Object classification method and device, electronic equipment and storage medium

Abstract

The invention provides an object classification method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, in particular to the technical field of big data. The specific implementation scheme is as follows: determining a plurality of first objects from a plurality of to-be-processed objects based on a first initial relationship; wherein the first initial relationship represents an interactive behavior relationship between each to-be-processed object and the target platform; determining a plurality of second objects from the plurality of to-be-processed objects based on the second initial relationship; wherein the second initial relation represents the relation between the to-be-processed objects and the attributes of the to-be-processed objects; clustering the plurality of first objects and the plurality of second objects, classifying the plurality of first objects and the plurality of second objects into at least two target groups, and executing the same interaction operation on each target object in the same target group; the target objects in each target group and the target platform have similar interaction behaviors.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, in particular to the field of big data technology, and specifically to an object classification method, device, electronic device, and storage medium. Background Art

[0002] With the popularization of e-commerce, the number of users participating in online transactions continues to increase, and the complexity of the relationships between users is also increasing, and users' attention to the security of online transactions is also increasing. Summary of the invention

[0003] The present disclosure provides an object classification method, device, electronic device and storage medium.

[0004] According to one aspect of the present disclosure, an object classification method is provided, comprising: based on a first initial relationship, determining a plurality of first objects from a plurality of objects to be processed; wherein the first initial relationship characterizes an interactive behavior relationship between each object to be processed and a target platform; based on a second initial relationship, determining a plurality of second objects from a plurality of objects to be processed; wherein the second initial relationship characterizes a relationship between each object to be processed and an attribute of each object to be processed; by clustering the plurality of first objects and the plurality of second objects, the plurality of first objects and the plurality of second objects are classified into at least two target groups, so as to perform the same interactive operation on each target object in the same target group; each target object in each target group has similar interactive behavior with the target platform.

[0005] According to another aspect of the present disclosure, there is provided an object classification device, comprising: a first determination module, a second determination module and a clustering module. The first determination module is used to determine a plurality of first objects from a plurality of objects to be processed based on a first initial relationship; wherein the first initial relationship characterizes the interactive behavior relationship between each object to be processed and a target platform. The second determination module is used to determine a plurality of second objects from a plurality of objects to be processed based on a second initial relationship; wherein the second initial relationship characterizes the relationship between each object to be processed and the attributes of each object to be processed. The clustering module is used to classify a plurality of first objects and a plurality of second objects into at least two target groups by clustering the plurality of first objects and the plurality of second objects, so as to perform the same interactive operation on each target object in the same target group; each target object in each target group has similar interactive behavior with the target platform.

[0006] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as described above.

[0007] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the method as described above.

[0008] According to another aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein the computer program implements the method described above when executed by a processor.

[0009] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0011] Figure 1 An exemplary system architecture to which the object classification method and apparatus according to an embodiment of the present disclosure can be applied is schematically shown;

[0012] Figure 2 A flowchart of an object classification method according to an embodiment of the present disclosure is schematically shown;

[0013] Figure 3 A schematic diagram schematically shows clustering based on weak relationships between objects to be processed according to an embodiment of the present disclosure;

[0014] Figure 4 A schematic diagram schematically shows clustering based on strong relationships between objects to be processed according to an embodiment of the present disclosure;

[0015] Figure 5 A schematic diagram schematically shows object classification based on clustering results obtained by classifying and clustering strong and weak relationships between objects to be processed according to an embodiment of the present disclosure;

[0016] Figure 6 A block diagram schematically shows an object classification device according to an embodiment of the present disclosure; and

[0017] Figure 7 A block diagram of an electronic device suitable for implementing an object classification method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0018] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0019] With the popularity of e-commerce stores, in order to protect the transaction security of online transaction users, it is necessary to conduct regular risk control management on the stores on the online trading platform. However, with the surge in the number of online stores, the relationships between online stores are complicated, which brings certain difficulties to the implementation of effective risk control management measures.

[0020] For example, when the association strength is expressed by weight, it is difficult to accurately classify online stores due to the uneven association strength. Moreover, due to the complex relationships between online stores, even if a complex clustering algorithm is executed based on weights, there is still a lot of interference information that affects the execution of the clustering algorithm. Therefore, not only is the processing efficiency low, but the classification accuracy is also poor.

[0021] In view of this, the present disclosure provides an object classification method, which processes initial relationships with different association strengths separately, and then clusters the first object and the second object obtained by respectively obtaining at least two target groups, so as to perform the same interaction operation on each target object in the same target group, thereby reducing the impact of differences in association strength on clustering and further improving the efficiency and accuracy of classification processing.

[0022] Figure 1 An exemplary system architecture to which the object classification method and apparatus according to an embodiment of the present disclosure can be applied is schematically shown.

[0023] It should be noted that Figure 1 The examples shown are only examples of system architectures to which the embodiments of the present disclosure can be applied, in order to help those skilled in the art understand the technical content of the present disclosure, but do not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios. For example, in another embodiment, the exemplary system architecture to which the object classification method and apparatus can be applied may include a terminal device, but the terminal device may implement the object classification method and apparatus provided by the embodiments of the present disclosure without interacting with the server.

[0024] like Figure 1As shown, the system architecture 100 according to this embodiment may include a terminal device 101, a network 102 and a server 103. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired and / or wireless communication links, etc.

[0025] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as knowledge reading applications, web browser applications, search applications, instant messaging tools, email clients and / or social platform software, etc. (only as examples).

[0026] The terminal device 101 may be any electronic device having a display screen and supporting web browsing, including but not limited to a smart phone, a tablet computer, a laptop computer, a desktop computer, and the like.

[0027] The server 103 may be a server that provides various services, such as a background management server (only as an example) that provides support for the content browsed by the user using the terminal device 101. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to the user request) to the terminal device.

[0028] It should be noted that the object classification method provided in the embodiment of the present disclosure can generally be executed by the terminal device 101. Accordingly, the object classification device provided in the embodiment of the present disclosure can also be set in the terminal device 101.

[0029] Alternatively, the object classification method provided in the embodiment of the present disclosure may also be generally performed by the server 103. Accordingly, the object classification device provided in the embodiment of the present disclosure may generally be arranged in the server 103. The object classification method provided in the embodiment of the present disclosure may also be performed by a server or server cluster that is different from the server 103 and can communicate with the terminal device 101 and / or the server 103. Accordingly, the object classification device provided in the embodiment of the present disclosure may also be arranged in a server or server cluster that is different from the server 103 and can communicate with the terminal device 101 and / or the server 103.

[0030] For example, the terminal device 101 can collect the interaction information between each object to be processed and the target platform, and send the interaction information to the server 103 through the network 102. The server 103 constructs an association spectrum 110 between each object to be processed based on the interaction information, and classifies each processing object into multiple groups 120 by executing the object classification method provided by the embodiment of the present disclosure, for example: group G1, group G2, group G3, ..., etc. The objects in each group have similar interaction behaviors with the target platform. The server 103 can feed back the classification results to the terminal device 101 so as to perform the same interaction operation on each object in the same group.

[0031] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision, disclosure and application of user personal information involved comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good morals.

[0032] In the technical solution of the present disclosure, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

[0033] Figure 2 The flowchart of the object classification method according to the embodiment of the present disclosure is schematically shown.

[0034] like Figure 2 As shown, the method 200 includes operations S210 to S230.

[0035] In operation S210 , a plurality of first objects are determined from a plurality of objects to be processed based on a first initial relationship.

[0036] In operation S220 , a plurality of second objects are determined from the plurality of objects to be processed based on the second initial relationship.

[0037] In operation S230 , the plurality of first objects and the plurality of second objects are classified into at least two target groups by clustering the plurality of first objects and the plurality of second objects, so as to perform the same interactive operation on each target object in the same target group.

[0038] According to an embodiment of the present disclosure, the target platform may be an online transaction platform, and the object to be processed may be a store registered on the target platform for online transactions.

[0039] According to the embodiments of the present disclosure, the first initial relationship represents the interactive behavior relationship between each object to be processed and the target platform. For example, when two shops conduct online transactions on the target platform, they use similar user avatars and user names, or use the same IP address of the same network segment during the same period, or use the same device for online transactions, etc., indicating that there is a first initial relationship between the two shops.

[0040] According to an embodiment of the present disclosure, the second initial relationship represents the relationship between each object to be processed and its attributes, including but not limited to the registered address, name, account, legal person, etc. of the store.

[0041] For example: when a shop is registered on an online trading platform, it is registered in the name of Company A, and the registered address is No. 001, Building C, District B, City A. Then the shop has a second initial relationship with Company A and "No. 001, Building C, District B, City A".

[0042] According to an embodiment of the present disclosure, the strength of association between the objects to be processed based on the first initial relationship association is relatively weak, and the strength of association between the objects to be processed based on the second initial relationship association is relatively strong. Strong association and weak association are relative, and the definition of strong and weak relationships depends on the degree of influence on the classification of the objects to be processed.

[0043] For example, when Shop S1 and Shop S2 share the same registration address to register on the online platform, it can be determined that Shop S1 and Shop S2 are operated by the same team. At this time, Shop S1 and Shop S2 have a strong correlation. When the similarity of the IP addresses used by Shop S3 and Shop S4 in the same period is 70%, it can only be inferred that the operating teams of Shop S3 and Shop S4 may be the same. At this time, Shop S3 and Shop S4 have a weak correlation.

[0044] Since the objects to be processed with different association strengths of the initial relationship are processed separately first, the influence of the difference in association strength on the clustering process can be effectively reduced. When clustering the multiple first objects and the multiple second objects, there is no need to introduce weights to represent the association strength, which reduces the computational difficulty of the clustering process, and the multiple first objects and the multiple second objects are classified into at least two target groups, so as to perform the same interactive operation on each target object in the same target group.

[0045] According to the embodiments of the present disclosure, each target object in each target group has similar interactive behaviors with the target platform. For example, each store in the same target group may be affiliated with a company under the same legal person, and use the same network IP address for online transactions during operation. When abnormal online transaction behavior is detected in any of the stores, the same interactive operation can be performed on all the stores in the target group, such as reducing the daily online transaction limit or suspending online transactions, etc., to protect the transaction security of online transaction users to the greatest extent.

[0046] According to an embodiment of the present disclosure, initial relationships with different association strengths are processed separately, and then the first objects and the second objects obtained by clustering are clustered again to obtain at least two target groups, so as to perform the same interactive operation on each target object in the same target group. This reduces the impact of differences in association strength on clustering, and further improves the efficiency and accuracy of classification processing.

[0047] Reference below Figure 3~Figure 5 , combined with specific embodiments Figure 2 The method shown is further explained.

[0048] According to an embodiment of the present disclosure, the above operation S210 may include the following operations: constructing a first target graph based on the first initial relationship; and clustering a plurality of first object nodes based on the first target graph to determine a plurality of first objects.

[0049] According to an embodiment of the present disclosure, the first target graph includes a plurality of first object nodes representing similar interactive behaviors. For example, the first target graph may be an isomorphic graph constructed with object identifiers as nodes. The first target graph may be an undirected graph, which is not specifically limited in the embodiment of the present disclosure. The edges between object nodes may represent similar interactive behaviors between object nodes. For example, each object node may represent a shop ID. When the registered user names between two shops with different IDs are the same, an attribute edge may be constructed between the two shop IDs, and the attribute of the edge is the user name.

[0050] Since the relationship between the first object nodes in the first target graph is relatively weak, when clustering the first object nodes based on the first target graph, the first target graph may be pruned based on a predetermined rule. The predetermined rule may be determined based on the degree of influence of the edge attribute on the clustering.

[0051] For example, when registering shops on an electronic trading platform, many unrelated shops may share similar names. Therefore, when two object nodes are connected by user names as edges, the relationship of such attribute edges can be deleted or ignored before or during clustering, and the edges will not participate in the clustering calculation, thereby further improving the clustering efficiency.

[0052] According to an embodiment of the present disclosure, by constructing a first target graph with multiple first object nodes having similar interactive behaviors, since the association strengths between the first object nodes in the graph are similar, the impact of excessive differences in association strengths between the object nodes on clustering accuracy can be reduced, thereby further improving clustering accuracy and clustering efficiency.

[0053] According to an embodiment of the present disclosure, constructing a first target graph based on the first initial relationship may include the following operations: obtaining the interaction behavior information between each object to be processed and the target platform within a predetermined time period; calculating the behavior similarity between each object to be processed based on the interaction behavior information; and using the identifier of each object to be processed as the first object node, constructing edges between each object to be processed based on the behavior similarity to obtain the first target graph.

[0054] According to an embodiment of the present disclosure, the interactive behavior information may include multiple types of interactive behaviors in different dimensions, for example, the interactive behavior types in the time dimension may include: store business hours, store product update time, etc. The interactive behavior types in the device dimension may include: interactive device ID, IP address, etc. used by the store operator. The interactive behavior types in the attribute dimension may include: user name, avatar, etc. of the store operator.

[0055] According to an embodiment of the present disclosure, the behavior similarities between the objects to be processed may be calculated respectively according to the types of interactive behaviors.

[0056] For example, the IP addresses used by shop ID1 to operate the shop during the period t1 to t3 include: IP1 to IP7, and the IP addresses used by shop ID3 to operate the shop during the period t1 to t3 include: IP1, IP5 to IP7. The similarity algorithm can be used to calculate the similarity of the IP addresses. When the similarity of the IP addresses is greater than or equal to a preset similarity threshold, it can be determined that there is an association relationship between shop ID1 and shop ID3 based on the similarity of the IP addresses.

[0057] Figure 3 The schematic diagram schematically shows clustering based on weak relationships between objects to be processed according to an embodiment of the present disclosure.

[0058] like Figure 3As shown, the embodiment 300 includes a first target graph 311 and a first clustering result 312. The first target graph 311 includes 7 object nodes representing shop ID1 to shop ID7. Shop ID1 and shop ID3, and shop ID6 and shop ID7 are connected with edges of IP attributes, indicating similar IP addresses used during the operation of the shops. Shop ID3 and shop ID6 are connected with an edge of business hours attributes, indicating that the business hours of shop ID3 and shop ID6 are similar. Shop ID1 and shop ID5 are connected with an edge of update time attributes, indicating that the product update time of shop ID1 and shop ID5 is similar. Shop ID4 and shop ID5 are connected with an edge of user avatar attributes, indicating that the user avatars of the operators of shop ID4 and shop ID5 are similar. Shop ID4 and shop ID2 are connected with an edge of user name attributes, indicating that the user names of the operators of shop ID4 and shop ID2 are similar. Shop ID4 and Shop ID7 are connected by an edge with both the user avatar attribute and the user name attribute, indicating that the operators of Shop ID4 and Shop ID7 have both similar user names and similar user avatars.

[0059] According to an embodiment of the present disclosure, based on the first target graph, multiple objects to be processed are clustered to determine multiple first objects, which may include the following operations: based on the first target graph, the modularity change information is calculated for each candidate clustering scheme, and an intermediate clustering scheme is determined from the candidate clustering schemes; based on the intermediate clustering scheme, each intermediate object is clustered to obtain an optimized relationship; and based on the optimized relationship, the operation of calculating the modularity change information for each candidate clustering scheme is returned to execute until the modularity change information converges to obtain multiple first objects.

[0060] According to an embodiment of the present disclosure, modularity represents the difference in similarity between clustered objects and non-clustered objects in clustering results for at least two candidate clustering solutions.

[0061] For example, candidate clustering scheme A may be to treat each object node as a community, and candidate clustering scheme B may be to merge each object into the community where its associated neighbor nodes are located. Then, the change information of modularity can be calculated according to formula (1):

[0062] (1)

[0063] Among them, k i,in It represents the number of edges between node i and all nodes in the community where the neighboring node to be moved to belongs; represents the number of edges connecting node i with all nodes in the first target graph; m represents the number of edges remaining in the first target graph after removing the edges within each community.

[0064] When the calculation result of formula (1) is greater than 0, it means that the clustering degree of candidate clustering scheme B is higher. Therefore, candidate clustering scheme B can be used as an intermediate clustering scheme, and so on.

[0065] According to an embodiment of the present disclosure, by continuously iterating and calculating the change information of the modularity of different candidate clustering schemes, when the change information of the modularity converges, it indicates that the iteration stops, which means that the clustering result is stable and multiple first objects after clustering can be obtained.

[0066] According to an embodiment of the present disclosure, the convergence of the modularity change information may be that the calculation result of formula (1) tends to a stable value.

[0067] For example: based on the change information of modularity, the first target graph 311 is clustered to obtain the first clustering result 312, which may include: multiple first objects "shop ID1, shop ID3, shop ID4, shop ID6 and shop ID7" and the association relationship between the first objects.

[0068] According to the embodiments of the present disclosure, by clustering and optimizing the object nodes in the first target graph, the association strength between the object nodes is further improved, and the difference in association strength between the object nodes with strong relationships is reduced.

[0069] According to an embodiment of the present disclosure, the above operation S220 may include the following operations: constructing a second target graph based on the second initial relationship; clustering a plurality of second object nodes based on the second target graph to determine a plurality of second objects.

[0070] According to an embodiment of the present disclosure, the second target graph may be a heterogeneous graph constructed with object nodes and attribute nodes. The second target graph includes a plurality of second object nodes and a plurality of attribute nodes representing associations. An edge between an object node and an attribute node in the second target graph indicates that the object node has the attribute.

[0071] For example: the registered address of shop ID1 is Building B, District A. In the second target graph, shop ID1 can be an object node, “Building B, District A” can be an attribute node, and an edge is connected between shop ID1 and “Building B, District A”.

[0072] According to an embodiment of the present disclosure, constructing a second target graph based on the second initial relationship may include the following operations: obtaining platform registration information of each object to be processed and the target platform within a predetermined period of time; extracting attribute information of each object to be processed from the platform registration information; and constructing a second target graph with each object to be processed and each attribute information as nodes and the corresponding relationship between each object to be processed and each attribute information as edges.

[0073] According to an embodiment of the present disclosure, the platform registration information may include information that a store on the e-commerce platform needs to provide to the platform if it wants to obtain operating qualifications, such as: company name, legal person, registered address, account, etc.

[0074] Figure 4 A schematic diagram schematically shows clustering based on the strong relationship between objects to be processed according to an embodiment of the present disclosure.

[0075] like Figure 4 As shown, in this embodiment 400 , a second target graph 411 and a second clustering result 412 may be included.

[0076] In the second target graph 411, shop ID3 has edges with company Co1 and registered address Ad1, indicating that the registered company of shop ID3 is "Co1" and the registered address of the company is "Ad1". Shop ID1 has edges with registered address Ad1, account Ac1 and legal person Lp1, indicating that the registered address of shop ID1 is "Ad1", the transaction account information is "Ac1" and the legal person of the registered company is "Lp1". Shop ID2 has edges with legal person Lp2 and company Co2, indicating that the registered company of shop ID2 is "Co2" and the legal person of the registered company is "Lp2". Shop ID4 has an edge with legal person Lp1, indicating that the legal person of the registered company of shop ID4 is "Lp1".

[0077] It should be noted that, usually when a store registers on an e-commerce platform, the above attribute information is a required item. In order to clearly show the relationship between the nodes, individual attribute nodes that are not associated with other nodes are omitted in the second target graph 411.

[0078] According to an embodiment of the present disclosure, based on the second target graph, multiple object nodes are clustered to determine multiple second objects, which may include the following operations: traversing the multiple second object nodes and multiple attribute nodes in the second target graph in sequence; and clustering at least two second object nodes associated with the same attribute node to obtain multiple second objects.

[0079] For example, in the second target graph 411, both shop ID3 and shop ID1 have the same registered address as Ad1. Therefore, when clustering, shop ID3 and shop ID1 can be clustered into one category, indicating that there is a strong association between shop ID3 and shop ID1. Similarly, a second clustering result 412 is obtained. The second clustering result 412 may include shop ID1, shop ID3, and shop ID4 that have a strong association, as well as the association relationship between the shops.

[0080] According to an embodiment of the present disclosure, since the credibility of the platform registration information is relatively high, based on the second target graph, a weak component clustering algorithm can be used to cluster the second object nodes and the attribute nodes, thereby improving the calculation speed of the clustering process.

[0081] According to an embodiment of the present disclosure, the above-mentioned operation S230 may include the following operations: constructing a third target graph with each first object and each second object as nodes, and with the relationship between each first object and the relationship between each second object as edges; and using the identifier of each target object as the initial label of the node, updating the initial label of each node by reading the initial categories of each neighbor node associated with each node until the maximum number of updates is reached, thereby obtaining at least two groups.

[0082] Figure 5 A schematic diagram schematically shows object classification based on clustering results obtained by classifying and clustering strong and weak relationships between objects to be processed according to an embodiment of the present disclosure.

[0083] like Figure 5 As shown, the embodiment 500 includes a third target graph 531 and a third clustering result 532 .

[0084] According to an embodiment of the present disclosure, in the third target graph 531, an isomorphic graph is constructed with objects as nodes and the relationships between objects as edges. Since the third target graph 531 is constructed based on the first clustering result 312 and the second clustering result 412 described above, the difference between the optimized weak relationship and the strong relationship has been reduced. In the case of a large amount of data, the relationship between the objects in the third target graph 531 can be further optimized based on the change in modularity to obtain a third clustering result 532. The third clustering result 532 can also be obtained by initializing the labels of each object node and performing clustering using the LPA algorithm.

[0085] According to an embodiment of the present disclosure, updating the initial label of each node by reading the initial categories of each neighbor node associated with each node may include the following operations: obtaining the number of each initial category by reading the initial categories of each neighbor node associated with each node; and updating the initial label of each node to the initial category with the largest number.

[0086] For example: First, initialize the labels of all nodes in the third target graph 531, which may be a store ID.

[0087] Then, in each round of iteration, all nodes can be traversed and each node can be labeled. For example, for shop ID1, the voting results of all neighboring nodes associated with "shop ID1" on the node label of "shop ID1" can be counted. If the label with the most votes is different from the current node label of "shop ID1", the node label of "shop ID1" will be updated to the label with the most votes.

[0088] According to the present disclosure, the voting process of the neighbor node for the node label of "Shop ID1" may include: obtaining the community labels of all neighbor nodes associated with the "Shop ID1" node, and the community label may be a category label, that is, the types of shops in the same community are the same. For example: the same operating model. By reading the community labels of all neighbor nodes, the same community labels are accumulated and the community label with the largest number of votes is calculated. When the cumulative results of multiple community labels are the same, the community label with the largest number of votes can be randomly determined to update the label of Shop ID1.

[0089] In the third clustering result 532, each object to be processed is divided into group G1 and group G2. Group G1 includes shop ID1, shop ID3 and shop ID4. Group G2 includes shop ID6 and shop ID7.

[0090] The same or similar interactive operations may be adopted for each object in the same group. The interactive operations may include positive recommendation operations or reverse control operations.

[0091] For example, group G1 represents a group of shops with good credit. In the product recommendation page, shops ID1, ID3 and ID4 can be ranked at the top.

[0092] For example, group G2 represents a group of shops with poor credit. When abnormal transaction behavior is detected in shop ID6, transaction control operations are performed on shop ID6 and shop ID7 at the same time, such as suspending transactions.

[0093] According to an embodiment of the present disclosure, based on the first clustering result and the second clustering result, a third target graph is constructed with object nodes, and then the third target graph is clustered by clustering the target objects, thereby reducing the impact of the differences between the strong and weak associations between the target objects on the clustering, thereby eliminating the need to use reference weights to distinguish between strong and weak relationships, simplifying the complexity of clustering calculations, and improving clustering efficiency and clustering accuracy.

[0094] Figure 6 A block diagram of an object classification device according to an embodiment of the present disclosure is schematically shown.

[0095] like Figure 6As shown, the object classification device 600 includes a first determination module 610 , a second determination module 620 and a clustering module 630 .

[0096] The first determination module 610 is used to determine a plurality of first objects from a plurality of objects to be processed based on a first initial relationship; wherein the first initial relationship represents an interactive behavior relationship between each object to be processed and a target platform.

[0097] The second determination module 620 is used to determine a plurality of second objects from a plurality of objects to be processed based on a second initial relationship; wherein the second initial relationship represents a relationship between each object to be processed and an attribute of each object to be processed.

[0098] The clustering module 630 is used to classify the multiple first objects and the multiple second objects into at least two target groups by clustering the multiple first objects and the multiple second objects, so as to perform the same interactive operation on each target object in the same target group; each target object in each target group has similar interactive behavior with the target platform.

[0099] According to an embodiment of the present disclosure, the first determining module may include a first constructing submodule and a first clustering submodule.

[0100] The first construction submodule is used to construct a first target graph based on the first initial relationship; wherein the first target graph includes a plurality of first object nodes representing similar interaction behaviors.

[0101] The first clustering submodule is used to cluster a plurality of first object nodes based on the first target graph to determine a plurality of first objects.

[0102] According to an embodiment of the present disclosure, the first construction submodule includes: a first acquisition unit, a first calculation unit and a first construction unit.

[0103] The first acquisition unit is used to acquire the interaction behavior information between each object to be processed and the target platform within a predetermined period of time.

[0104] The first calculation unit is used to calculate the behavior similarity between the objects to be processed according to the interaction behavior information.

[0105] The first construction unit is used to construct edges between the objects to be processed based on the identification of each object to be processed as a first object node and based on the behavior similarity to obtain a first target graph.

[0106] According to an embodiment of the present disclosure, the first clustering submodule includes a second calculating unit, a second clustering unit and a first obtaining unit.

[0107] The second calculation unit is used to calculate the change information of the modularity for each candidate clustering scheme based on the first target graph, and determine the intermediate clustering scheme from the candidate clustering schemes; wherein the modularity represents the similarity difference between the clustered objects and the non-clustered objects in the clustering results of at least two candidate clustering schemes.

[0108] The second clustering unit is used to cluster the intermediate objects based on the intermediate clustering scheme to obtain optimized relationships.

[0109] The first obtaining unit is used to return to execute the operation of calculating the change information of the modularity for each candidate clustering scheme based on the optimized relationship until the change information of the modularity converges to obtain multiple first objects.

[0110] According to an embodiment of the present disclosure, the second determination module includes: a second construction submodule and a second clustering submodule.

[0111] The second construction submodule is used to construct a second target graph based on the second initial relationship; wherein the second target graph includes a plurality of second object nodes and a plurality of attribute nodes representing associations.

[0112] The second clustering submodule is used to cluster the plurality of second object nodes based on the second target graph to determine the plurality of second objects.

[0113] According to an embodiment of the present disclosure, the second clustering submodule includes: a traversal unit and a third clustering unit.

[0114] The traversal unit is used to traverse the plurality of second object nodes and the plurality of attribute nodes in the second target graph in sequence.

[0115] The third clustering unit is used to cluster at least two second object nodes associated with the same attribute node to obtain a plurality of second objects.

[0116] According to an embodiment of the present disclosure, the second construction submodule includes: a second acquisition unit, an extraction unit and a second construction unit.

[0117] The second acquisition unit is used to acquire the platform registration information of each object to be processed and the target platform within a predetermined period of time.

[0118] The extraction unit is used to extract the attribute information of each object to be processed from the platform registration information.

[0119] The second construction unit is used to construct a second target graph by taking each object to be processed and each attribute information as a node and taking the corresponding relationship between each object to be processed and each attribute information as an edge.

[0120] According to an embodiment of the present disclosure, the clustering module includes: a third construction submodule and a classification submodule.

[0121] The third construction submodule is used to construct a third target graph using the first objects and the second objects as nodes and the relationships between the first objects and the relationships between the second objects as edges.

[0122] The classification submodule is used to use the identifier of each target object as the initial label of the node, and to update the initial label of each node by reading the initial categories of each neighboring node associated with each node until the maximum number of updates is reached to obtain at least two groups.

[0123] According to an embodiment of the present disclosure, the classification submodule includes: a reading unit and an updating unit.

[0124] The reading unit is used to obtain the quantity of each initial category by reading each initial category of each neighboring node associated with each node.

[0125] The updating unit is used to update the initial label of each node to the initial category with the largest number.

[0126] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0127] According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as described above.

[0128] According to an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause a computer to execute the method as described above.

[0129] According to an embodiment of the present disclosure, a computer program product includes a computer program, and when the computer program is executed by a processor, the computer program implements the method as described above.

[0130] Figure 7 A schematic block diagram of an example electronic device 700 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0131] like Figure 7 As shown, the device 700 includes a computing unit 701, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the device 700 can also be stored. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0132] A number of components in the device 700 are connected to the I / O interface 705, including: an input unit 706, such as a keyboard, a mouse, etc.; an output unit 707, such as various types of displays, speakers, etc.; a storage unit 708, such as a disk, an optical disk, etc.; and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the device 700 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0133] The computing unit 701 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 701 performs the various methods and processes described above, such as the object classification method. For example, in some embodiments, the object classification method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 708. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 700 via the ROM 702 and / or the communication unit 709. When the computer program is loaded into the RAM 703 and executed by the computing unit 701, one or more steps of the object classification method described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to perform the object classification method in any other appropriate manner (e.g., by means of firmware).

[0134] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0135] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0136] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0137] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0138] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0139] A computer system may include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises through computer programs running on respective computers and having a client-server relationship to each other. The server may be a cloud server, a server in a distributed system, or a server combined with a blockchain.

[0140] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0141] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for object classification, comprising: Based on the first initial relationship, determining a plurality of first objects from a plurality of objects to be processed; wherein the first initial relationship represents the interaction behavior relationship between each of the objects to be processed and the target platform; Based on the second initial relationship, determining a plurality of second objects from the plurality of objects to be processed; wherein the second initial relationship represents the relationship between each of the objects to be processed and the attributes of each of the objects to be processed; The multiple first objects and the multiple second objects are clustered to classify the multiple first objects and the multiple second objects into at least two target groups, so as to perform the same interactive operation on each target object in the same target group; each target object in each target group has similar interactive behavior with the target platform.

2. The method according to claim 1, wherein the determining a plurality of first objects from a plurality of objects to be processed based on the first initial relationship comprises: Based on the first initial relationship, construct a first target graph; wherein the first target graph includes a plurality of first object nodes representing similar interaction behaviors; and Based on the first target graph, the plurality of first object nodes are clustered to determine the plurality of first objects.

3. The method according to claim 2, wherein: The step of constructing a first target graph based on the first initial relationship includes: Obtaining interaction behavior information between each object to be processed and the target platform within a predetermined period of time; Calculating the behavior similarity between the objects to be processed according to the interaction behavior information; and The identifier of each object to be processed is used as the first object node, and based on the behavior similarity, the edges between each object to be processed are constructed to obtain the first target graph.

4. The method according to claim 2, wherein: The clustering of the plurality of objects to be processed based on the first target graph to determine the plurality of first objects includes: Based on the first target graph, for each candidate clustering scheme, respectively calculate the change information of the modularity, and determine the intermediate clustering scheme from the candidate clustering schemes; wherein the modularity represents the similarity difference between the clustered objects and the non-clustered objects in the clustering results of at least two candidate clustering schemes; Based on the intermediate clustering scheme, clustering the intermediate objects to obtain optimized relationships; and Based on the optimized relationship, the operation of calculating the change information of the modularity for each candidate clustering scheme is returned to be executed until the change information of the modularity converges to obtain the multiple first objects.

5. The method according to claim 1, wherein: The step of determining a plurality of second objects from the plurality of objects to be processed based on the second initial relationship comprises: Based on the second initial relationship, construct a second target graph; wherein the second target graph includes a plurality of second object nodes and a plurality of attribute nodes representing associations; and Based on the second target graph, a plurality of the second object nodes are clustered to determine the plurality of second objects.

6. The method according to claim 5, wherein: The step of clustering the plurality of object nodes based on the second target graph to determine the plurality of second objects comprises: sequentially traversing the plurality of second object nodes and the plurality of attribute nodes in the second target graph; and At least two second object nodes associated with the same attribute node are clustered to obtain the plurality of second objects.

7. The method according to claim 5, wherein: The step of constructing a second target graph based on the second initial relationship includes: Acquire platform registration information of each object to be processed and the target platform within a predetermined period of time; Extracting attribute information of each object to be processed from the platform registration information; and The second target graph is constructed by taking each of the objects to be processed and each of the attribute information as nodes, and taking the corresponding relationship between each of the objects to be processed and each of the attribute information as edges.

8. The method according to claim 1, wherein: The clustering of the plurality of first objects and the plurality of second objects to classify the plurality of first objects and the plurality of second objects into at least two target groups includes: constructing a third target graph by taking the first objects and the second objects as nodes and taking the relationships between the first objects and the relationships between the second objects as edges; and The identifier of each target object is used as the initial label of the node, and the initial label of each node is updated by reading the initial categories of each neighbor node associated with each node until the maximum number of updates is reached, thereby obtaining the at least two groups.

9. The method according to claim 8, wherein: The updating of the initial label of each node by reading the initial category of each neighbor node associated with each node includes: By reading each initial category of each neighboring node associated with each node, the number of each initial category is obtained; and Update the initial label of each node to the initial category with the largest number.

10. An object classification device, comprising: A first determination module, configured to determine a plurality of first objects from a plurality of objects to be processed based on a first initial relationship; wherein the first initial relationship represents an interactive behavior relationship between each of the objects to be processed and a target platform; A second determination module is used to determine a plurality of second objects from the plurality of objects to be processed based on a second initial relationship; wherein the second initial relationship represents the relationship between each of the objects to be processed and the attributes of each of the objects to be processed; A clustering module is used to classify the multiple first objects and the multiple second objects into at least two target groups by clustering the multiple first objects and the multiple second objects, so as to perform the same interactive operation on each target object in the same target group; each target object in each target group has similar interactive behavior with the target platform.

11. The device according to claim 10, wherein: The first determining module comprises: A first construction submodule is configured to construct a first target graph based on the first initial relationship; wherein the first target graph includes a plurality of first object nodes representing similar interaction behaviors; and The first clustering submodule is used to cluster the multiple first object nodes based on the first target graph to determine the multiple first objects.

12. The apparatus according to claim 11, wherein the first construction submodule comprises: A first acquisition unit is used to acquire the interaction behavior information between each object to be processed and the target platform within a predetermined period of time; A first calculation unit, configured to calculate the behavior similarity between the objects to be processed according to the interaction behavior information; as well as The first construction unit is used to construct edges between the objects to be processed based on the behavior similarity by taking the identifiers of the objects to be processed as the first object nodes to obtain the first target graph.

13. The method according to claim 11, wherein: The first clustering submodule comprises: a second calculation unit, configured to calculate the change information of the modularity for each candidate clustering scheme based on the first target graph, and determine an intermediate clustering scheme from the candidate clustering schemes; wherein the modularity represents the difference in similarity between the clustered objects and the non-clustered objects in the clustering results of at least two candidate clustering schemes; A second clustering unit is used to cluster the intermediate objects based on the intermediate clustering scheme to obtain an optimized relationship; and The first obtaining unit is used to return to execute the operation of calculating the change information of the modularity for each candidate clustering scheme based on the optimized relationship, until the change information of the modularity converges, thereby obtaining the multiple first objects.

14. The method according to claim 10, wherein: The second determining module includes: A second construction submodule, configured to construct a second target graph based on a second initial relationship; wherein the second target graph includes a plurality of second object nodes and a plurality of attribute nodes representing associations; and The second clustering submodule is used to cluster the plurality of second object nodes based on the second target graph to determine the plurality of second objects.

15. The method according to claim 14, wherein: The second clustering submodule comprises: a traversal unit, configured to sequentially traverse the plurality of second object nodes and the plurality of attribute nodes in the second target graph; and The third clustering unit is used to cluster at least two second object nodes associated with the same attribute node to obtain the multiple second objects.

16. The method according to claim 14, wherein: The second construction submodule includes: A second acquisition unit is used to acquire platform registration information of each object to be processed and the target platform within a predetermined period of time; an extraction unit, configured to extract attribute information of each object to be processed from the platform registration information; and The second construction unit is used to construct the second target graph by taking each of the objects to be processed and each of the attribute information as nodes and taking the corresponding relationship between each of the objects to be processed and each of the attribute information as edges.

17. The method according to claim 10, wherein: The clustering module comprises: A third construction submodule is used to construct a third target graph using the first objects and the second objects as nodes and the relationships between the first objects and the relationships between the second objects as edges; and The classification submodule is used to use the identifier of each target object as the initial label of the node, update the initial label of each node by reading the initial categories of each neighbor node associated with each node, until the maximum number of updates is reached, and obtain the at least two groups.

18. The method according to claim 17, wherein: The classification submodule comprises: A reading unit, configured to obtain the number of each initial category by reading each initial category of each neighboring node associated with each node; and The updating unit is used to update the initial label of each node to the initial category with the largest number.

19. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 9.

20. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-9.

21. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 9.

Images