A product object evaluation method, device, equipment and storage medium

Abstract

The present invention discloses a product object evaluation method, device, equipment and storage medium. The method includes: determining the neighbor product objects of the target product object according to the relationship edge data in the product object knowledge graph; obtaining the target attribute data of the target product object and the neighbor attribute data of the neighbor product object from the entity point data of the product object knowledge graph; determining the evaluation information of the target product object according to the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph. The technical solution of the present invention can improve the accuracy of product object evaluation information.

Description

Technical Field

[0001] The embodiments of the present invention relate to the computer field, and in particular to a product object evaluation method, device, equipment and storage medium. Background Art

[0002] Evaluating product objects and obtaining evaluation information of product objects can facilitate users to make choices about product objects. Therefore, how to accurately evaluate product objects is crucial. Summary of the invention

[0003] Embodiments of the present invention provide a product object evaluation method, apparatus, device and storage medium to improve the accuracy of product object evaluation information.

[0004] In a first aspect, an embodiment of the present invention provides a product object evaluation method, comprising:

[0005] Determine the neighbor product objects of the target product object based on the relationship edge data in the product object knowledge graph;

[0006] Acquire target attribute data of the target product object and neighbor attribute data of the neighbor product object from entity point data of the product object knowledge graph;

[0007] Evaluation information of the target product object is determined according to the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph.

[0008] In a second aspect, an embodiment of the present invention further provides a product object evaluation device, comprising:

[0009] A neighbor product object determination module is used to determine the neighbor product objects of the target product object based on the relationship edge data in the product object knowledge graph;

[0010] An attribute data acquisition module, used to acquire target attribute data of the target product object and neighbor attribute data of the neighbor product object from entity point data of the product object knowledge graph;

[0011] An evaluation information determination module is used to determine the evaluation information of the target product object based on the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph.

[0012] In a third aspect, an embodiment of the present invention further provides an electronic device, the electronic device comprising:

[0013] one or more processors;

[0014] A storage device for storing one or more programs;

[0015] When the one or more programs are executed by the one or more processors, the one or more processors implement the product object evaluation method described in any embodiment of the present invention.

[0016] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the product object evaluation method described in any embodiment of the present invention.

[0017] The technical solution provided by the embodiment of the present invention can determine the neighbor product object entity points of the target product object entity point in the product object graph through the relationship edge data in the product object knowledge graph, and can determine the target attribute data of the target product object and the neighbor attribute data of the neighbor product objects according to the entity point data in the product object knowledge graph; and then determine the evaluation information of the target product object according to the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph. The above scheme solves the problem that when evaluating the target product object, only the attribute data of the target product object itself is considered, while ignoring the various associations of the target product object that may affect the evaluation information of the target product object, so that the evaluation information of the target product object is different from the actual situation. This scheme uses the product object knowledge graph to calculate the target attribute data of the target product object in combination with the associations of the target product object when determining the evaluation information of the target product object, thereby forming a more comprehensive evaluation system for the target product object and improving the accuracy of the evaluation information of the target product object. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of a product object evaluation method provided in Embodiment 1 of the present invention;

[0019] Figure 2 A flowchart of a product object evaluation method provided in Embodiment 2 of the present invention;

[0020] Figure 3 a is a schematic diagram of a closed triangle in the product object knowledge graph;

[0021] Figure 3 b is a schematic diagram of an open triangle in the product object knowledge graph;

[0022] Figure 4 A flowchart of a product object evaluation method provided in Embodiment 3 of the present invention;

[0023] Figure 5 A schematic diagram of the structure of a product object evaluation device provided in Embodiment 4 of the present invention;

[0024] Figure 6This is a schematic diagram of the structure of an electronic device provided in Embodiment 5 of the present invention. DETAILED DESCRIPTION

[0025] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0026] Embodiment 1

[0027] Figure 1 This is a flowchart of a product object evaluation method provided in Embodiment 1 of the present invention. This embodiment can be applied to the situation of how to evaluate a product object. The method can be executed by a product object evaluation device provided in an embodiment of the present invention. The device can be implemented in software and / or hardware. The device can be configured in an electronic device. The method specifically includes:

[0028] S110. Determine the neighbor product objects of the target product object based on the relationship edge data in the product object knowledge graph.

[0029] The product object refers to a product that can be selected by the user based on the product evaluation information, such as a fund product. The target product object refers to a product object that has evaluation requirements.

[0030] The product object knowledge graph includes two elements: entity points and relationship edges. Entity points can be used to represent product objects, and relationship edges can be used to represent the relationship between entity points, specifically the relationship between adjacent product objects. For example, the relationship between adjacent product objects can be the relationship of the same manager or the relationship of belonging to the same industry, etc. Optionally, the database format of the product object knowledge graph can be JanusGraph, HugeGraph or other data storage formats.

[0031] Correspondingly, the product object knowledge graph includes two types of data: entity point data and relationship edge data. Among them, the entity point data is the relevant data of the entity point in the product object knowledge graph, which may include the attribute data of the product object corresponding to the entity point; the relationship edge data is the relevant data of the relationship edge in the product object knowledge graph, which may include the relationship attribute data between the product objects associated with the relationship edge. The relationship attribute data may include the common attribute data of the product objects associated with the relationship edge, such as the management party, industry, etc. Attribute data refers to data that can reflect the characteristics and value of the product object. For example, the attribute data of the product object may include but is not limited to the name, scale, sales status, and relevant information of the manager of the product object.

[0032] For example, if the product object is a fund product, the attribute data may include fund product information, fund product manager information, fund product company information, etc. Fund product information mainly includes fund product daily market information, scale data, income data, risk data and other related information describing fund product performance; fund product manager information mainly includes fund product manager work experience, years of work and fund performance and other related information describing the personal ability of fund product managers; fund product company information mainly includes company qualifications, company scale, years of establishment and other information that may affect the quality of fund products and the level of fund product managers.

[0033] Exemplarily, a method for constructing a product object knowledge graph may be to extract sample product objects, relationships between sample product objects, sample attribute data of sample product objects, and relationship attribute data from detailed description information of sample product objects; and construct a product object knowledge graph based on the sample product objects, relationships between sample product objects, sample attribute data of sample product objects, and relationship attribute data.

[0034] Among them, the sample product object refers to the product object used to construct the knowledge graph, and the detailed description information of the sample product object refers to the relevant information describing the sample product object, which may include but is not limited to the basic information of the sample product object, the management information of the sample product object, etc.

[0035] Optionally, sample product objects and the relationships between sample product objects can be extracted from the detailed description information of sample product objects according to entity point and relationship extraction rules. The sample product objects are used as entity points in the product object knowledge graph, and the relationships between sample product objects are used as relationship edges in the product object knowledge graph.

[0036] At the same time, relevant attribute data of the sample product object, i.e., sample attribute data, can be extracted from the detailed description information of the sample product object, and entity point data can be constructed based on the extracted sample attribute data; relationship attribute data can be extracted from the detailed description information of the sample product object based on the relationship between the sample products, and relationship edge data can be constructed based on the extracted relationship attribute data. Optionally, based on the entity points of the sample product object and the relationship edge data between the sample product objects, a basic data triple for generating a product object knowledge graph can be obtained, and the triple is represented by <entity 1, relationship, entity 2>. Among them, entity 1 and entity 2 refer to sample product objects with an association relationship, and the relationship is the relationship edge data between entity 1 and entity 2.

[0037] At this point, the construction of the product object knowledge graph is completed.

[0038] Preferably, after constructing the product object knowledge graph, detailed description information of the product objects can continue to be collected in real time, and the knowledge graph can be updated in real time based on the detailed description information of the product objects collected in real time.

[0039] A neighbor product object refers to a product object that has a relationship edge with the target product object.

[0040] Specifically, the identification information of the target product object can be used as an index to search for the relationship edge data in the constructed product object knowledge graph to determine the neighbor product objects of the target product object. The identification information of the target product object is used to uniquely identify the target product object, which can be the ID of the target product object, or the unique identifier of the target product object in the product object knowledge graph (such as the number of the entity point corresponding to the target product object).

[0041] S120. Obtain target attribute data of the target product object and neighbor attribute data of the neighbor product objects from the entity point data of the product object knowledge graph.

[0042] Specifically, the entity points corresponding to the target product object and the neighbor product objects are determined in the product object knowledge graph. The data contained in the entity points corresponding to the target product object are the target attribute data of the target product object, and the data contained in the entity points corresponding to the neighbor product object are the neighbor attribute data of the neighbor product object.

[0043] Preferably, after obtaining the target attribute data and the neighbor attribute data, the target attribute data and the neighbor attribute data can also be subjected to main attribute data extraction, and unimportant attribute data can be eliminated. The extracted target attribute data and the neighbor attribute data only contain main attribute data. Exemplarily, the extraction method can be principal component analysis. Extracting the main attribute data from the target attribute data and the neighbor attribute data can save computing resources and shorten the operation time when determining the evaluation information of the target product object.

[0044] S130: Determine evaluation information of the target product object according to the target attribute data, neighbor attribute data and / or the topological structure of the product object knowledge graph.

[0045] The evaluation information of the target product object may be an evaluation value calculated according to a preset calculation rule, and the higher the evaluation value, the better the target product object; the evaluation information of the target product object may also be an evaluation grade calculated according to a preset calculation rule, and the evaluation grade may be preset, and the higher the evaluation grade, the better the target product object. The user may select the target product object based on the evaluation information of the target product object.

[0046] Optionally, the target attribute data and the neighbor attribute data may be input into a pre-trained evaluation model, and the evaluation model outputs evaluation information on the target product object.

[0047] Alternatively, the relevant structural data of the target attribute data in the product object knowledge graph can be determined based on the topological structure of the product object knowledge graph; and then the evaluation information of the target product object can be determined based on the relevant structural data and the target attribute data.

[0048] Alternatively, the relevant structural data of the target attribute data in the product object knowledge graph may be determined according to the topological structure of the product object knowledge graph, and the evaluation information of the target product object may be determined based on the relevant structural data, target attribute data, and neighbor attribute data.

[0049] The technical solution provided in this embodiment can determine the neighbor product object entity points of the target product object entity point in the product object graph through the relationship edge data in the product object knowledge graph, and can determine the target attribute data of the target product object and the neighbor attribute data of the neighbor product objects based on the entity point data in the product object knowledge graph; and then determine the evaluation information of the target product object based on the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph. The above scheme solves the problem that when evaluating the target product object, only the attribute data of the target product object itself is considered, while ignoring the various associations of the target product object that may affect the evaluation information of the target product object, so that the evaluation information of the target product object is different from the actual situation. This scheme uses the product object knowledge graph to calculate the target attribute data of the target product object in combination with the associations of the target product object when determining the evaluation information of the target product object, thereby forming a more comprehensive evaluation system for the target product object and improving the accuracy of the evaluation information of the target product object.

[0050] Embodiment 2

[0051] Figure 2 This is a flowchart of a product object evaluation method provided in Embodiment 2 of the present invention. This embodiment is optimized based on the above embodiment and provides an optional embodiment for determining the evaluation information of the target product object. Specifically, Figure 2 As shown, the product object evaluation method provided in this embodiment may include:

[0052] S210. Determine the neighbor product objects of the target product object based on the relationship edge data in the product object knowledge graph.

[0053] S220. Obtain target attribute data of the target product object and neighbor attribute data of the neighbor product objects from the entity point data of the product object knowledge graph.

[0054] S230. Determine basic characteristics of the target product object based on the target attribute data.

[0055] The basic features of the target product object may be related features obtained by processing the target attribute data and can be used to characterize the target product object. Optionally, the basic features may be represented in the form of a matrix or a vector.

[0056] Specifically, the basic features of the target product object are pre-set, and the number and type of the basic features can be set according to the actual application occasions of the target product object. The basic features of the target product object are calculated by using the feature extraction algorithm and taking the target attribute data as the algorithm input.

[0057] Preferably, the basic characteristics of the target product object can be determined based on the characteristics of the target product object itself and the characteristics of the management party. Specifically, this can be achieved through the following sub-steps:

[0058] S2301. Determine the own characteristics according to the private data of the target product object in the target attribute data.

[0059] The private data of the target product object refers to the attribute data of the target product object itself, and the own characteristics are basic characteristics determined based on the private data of the target product object.

[0060] Specifically, the private data of the target product object is extracted from the entity point data of the target product object in the product object knowledge graph. The private data is used as the algorithm input through the feature extraction algorithm to calculate the own features of the target product object.

[0061] S2302. Determine management party characteristics according to management party data of the target product object in the target attribute data.

[0062] The management party data refers to the relevant data of the target product manager and the target product management company. For example, the relevant data of the target product manager can be the target product manager's education, years of work experience, and work efficiency, etc.; the relevant data of the target product management company can be the company size, company reputation, and number of company employees, etc.

[0063] Specifically, the management party data of the target product object is extracted from the entity point data of the target product object in the product object knowledge graph. Through the feature extraction algorithm, the management party data is used as the algorithm input to calculate the management party features of the target product object.

[0064] S2303. Determine the basic characteristics of the target product object based on its own characteristics and the characteristics of the management party.

[0065] Specifically, the acquired self-features and management-side features are summarized, and a matrix A composed of basic features is formed based on the summarized features. Matrix A is the basic features of the determined target product object.

[0066] Preferably, before summarizing the basic features of the target product object, the basic features may be screened to eliminate useless features in the basic features, and the basic features after feature screening are summarized to form a matrix A composed of basic features. The feature screening method may be a chi-square test method, an information gain method, or a correlation coefficient method.

[0067] Exemplarily, if the product object is a fund product, the fund product characteristics include the fund product characteristics and the fund product management characteristics. The fund product management characteristics include the fund product manager characteristics and the fund product management company characteristics. According to the fund product information, the fund product characteristics are obtained through the feature extraction algorithm. The fund product characteristics include but are not limited to: yield, Sharpe ratio, information ratio, volatility, maximum drawdown rate, etc. According to the fund product manager information, the fund product manager characteristics are obtained through the feature extraction algorithm. The fund product manager characteristics include but are not limited to: the manager's education, the manager's years of work, the historical average yield of the fund products managed by the manager, the annualized yield of the fund products currently managed by the manager, etc. According to the fund product company information, the fund product management company characteristics are obtained through the feature extraction algorithm. The fund product management company characteristics include but are not limited to: the company's establishment years, the company's asset size, the number of company managers, the company's yield, etc. The obtained fund product characteristics are subjected to feature screening, and the screened fund product characteristics are summarized to form a matrix A composed of fund product characteristics.

[0068] S240: Determine the neighbor characteristics of the target product object according to the neighbor attribute data.

[0069] Among them, the neighbor features refer to the basic features of the neighbor product objects of the target product object.

[0070] Specifically, the neighbor product objects of the target product object are determined according to the product object knowledge graph, and the neighbor features are pre-set. The number and type of neighbor features can be set according to the actual application scenarios of the neighbor product objects. Through the feature extraction algorithm, the neighbor attribute data is used as the algorithm input to calculate the neighbor features. The neighbor features are summarized to form a matrix B composed of neighbor features. The number of rows of matrix B is the number of neighbor product objects, and the number of columns is the number of neighbor features.

[0071] S250. Determine the graph features of the target product object in the product object knowledge graph according to the topological structure of the product object knowledge graph.

[0072] Among them, graph features refer to feature data that can be used to analyze the network morphology in the product object knowledge graph. Graph features include the degree, centrality and clustering coefficient of the product object.

[0073] Preferably, the graph features of the target product object can be determined based on the topological structure of the product object knowledge graph. Specifically, this can be achieved through the following sub-steps:

[0074] S2501. Determine the degree, centrality and clustering coefficient of the target product object according to the topological structure of the product object knowledge graph.

[0075] Among them, the degree is the degree of the entity point in the product object knowledge graph, that is, the number of relationships of the target product object in the product object knowledge graph. Centrality refers to the importance of the target product object in the product object knowledge graph. The clustering coefficient refers to the degree of aggregation of the corresponding entity points of the target product object in the product object knowledge graph.

[0076] Specifically, the degree of the target product object refers to the total number of edges connecting all entity points corresponding to the target product object in the product object knowledge graph. The total number of edges is the number of relationships between the target product object in the product object knowledge graph. The entity point corresponding to the target product object in the product object knowledge graph is determined, and then the total number of edges corresponding to the entity point of the target product object is further determined based on the topological structure of the product object knowledge graph, thereby obtaining the degree of the target product object. The centrality of the target product object refers to the proportion of the number of paths passing through the entity point corresponding to the target product object among all the shortest edges between two entity points in the product object knowledge graph to the total number of shortest paths. The centrality calculation formula for the target product object is:

[0077]

[0078] Among them, C B is the centrality of the target product object, σ st (v) represents the number of shortest paths from entity point s to entity point t passing through entity point v, σ st Represents the number of all shortest paths from entity point s to entity point t. Entity point v is the entity point corresponding to the target product object, and entity point s and entity point t are any two entity points in the product object knowledge graph.

[0079] The calculation formula of the clustering coefficient of the target product object in the product object knowledge graph is:

[0080]

[0081] Among them, T c represents the number of closed triangles in the product object knowledge graph, T oRepresents the number of open triangles in the product object knowledge graph. Figure 3 As shown, Figure 3 a is a schematic diagram of a closed triangle in the product object knowledge graph; Figure 3 b is a schematic diagram of an open triangle in the product object knowledge graph. A closed triangle means that there are three edges between three entity points, and any two entity points are directly connected by an edge; an open triangle means that there are only two edges between three entity points.

[0082] S2502. Determine the graph features of the target product object in the product object knowledge graph based on at least one of the degree, centrality and clustering coefficient.

[0083] Specifically, the calculated degree, centrality and clustering coefficient are used as the graph features to be selected, and the graph features to be selected are screened to eliminate unnecessary features, and the screened graph features to be selected are used as the graph features in the product object knowledge graph. The graph features are summarized to form a matrix C composed of graph features.

[0084] According to the topological structure of the product object knowledge graph, the graph characteristics of the target product object are determined. The graph characteristics of the target product object can be obtained by quantitatively calculating the degree, centrality and clustering coefficient of the target product object in the product object knowledge graph, and the importance of the target product object in the product object knowledge graph is determined based on the graph characteristics of the target product object. The importance of the target product object in the product object knowledge graph obtained by the degree, centrality and clustering coefficient is used as an indicator for determining the evaluation information of the target product object, which can make the evaluation of the target product object more accurate.

[0085] S260: Determine evaluation information of the target product object based on basic features of the target product object, as well as neighbor features and / or graph features.

[0086] Specifically, the matrix A composed of the basic features of the target product object, the matrix B composed of the neighbor features and / or the matrix C composed of the graph features are summarized to obtain the overall feature data of the target product, and each feature in the overall feature data of the target product is weighted, and the weight of each feature can be preset according to the actual situation. The overall feature data of the target product is then weighted and summed, and the evaluation information of the target product object is determined according to the weighted summation result.

[0087] Exemplarily, basic features of the target product object, as well as neighbor features and / or graph features are fused to obtain evaluation information of the target product object.

[0088] Specifically, the overall feature data of the target product is summarized based on the matrix A composed of the basic features of the target product object, the matrix B composed of the neighbor features, and the matrix C composed of the graph features:

[0089] F=[A,B,C]

[0090] Among them, the matrix F is the overall feature data of the target product. If the matrix F contains n features in total, define the weight matrix W:

[0091] W=[w 1 ,w 2 ,...,w n ]

[0092] Among them, w i It refers to the calculation weight of the feature i contained in the matrix F, and 1≤i≤n. Then the features in the overall feature data of the target product object are weighted and summed to obtain the evaluation information of the target product object. The calculation formula of the evaluation information of the target product object is:

[0093]

[0094] Wherein, E is the calculation result of the evaluation information of the target product object. The larger the value of E, the higher the evaluation value of the target product object, and the better the target product object.

[0095] By fusing the multi-dimensional features of the target product object, the evaluation information of the target product object can be obtained. Each multi-dimensional feature can be weighted according to the importance of the multi-dimensional features of the target product object, and the features in the overall feature data of the target product object can be weighted and summed to obtain the evaluation information of the target product object. The evaluation information of the target product object can be quantified, thereby improving the accuracy of the evaluation information and making the evaluation information more intuitive.

[0096] The technical solution provided in this embodiment performs feature calculation based on the target attribute data to determine the basic features of the target product object, performs feature calculation based on the neighbor attribute data to determine the neighbor features of the target product object, and then determines the graph features based on the topological structure of the product object knowledge graph, and determines the evaluation information of the target product object through the basic features of the target product object, the neighbor features of the target product object, and the graph features. The above scheme aggregates the attribute data of the target product object into multi-dimensional evaluation features of the target product object. When determining the evaluation information of the target product object through the multi-dimensional evaluation features of the target product object, it can solve the problem of a slow calculation process when determining the evaluation information of the target product object through the attribute data of the target product object because the attribute data of the target product object is too complicated; it improves the efficiency of obtaining the evaluation information of the target product object.

[0097] Embodiment 3

[0098] Figure 4 This is a flowchart of a product object evaluation method provided in Embodiment 3 of the present invention. This embodiment is optimized on the basis of the above embodiment, and provides a preferred embodiment of fusing the basic features of each neighbor product object to determine the basic features of each neighbor product object when the number of neighbor product objects is at least two. Specifically, Figure 4 As shown, the product object evaluation method provided in this embodiment may include:

[0099] S310. Determine the neighbor product objects of the target product object based on the relationship edge data in the product object knowledge graph.

[0100] S320. Obtain target attribute data of the target product object and neighbor attribute data of the neighbor product objects from the entity point data of the product object knowledge graph.

[0101] S330. Determine basic characteristics of the target product object based on the target attribute data.

[0102] S340: If the number of neighbor product objects is at least two, determine basic features of each neighbor product object based on attribute data of each neighbor.

[0103] Specifically, when there are two or more neighbor product objects of the target product object, the entity point corresponding to each neighbor product object in the product object knowledge graph, i.e., the neighbor entity point, is determined. The neighbor attribute data of the neighbor product object is determined based on the data of the neighbor entity point. Through the feature extraction algorithm, each neighbor attribute data is used as the algorithm input to calculate the basic features of each neighbor product object.

[0104] S350: Fuse the basic features of each neighbor product object to obtain the neighbor features of the target product object.

[0105] Specifically, the basic features of each neighbor product object are integrated by weighted summation, and the neighbor features of the target product object are determined according to the result of weighted summation. The calculation formula of the neighbor features of the target product object is:

[0106]

[0107] where w k represents the weight coefficient of the kth neighbor product object on feature i; F (i,k) Indicates the value of the kth neighbor product object on the i-th feature, the result of weighted summation B iis the value of all neighbor product objects on feature i. The fused neighbor features are summarized to form a matrix B composed of neighbor features. The number of rows in matrix B is the number of neighbor product objects, and the number of columns is the number of neighbor features.

[0108] S360. Determine the graph features of the target product object in the product object knowledge graph according to the topological structure of the product object knowledge graph.

[0109] S370: Determine evaluation information of the target product object based on basic features of the target product object, as well as neighbor features and / or graph features.

[0110] The technical solution provided in this embodiment, when the number of neighbor product objects of the target product object is at least two, determines the attributes of each neighbor product through the product object knowledge graph, and determines the basic characteristics of each neighbor product object based on the attribute data of each neighbor, fuses and calculates the basic characteristics of each neighbor product object to obtain the neighbor characteristics of the target product object, and determines the evaluation information of the target product object based on the neighbor characteristics of the target product object obtained by the fusion calculation. The above scheme solves the problem that when the number of neighbor product objects of the target product object is too large, the evaluation information of the target product object is determined only by the neighbor characteristics of one neighbor product object, which will lead to inaccurate evaluation information due to incomplete information. This scheme realizes the effect of obtaining the neighbor characteristics by fusing the basic characteristics of each neighbor product object, comprehensively considering the influence of the basic characteristics of each neighbor product object on the evaluation information of the target product, and improving the accuracy of the evaluation information of the target product.

[0111] Embodiment 4

[0112] Figure 5 This is a schematic diagram of the structure of a product object evaluation device provided in Embodiment 4 of the present invention. This embodiment is applicable to situations where product objects are evaluated, such as Figure 5 As shown, the product object evaluation device includes: a neighbor product object determination module 410, an attribute data acquisition module 420 and an evaluation information determination module 430.

[0113] Wherein, the neighbor product object determination module 410 is used to determine the neighbor product objects of the target product object according to the relationship edge data in the product object knowledge graph;

[0114] The attribute data acquisition module 420 is used to acquire target attribute data of the target product object and neighbor attribute data of the neighbor product object from the entity point data of the product object knowledge graph;

[0115] The evaluation information determination module 430 is used to determine the evaluation information of the target product object based on the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph.

[0116] The technical solution provided in this embodiment can determine the neighbor product object entity points of the target product object entity point in the product object graph through the relationship edge data in the product object knowledge graph, and can determine the target attribute data of the target product object and the neighbor attribute data of the neighbor product objects based on the entity point data in the product object knowledge graph; and then determine the evaluation information of the target product object based on the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph. The above scheme solves the problem that when evaluating the target product object, only the attribute data of the target product object itself is considered, while ignoring the various associations of the target product object that may affect the evaluation information of the target product object, so that the evaluation information of the target product object is different from the actual situation. This scheme uses the product object knowledge graph to calculate the target attribute data of the target product object in combination with the associations of the target product object when determining the evaluation information of the target product object, thereby forming a more comprehensive evaluation system for the target product object and improving the accuracy of the evaluation information of the target product object.

[0117] The evaluation information determination module 430 further includes:

[0118] A basic feature determination unit, used to determine the basic features of the target product object according to the target attribute data;

[0119] A neighbor feature determination unit, used to determine the neighbor features of the target product object according to the neighbor attribute data;

[0120] A graph feature determination unit, used to determine the graph features of the target product object in the product object knowledge graph according to the topological structure of the product object knowledge graph;

[0121] The target product object evaluation unit is used to determine the evaluation information of the target product object according to the basic characteristics of the target product object, as well as the neighbor characteristics and / or the graph characteristics.

[0122] Among them, the above basic characteristics are also specifically used for:

[0123] Determine the own characteristics according to the private data of the target product object in the target attribute data;

[0124] Determine management party characteristics according to management party data of the target product object in the target attribute data;

[0125] Determine the basic characteristics of the target product object based on its own characteristics and the characteristics of the management party.

[0126] Furthermore, the neighbor feature determination unit is further specifically used for:

[0127] If the number of neighbor product objects is at least two, then the neighbor characteristics of the target product object are determined according to the neighbor attribute data, including:

[0128] Determine the basic characteristics of each neighbor product object based on the attribute data of each neighbor;

[0129] The basic features of each neighbor product object are fused to obtain the neighbor features of the target product object.

[0130] Exemplarily, the above-mentioned graph feature determination unit is also specifically used for:

[0131] According to the topological structure of the product object knowledge graph, the degree, centrality and clustering coefficient of the target product object are determined;

[0132] According to at least one of the degree, centrality and clustering coefficient, a graph feature of the target product object in the product object knowledge graph is determined.

[0133] Furthermore, the target product object evaluation unit is also specifically used for:

[0134] The basic features of the target product object, as well as the neighbor features and / or graph features are fused to obtain evaluation information of the target product object.

[0135] Exemplarily, the product object evaluation device provided in this embodiment further includes:

[0136] A sample product object extraction module, used to extract sample product objects, relationships between sample product objects, sample attribute data of sample product objects, and relationship attribute data from detailed description information of sample product objects;

[0137] The product object knowledge graph construction module is used to construct a product object knowledge graph based on sample product objects, the relationships between sample product objects, sample attribute data of sample product objects, and relationship attribute data.

[0138] The product object evaluation device provided in this embodiment can be applied to the product object evaluation method provided in any of the above embodiments, and has corresponding functions and beneficial effects.

[0139] Embodiment 5

[0140] Figure 6 A schematic diagram of the structure of an electronic device provided in Embodiment 5 of the present invention is shown in FIG. Figure 6 As shown, the electronic device includes a processor 50, a memory 51, an input device 52 and an output device 53; the number of processors 50 in the electronic device can be one or more. Figure 6A processor 50 is taken as an example; the processor 50, the memory 51, the input device 52 and the output device 53 in the electronic device can be connected by a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0141] The memory 51 is a computer-readable storage medium that can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the product object evaluation method in the embodiment of the present invention. The processor 50 executes various functional applications and data processing of the electronic device by running the software programs, instructions and modules stored in the memory 51, that is, implements the above-mentioned product object evaluation method.

[0142] The memory 51 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 51 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 51 may further include a memory remotely arranged relative to the processor 50, and these remote memories may be connected to the electronic device 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.

[0143] The input device 52 may be used to receive input product object attribute data and generate product object evaluation related parameter input related to user settings and function control of the electronic device. The output device 53 may include a display device such as a display screen.

[0144] The electronic device provided in this embodiment can be applied to the product object evaluation method provided in any of the above embodiments, and has corresponding functions and beneficial effects.

[0145] Embodiment 6

[0146] Embodiment 6 of the present invention further provides a storage medium including computer executable instructions, wherein the computer executable instructions are used to execute a product object evaluation method when executed by a computer processor, the method comprising:

[0147] Determine the neighbor product objects of the target product object based on the relationship edge data in the product object knowledge graph; obtain the target attribute data of the target product object and the neighbor attribute data of the neighbor product objects from the entity point data of the product object knowledge graph; determine the evaluation information of the target product object based on the target attribute data, as well as the neighbor attribute data and / or the topological structure of the product object knowledge graph.

[0148] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0149] It is worth noting that in the above embodiments, the various units and modules included are only divided according to functional logic, but are not limited to the above divisions, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0150] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A product object evaluation method, characterized in that: include: Determine the neighbor product objects of the target product object based on the relationship edge data in the product object knowledge graph; Acquire target attribute data of the target product object and neighbor attribute data of the neighbor product object from entity point data of the product object knowledge graph; Determining evaluation information of the target product object according to the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph; Wherein, determining the evaluation information of the target product object according to the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph includes: Determining basic characteristics of the target product object according to the target attribute data; Determining neighbor characteristics of the target product object according to the neighbor attribute data; According to the topological structure of the product object knowledge graph, determine the graph features of the target product object in the product object knowledge graph; wherein the graph features are feature data used to analyze the network morphology in the product object knowledge graph, and the graph feature data include the degree, centrality and clustering coefficient of the product object; Evaluation information of the target product object is determined according to the basic features of the target product object, as well as the neighbor features and / or the graph features.

2. The method according to claim 1, characterized in that Determining the basic characteristics of the target product object according to the target attribute data includes: Determining the own characteristics according to the private data of the target product object in the target attribute data; Determining management party characteristics according to management party data of the target product object in the target attribute data; The basic characteristics of the target product object are determined according to the own characteristics and the management party characteristics.

3. The method according to claim 1, characterized in that If the number of the neighbor product objects is at least two, determining the neighbor characteristics of the target product object according to the neighbor attribute data includes: Determine the basic characteristics of each neighbor product object based on the attribute data of each neighbor; The basic features of each neighbor product object are fused to obtain the neighbor features of the target product object.

4. The method according to claim 1, characterized in that: The determining, according to the topological structure of the product object knowledge graph, the graph features of the target product object in the product object knowledge graph includes: Determining the degree, centrality and clustering coefficient of the target product object according to the topological structure of the product object knowledge graph; According to at least one of the degree, the centrality and the clustering coefficient, a graph feature of the target product object in the product object knowledge graph is determined.

5. The method according to claim 1, characterized in that Determining the evaluation information of the target product object according to the basic features of the target product object, the neighbor features and / or the graph features includes: The basic features of the target product object, the neighbor features and / or the graph features are fused to obtain evaluation information of the target product object.

6. The method according to claim 1, characterized in that Also includes: extracting sample product objects, relationships between sample product objects, sample attribute data of sample product objects, and relationship attribute data from detailed description information of sample product objects; The product object knowledge graph is constructed based on the sample product objects, the relationships between the sample product objects, the sample attribute data of the sample product objects, and the relationship attribute data.

7. A product object evaluation device, characterized in that: include: A neighbor product object determination module is used to determine the neighbor product objects of the target product object based on the relationship edge data in the product object knowledge graph; An attribute data acquisition module, used to acquire target attribute data of the target product object and neighbor attribute data of the neighbor product object from entity point data of the product object knowledge graph; An evaluation information determination module, configured to determine the evaluation information of the target product object according to the target attribute data, the neighbor attribute data and / or the topological structure of the product object knowledge graph; Wherein, the evaluation information determination module includes: A basic feature determination unit, used to determine the basic features of the target product object according to the target attribute data; A neighbor feature determination unit, configured to determine the neighbor feature of the target product object according to the neighbor attribute data; A graph feature determination unit, configured to determine the graph features of the target product object in the product object knowledge graph according to the topological structure of the product object knowledge graph; wherein the graph features are feature data for analyzing the network morphology in the product object knowledge graph, and the graph feature data include the degree, centrality and clustering coefficient of the product object; The target product object evaluation unit is used to determine the evaluation information of the target product object according to the basic characteristics of the target product object, the neighbor characteristics and / or the graph characteristics.

8. An electronic device, characterized in that: The electronic device comprises: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the product object evaluation method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the product object evaluation method as described in any one of claims 1 to 6 is implemented.

Images