 4, characterized in that: Before selecting a target cluster from a preset reference cluster according to the target investment feature and the preset reference investment feature, the method further includes: Constructing the reference cluster specifically includes: Obtain the investment characteristics of the historical resource pool to obtain the historical investment characteristics; Selecting a preliminary investment characteristic from any one of the historical investment characteristics; Constructing clusters according to the preliminary investment characteristics to obtain preliminary clusters; Calculating similarity between the preliminary investment features and the historical investment features to obtain preliminary feature similarity; The preliminary cluster is updated according to the preliminary feature similarity and the historical investment feature to obtain the reference cluster.

7. The method according to claim 6, characterized in that The updating of the preliminary cluster according to the preliminary feature similarity and the historical investment feature to obtain the reference cluster includes: screening selected investment features from the historical investment features based on the preliminary feature similarity; updating the preliminary cluster according to the selected investment characteristics to obtain a candidate cluster; Acquire investment features of the candidate cluster to obtain at least one candidate investment feature; Calculating the feature mean of at least one of the candidate investment features to obtain an updated investment feature; wherein the updated investment feature is the investment feature corresponding to the center of the candidate cluster; The candidate cluster is updated according to the updated investment feature and the historical investment feature to obtain the reference cluster.

8. A resource pool classification device, characterized in that: The device comprises: A data acquisition module is used to acquire target investment data of a target resource pool; wherein the target investment data includes: investor attribute information, investment product information and investment relationship category, wherein the investor attribute information is the attribute information of the investment object, and the investment relationship category represents the relationship category between the investment object and the investment product; A graph network construction module, used to construct a graph network according to the investor attribute information, the investment product information and the investment relationship category to obtain an investment relationship graph network; A feature extraction module, used to extract features from the target investment data according to the investment relationship graph network to obtain target investment features of the target resource pool; A cluster screening module, used to screen out a target cluster from a preset reference cluster according to the target investment characteristics and preset reference investment characteristics; wherein the reference investment characteristics are derived from the reference cluster; A classification module is used to classify the target resource pool according to the target cluster.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the resource pool classification method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the resource pool classification method according to any one of claims 1 to 7 is implemented.

Images