Bilateral network division

Abstract

Disclosed in embodiments of the present description are a bilateral network division method and apparatus, an electronic device, and a storage medium. The bilateral network division method comprises: acquiring historical interaction data of a first object group, and determining a bilateral object group on the basis of the historical interaction data, wherein the bilateral object group comprises at least one first object in the first object group and at least one second object in a second object group; next, using individual objects in the bilateral object group as nodes and data attribute information of target interaction data of the individual objects as feature vectors for connecting the nodes, constructing a probability model, and assigning node attribute information to the nodes; then, on the basis of a target classification rule, determining an optimization mode of the probability model, and on the basis of the optimization mode, processing and expanding the nodes and / or the feature vectors in the probability model; and finally, performing node clustering on the feature vectors in the processed probability model, and classifying the nodes into at least one bilateral network set.

Description

Bilateral network partitioning Technical Field

[0001] The embodiments of this specification belong to the field of data processing technology, and particularly relate to a bilateral network partitioning method, device, electronic device, and storage medium. Background Art

[0002] With the rapid development of information technology, more and more companies are incorporating bilateral networks into their operational model decisions. A bilateral network is composed of two distinct types of participants, who interact and collaborate with each other, fostering their growth and development. However, because bilateral networks often involve two or more distinct groups of participants, there is currently no reliable way to accurately categorize them, leading to ineffective corporate decision-making based on these networks. Summary of the Invention

[0003] The embodiments of this specification provide a bilateral network partitioning method, device, electronic device, and storage medium, and the technical solutions are as follows.

[0004] In a first aspect, an embodiment of this specification provides a bilateral network partitioning method, including:

[0005] Acquire historical interaction data of a first object group, and determine a bilateral object group based on the historical interaction data, wherein the bilateral object group includes at least one first object in the first object group and at least one second object in the second object group, and the second object has an interaction history with at least one first object; construct a probability model with each individual object in the bilateral object group as a node and data attribute information of the target interaction data of the individual object as a feature vector connecting the nodes, and assign node attribute information to each node; determine an optimization method for the probability model based on a target classification rule, and process and expand the nodes and / or feature vectors in the probability model based on the optimization method; perform node clustering on each feature vector in the processed probability model, and divide each node into at least one bilateral network set.

[0006] In a second aspect, an embodiment of the present specification provides a bilateral network partitioning device, comprising: an acquisition module, configured to acquire historical interaction data of a first object group, and determine a bilateral object group based on the historical interaction data, wherein the bilateral object group includes at least one first object in the first object group and at least one second object in the second object group, and the second object has an interaction history with at least one first object; a construction module, configured to construct a probability model with each individual object in the bilateral object group as a node and data attribute information of the target interaction data of the individual object as a feature vector connecting the nodes, and assign node attribute information to each node; an expansion module, configured to determine an optimization method of the probability model based on a target classification rule, and process and expand the nodes and / or feature vectors in the probability model based on the optimization method; and a partitioning module, configured to perform node clustering on each feature vector in the processed probability model, and partition each node into at least one bilateral network set.

[0007] In a third aspect, an embodiment of this specification further provides an electronic device, which may include: a processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the above-mentioned bilateral network partitioning method steps.

[0008] In a fourth aspect, an embodiment of this specification provides a computer storage medium, which stores a plurality of instructions suitable for being loaded by a processor and executing the above-mentioned bilateral network partitioning method steps.

[0009] The technical solutions provided in some embodiments of this specification offer at least the following beneficial effects: In one or more embodiments of this specification, after determining a bilateral subject group based on historical interaction data of a first subject group, a probabilistic model can be constructed based on the bilateral subject group, and the nodes and feature vectors in the probabilistic model can be expanded and optimized according to target classification rules to improve the probabilistic model. Ultimately, by clustering the nodes in the probabilistic model, bilateral network sets can be delineated with high accuracy, thereby improving the effectiveness of enterprise decision-making based on bilateral networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of this specification, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0011] FIG1 is a schematic diagram of a system architecture of a bilateral network partitioning method provided in an embodiment of this specification.

[0012] FIG2 is a flow chart of a bilateral network partitioning method provided in an embodiment of this specification.

[0013] FIG3 is a schematic diagram showing an example of the process of bilateral market segmentation provided in an embodiment of this specification.

[0014] FIG4 is a flowchart of another bilateral network partitioning method provided in an embodiment of this specification.

[0015] FIG5 is a schematic structural diagram of a bilateral network partitioning device provided in an embodiment of this specification.

[0016] FIG6 is a schematic structural diagram of an electronic device provided in an embodiment of this specification. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of this specification will be described clearly and completely below in conjunction with the drawings in the embodiments of this specification.

[0018] Throughout this specification, the claims, and the accompanying drawings, the terms "first," "second," "third," and the like are used to distinguish between different items, not to describe a particular order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements but may optionally include steps or elements not listed, or may include other steps or elements inherent to the process, method, product, or apparatus.

[0019] The following description provides examples and does not limit the scope, applicability, or examples set forth in the claims. Changes may be made to the functions and arrangements of the elements described without departing from the scope of this specification. Various examples may appropriately omit, replace, or add various processes or components. For example, the described methods may be performed in an order different from the order described, and various steps may be added, omitted, or combined. Furthermore, features described with respect to some examples may be combined in other examples.

[0020] First, some of the embodiments of the present application will be explained below to facilitate understanding by those skilled in the art.

[0021] A bilateral network is a network composed of two distinct types of participants. These two types of participants are interdependent within the network, interacting and collaborating with each other. A characteristic of bilateral networks is that interactions and transactions between participants generate mutually beneficial effects, thereby promoting the network's growth and development. In a bilateral network, the actions of one party directly impact the interests of the other, and vice versa. For example, an e-commerce platform is a typical bilateral network, where buyers and sellers conduct transactions on the platform, and their interaction and collaboration foster the platform's prosperity.

[0022] Two-sided marketplace: A specific form of a two-sided network, where two distinct participants—supply and demand—transact to achieve mutual benefit. In a two-sided marketplace, the interaction between supply and demand creates a two-way transfer of value.

[0023] A probabilistic model is a mathematical model used to describe and infer relationships between random variables. Based on the principles of probability theory, it uses probability distributions and conditional probabilities to represent dependencies and uncertainties between variables. Probabilistic models are used in a variety of fields, such as statistics, machine learning, and artificial intelligence. They can be used to solve a variety of problems, including prediction, classification, clustering, and decision-making.

[0024] A graphical model is a probabilistic model that uses graphs to represent relationships between variables. It uses nodes and edges to represent variables and their dependencies, and is used to describe the conditional probability distribution between variables. In a graphical model, nodes represent random variables, and edges represent dependencies between variables.

[0025] Please refer to FIG1 , which shows a schematic diagram of the system architecture of a bilateral network partitioning method provided in an embodiment of this specification.

[0026] As shown in FIG1 , the system architecture of the bilateral network partitioning method may include at least a terminal 100 , a server 200 and a network.

[0027] Terminal 100 includes, but is not limited to, electronic devices such as smartphones, desktop computers, tablet computers, laptops, smart speakers, digital assistants, and smart wearable devices. It may also be software running on these electronic devices, such as applications. Optionally, the operating system running on the electronic device may include, but is not limited to, Android, iOS, Linux, Windows, etc. Optionally, terminal 100 provides users with a bilateral network partitioning service. Terminal 100 may obtain operating instructions from an application program interface and send a bilateral network establishment request to server 200, causing server 200 to perform bilateral network partitioning.

[0028] Server 200 can provide background services for terminal 100, construct a probability model based on the bilateral network establishment request sent by terminal 100, and divide at least one bilateral network set through optimization and clustering of the probability model. Specifically, server 200 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0029] The network is a medium for providing a communication link between the terminal 100 and the server 200, between the terminals 100 and 100, and between the servers 200 and 200. The network may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0030] In addition, it should be noted that what is shown in FIG1 is only a system provided by the present disclosure. In practical applications, other systems may also be included, for example, more terminals may be included.

[0031] In the embodiments of this specification, the terminal 100 and the server 200 may be directly or indirectly connected via wired or wireless communication, which is not limited in this disclosure.

[0032] Next, please refer to FIG. 2 , which shows an overall flow chart of a bilateral network partitioning method provided by an embodiment of this specification. The bilateral network partitioning method can be used in a server.

[0033] As shown in FIG2 , the bilateral network partitioning method may include at least the following steps.

[0034] Step 201: Acquire historical interaction data of a first object group, and determine a bilateral object group based on the historical interaction data.

[0035] The bilateral object group includes at least one first object in the first object group and at least one second object in the second object group, and the second object has an interaction history with at least one first object.

[0036] First, it's important to note that two-sided networks can be applied in a variety of fields. For example, in e-commerce, two-sided networks are defined as buyers and sellers. Buyers and sellers conduct transactions on a platform, and their interaction and collaboration foster the platform's growth.

[0037] As another example, two-sided networks can be applied to shared resources, such as ride-sharing platforms and accommodation-sharing platforms. Providers and demanders are the two groups in a two-sided network. By connecting these two groups, a two-sided network is constructed. Providers can offer their own resources, while demanders can find suitable resources through the platform, enabling shared utilization of resources.

[0038] As another example, two-sided networks can be applied to social media. Users and platforms are the two groups in a two-sided network. Social media platforms connect users and form a two-sided network by connecting them and their interactions. Users can socialize, share, and communicate through the platform, while also contributing content and activity to the platform.

[0039] As another example, two-sided networks can be applied to online auctions. Buyers and sellers are the two groups of entities in a two-sided network. Online auction platforms connect sellers and buyers, forming a two-sided network. Sellers can auction their items on the platform, while buyers can bid and purchase.

[0040] As another example, two-sided networks can be applied to talent markets. Job seekers and recruiters are two groups in a two-sided network. A talent market platform connects these two groups, forming a two-sided network. Job seekers can showcase their skills and experience on the platform, while recruiters can search for suitable talent.

[0041] As another example, two-sided networks can be applied to venture capital. Entrepreneurs and investors are two groups within a two-sided network, and the relationship between them can also be considered a two-sided network. Entrepreneurs need financial support and resources, while investors seek out promising projects to invest in.

[0042] The fields in which the bilateral network can be applied include but are not limited to the fields mentioned above. For the sake of convenience, this embodiment and the following embodiments will be explained based on the case where the bilateral network is applied in the field of e-commerce.

[0043] Specifically, the first object group can be consumers on the e-commerce platform, that is, a group of individuals who use the Internet platform to purchase products or services. The second object group can be supply shops on the e-commerce platform, that is, a group of individuals or shops that use the Internet platform to sell products or services. The server can obtain the historical interaction data of the first object group from the historical database of the e-commerce platform, that is, the transaction data of consumers on the platform. Based on these transaction data, it can be determined which consumers and which supply shops have conducted transactions. Ultimately, all consumers and all supply shops determined by the server based on the historical interaction data together constitute a bilateral object group. Among them, the specific selection range of the historical interaction data, such as the selected time period, the selected date, etc., can be set differently according to the actual bilateral area division needs.

[0044] Furthermore, since bilateral object groups are constructed to demarcate bilateral regions and determine the boundaries of regional subsets for each category, some individuals that clearly do not interact significantly with other objects can be eliminated during the construction of bilateral object groups. Therefore, the resulting bilateral object groups may not necessarily include all individuals in the first and / or second object groups.

[0045] Step 203 : construct a probability model with each individual object in the bilateral object group as a node and the data attribute information of the target interaction data of the individual object as a feature vector connecting the nodes, and assign node attribute information to each node.

[0046] Specifically, in the aforementioned steps, the bilateral object group determined based on the selected historical interaction data has certain limitations and cannot cover all individual objects that the bilateral object group should actually include. Therefore, to accurately demarcate the bilateral area, the server can treat each individual object in the bilateral object group as a node and use the data attribute information of each target interaction data as the feature vector connecting the nodes, that is, the relationship edge connecting the nodes, to construct a probability model. At the same time, the server also assigns node attribute information to each node in the probability model. The individual object can be either the first object or the second object. The probability model can specifically be a graph model, a tree model, a deep learning model, etc. The target interaction data can be the interaction data generated when the first object and the second object engage in a certain type of interaction. The data attribute information can be determined by the server based on the specific interaction type of the target interaction data, such as clothing transactions, food transactions, electronic device transactions, etc. Node attribute information can be determined based on the attributes of the individual object corresponding to the node, such as education level, current location, height, weight, age, store category, store location, store size, etc. It can also be determined based on the data type ratio of all interaction data of the individual object corresponding to the node. For example, if clothing transactions account for the largest proportion, the node will be assigned an attribute with a clothing bias; if food transactions account for the largest proportion, the node will be assigned an attribute with a food bias, and so on. The feature vector can be set to different specific vector values ​​based on the different data attribute information.

[0047] As an optional embodiment of this specification, step 203 uses each individual object in the bilateral object group as a node and the data attribute information of the target interaction data of the individual object as the feature vector connecting the nodes to construct a probability model, including: determining the individual interaction data of each individual object in the bilateral object group, clustering each individual interaction data based on the data type to obtain the target interaction data; and using each individual object as a node and the data attribute information of the target interaction data as the feature vector connecting the nodes to construct a probability model.

[0048] Specifically, the server constructs a probabilistic model to enable subsequent learning and optimization of the nodes and feature vectors within the probabilistic model, thereby expanding the number of different types of nodes and feature vectors. For a given individual object, if it interacts with the same individual object multiple times for the same type of transaction, this should be represented in the probabilistic model as a single feature vector, rather than as multiple feature vectors that would increase the model's complexity. Therefore, the server first determines the complete interaction data for each individual object in the bilateral object group, i.e., the individual interaction data. It then performs clustering calculations on all individual interaction data for each individual object, grouping individual interaction data of the same interaction type into a single target interaction data set. This ensures that no data of the same interaction type exists within the target interaction data for each individual object. Finally, the server constructs a probabilistic model based on the data attribute information corresponding to these target interaction data as feature vectors.

[0049] As an optional embodiment of this specification, after constructing the probability model in step 203, the method further includes: performing data distribution analysis on the probability model and deleting outlier samples in the probability model.

[0050] Specifically, the specific vector value of the feature vector can be determined based on the interaction type. After the clustering and other processing steps described above, the specific value of the feature vector may fluctuate discretely. The server also performs a data distribution analysis of discrete data on each feature vector in the probability model, determining the deviation of each vector value relative to the remaining vector values. Feature vectors with significant deviations, i.e., deviations greater than a preset deviation value, are removed as outliers to improve the accuracy of subsequent feature vector learning in the probability model.

[0051] Step 205: Determine an optimization method for the probability model based on target classification rules, and process and expand the nodes and / or feature vectors in the probability model based on the optimization method.

[0052] Specifically, enterprises typically segment their bilateral networks to analyze the market and adjust their business decisions. Different enterprises may have different market priorities. Therefore, before segmenting bilateral regions, enterprise managers can send classification information to the server via a terminal or other means. This information includes target classification rules, i.e., the specific classification criteria for the enterprise's bilateral regions. Based on the target classification rules, the server determines different optimization methods and then optimizes and learns the probability model to expand the nodes and / or feature vectors of the probability model.

[0053] As an example, the target classification rule can specifically characterize whether the classification of the bilateral region focuses only on the connection information of the bilateral object group, that is, the feature vector, without considering the attribute information of the individual, that is, the node, or whether it considers not only the connection information but also the attribute information of the individual. For the former, the server can use a random walk algorithm to learn the feature vectors of the nodes in the probability model, and after the learning is completed, infer other possible nodes and / or feature vectors. For the latter, the server can use a graph neural network to learn the feature vectors of the nodes in the probability model, and after the learning is completed, infer other possible nodes and / or feature vectors.

[0054] Step 207: perform node clustering on each of the processed feature vectors in the probability model, and divide each of the nodes into at least one bilateral network set.

[0055] Specifically, after learning and optimizing the probability model, the server can assume that the probability model has relatively comprehensive coverage of all possible nodes and feature vectors. At this point, the server clusters nodes not by transaction type, but rather by node. Nodes are clustered together not for redundancy removal but rather because they are considered to have strong correlations, forming a bilateral network set. This allows businesses to select the appropriate bilateral network set based on their needs and then determine the network boundaries based on the set. Specifically, the server determines, based on the nodes in the set, which consumers and commercial vendors the set covers, and can then use this information to inform decisions about the company's subsequent marketing strategies.

[0056] As an example, as shown in FIG3 , the overall process of the embodiment of this specification will be schematically explained by taking the bilateral market in the bilateral network as an example. The server determines the bilateral object group through the user's consumption transaction data, and then constructs a graph model. Then, the server optimizes the graph model by feature engineering to improve the nodes covered by the extended graph model and the feature vectors corresponding to the nodes. Finally, the server can identify the market boundaries by clustering the nodes, that is, dividing the market into bilateral area sets, so that managers can make decisions based on the boundary range of a specific bilateral area set.

[0057] As an optional embodiment of this specification, the method further includes: setting a label for the bilateral network set based on network set attribute information of the bilateral network set.

[0058] Specifically, each bilateral network set has network set attribute information. This network set attribute information can be determined based on the node attribute information of each node within the bilateral network set and the data attribute information of the feature vector. For example, if both the node attribute information and the data attribute information include winter clothing attributes, the network set attribute information can be winter clothing attributes. The server can then assign corresponding attribute tags to these bilateral network sets based on the network set attribute information, allowing enterprises to quickly find the bilateral network sets they need based on the tags.

[0059] As an optional embodiment of this specification, the method also includes: obtaining current decision information, determining a first network range corresponding to the current decision information; determining a target bilateral network set that matches the current decision information, and determining a second network range of the target bilateral network set; and generating decision recommendation information based on a non-intersecting range of the first network range and the second network range.

[0060] Specifically, the server can obtain the company's current decision information and, based on this information, determine the first network scope—that is, which types of consumers and stores the company intends to cover with this decision. Next, the server can identify, from each bilateral network set, a target bilateral network set that matches the attribute type targeted by the current decision information, such as winter clothing or hot pot restaurants, and determine the second network scope of the target bilateral network set. By comparing the first and second network scopes, the server can determine which individual objects in the bilateral network are not actually covered by the company's decision, or which individual objects covered by the decision are invalid. This information can then be used to generate decision-making recommendations and fed back to the terminal used by the decision-maker to help them optimize their decision-making plan.

[0061] In the embodiments of this specification, after determining bilateral subject groups based on the historical interaction data of a first subject group, a probabilistic model is constructed based on the bilateral subject groups. The nodes and feature vectors in the probabilistic model are then expanded and optimized according to target classification rules to improve the probabilistic model. Ultimately, by clustering the nodes in the probabilistic model, bilateral network sets can be delineated with high accuracy, thereby improving the effectiveness of enterprise decision-making based on bilateral networks.

[0062] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0063] As another optional embodiment of this specification, please refer to Figure 4, which shows an overall flow chart of another bilateral network partitioning method provided by an embodiment of this specification. The bilateral network partitioning method can be used in a server.

[0064] As shown in FIG4 , the bilateral network partitioning method may include at least the following steps.

[0065] Step 401: Determine the interaction type of the target second object, and set data acquisition conditions based on the interaction type.

[0066] The data acquisition conditions include time conditions and space conditions.

[0067] Specifically, those who want to perform bilateral network segmentation are generally supply shops, that is, merchants. Therefore, in the process of bilateral network segmentation, the probability model should be constructed and clustered based on data related to the merchant itself. Data that is completely unrelated to the merchant is not helpful for the market segmentation desired by the merchant. For example, if the merchant sells breakfast, the server should only consider data during the breakfast time period. For another example, if the merchant is located in a scenic area, other merchants in the same scenic area should be considered when performing bilateral area segmentation. Therefore, the server can determine the merchant who wants to perform bilateral network segmentation, that is, the interaction type of the target second object, and set different time conditions and space conditions.

[0068] Step 403: Acquire historical interaction data of the first object group based on the data acquisition condition.

[0069] Specifically, after the data acquisition condition is determined in the aforementioned step, the server will use the data acquisition condition as a standard to acquire the historical interaction data of the first object group from the historical database.

[0070] Step 405 : Calculate the correlation value between each data type of each historical interaction data and the interaction type, filter out invalid data whose correlation value is less than a preset correlation value, and obtain filtered data.

[0071] Specifically, the historical interaction data obtained by the server may still contain some data that is useless to merchants. To improve the accuracy of subsequent bilateral region divisions, the server should identify valid and research-worthy transaction data from the historical interaction data. Specifically, the server will determine the data type of each historical interaction data and, through correlation calculation, determine the correlation value between each data type and the interaction type. If the correlation value is less than the preset correlation value, the data is considered to have a low correlation with the merchant's interaction type and will be eliminated as invalid data, ultimately resulting in the retained filtered data.

[0072] Step 407: Construct a bilateral subject group based on the screening data.

[0073] Specifically, the server will determine the transaction objects of the two parties in each filtered data, and summarize the transaction objects corresponding to all filtered data to construct a bilateral object group.

[0074] Step 409 : construct a probability model with each individual object in the bilateral object group as a node and the data attribute information of the target interaction data of the individual object as a feature vector connecting the nodes, and assign node attribute information to each node.

[0075] Specifically, step 409 may refer to step 203 and will not be described in detail here.

[0076] Step 411: Determine a target classification rule in a preset classification rule library, and determine an optimization method for the probability model based on the target classification rule and the number of two-sided groups.

[0077] The classification rule base includes a first classification rule based on the classification of the data attribute information and a second classification rule based on the classification of the node attribute information and the data attribute information.

[0078] Specifically, the server can obtain classification information sent by enterprise decision-makers via terminals. This classification information can indicate the current classification requirements, namely, whether to consider data attribute information classification or node attribute information classification. The cloud server then selects the corresponding target classification rule from a preset classification rule library based on this classification information. In addition to obtaining the target classification rule, the server can also obtain the number of bilateral groups in a bilateral object group, that is, determine the number of individual objects in the bilateral object group. The server ultimately comprehensively considers the optimization method of the probability model based on the target classification rule and the number of bilateral groups.

[0079] As an example, the number of bilateral groups can be prioritized over the target classification rules. When the number of bilateral groups is less than a preset number, the scope of the two-sided market being studied is considered small, and the number of bilateral groups involved is small. In this case, the differences in target classification rules will not be considered, and manual feature engineering or matrix classification will be used as the optimization method to subsequently optimize the probability model. When the number of bilateral groups is not less than a preset number, the differences in target classification rules can be considered, that is, the random walk algorithm will be used as the optimization method for the first classification rule, and the graph neural network will be used as the optimization method for the second classification rule.

[0080] Step 413: Expand the nodes and / or feature vectors in the probability model based on the optimization process.

[0081] Specifically, step 413 may refer to step 205 and will not be described in detail here.

[0082] Step 415: perform node clustering on each of the processed feature vectors in the probability model, and divide each of the nodes into at least one bilateral network set.

[0083] Specifically, step 415 may refer to step 207 and will not be described in detail here.

[0084] In the embodiment of this specification, the data acquisition conditions will be determined according to the interaction type of the target second object, and then the historical interaction data will be obtained according to the data acquisition conditions. Then, the server will also screen out invalid data in the historical interaction data based on the correlation, and construct a bilateral object group based on the remaining screened data. After the bilateral object group is determined, a probability model is constructed based on the bilateral object group, and an optimization method is selected based on the target classification rules and the number of bilateral groups. The nodes and feature vectors in the probability model are expanded and optimized according to the optimization method to improve the probability model. Ultimately, by clustering the nodes in the probability model, bilateral network sets can be divided, and the accuracy of the division of the bilateral network sets is high, which makes the actual effect of the enterprise's decision-making based on the bilateral network better.

[0085] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0086] Next, please refer to Figure 5, which shows a schematic diagram of the structure of a bilateral network partitioning device provided in an embodiment of this specification. It should be noted that the bilateral network partitioning device shown in Figure 5 is used to execute the method of the embodiment shown in Figure 2 of this application. For ease of explanation, only the portion relevant to the embodiment of this application is shown. For specific technical details not disclosed, please refer to the embodiment shown in Figure 2 of this application.

[0087] As shown in Figure 5, the bilateral network partitioning device may at least include: an acquisition module 501, used to acquire historical interaction data of a first object group, and determine a bilateral object group based on the historical interaction data, wherein the bilateral object group includes at least one first object in the first object group and at least one second object in the second object group, and the second object has an interaction history with at least one first object; a construction module 502, used to construct a probability model with each individual object in the bilateral object group as a node and data attribute information of the target interaction data of the individual object as a feature vector connecting the nodes, and assign node attribute information to each node; an expansion module 503, used to determine an optimization method of the probability model based on a target classification rule, and process and expand the nodes and / or feature vectors in the probability model based on the optimization method; a partitioning module 504, used to perform node clustering on each feature vector in the processed probability model, and partition each node into at least one bilateral network set.

[0088] As an optional embodiment of this specification, the acquisition module 501 is specifically used to: determine the interaction type of the target second object, and set data acquisition conditions based on the interaction type, and the data acquisition conditions include time conditions and space conditions; acquire historical interaction data of the first object group based on the data acquisition conditions.

[0089] As an optional embodiment of this specification, the acquisition module 501 is specifically further used to: respectively calculate the correlation value between each data type of each historical interaction data and the interaction type, filter out invalid data whose correlation value is less than a preset correlation value, and obtain filtered data; and construct a bilateral object group based on the filtered data.

[0090] As an optional embodiment of this specification, the construction module 502 is specifically used to: determine the individual interaction data of each individual object in the bilateral object group, cluster the individual interaction data based on the data type to obtain target interaction data; and construct a probability model with each individual object as a node and the data attribute information of the target interaction data as the feature vector connecting the nodes.

[0091] As an option in the embodiment of this specification, the construction module 502 is further configured to perform data distribution analysis on the probability model and delete outlier samples in the probability model.

[0092] As an optional embodiment of this specification, the extension module 503 is specifically used to: determine the target classification rule in a preset classification rule library, and determine the optimization method of the probability model based on the target classification rule and the number of bilateral groups, the classification rule library including a first classification rule based on the classification of the data attribute information and a second classification rule based on the classification of the node attribute information and the data attribute information.

[0093] As an option of the embodiment of this specification, the apparatus further includes: a setting module, configured to set a label for the bilateral network set based on the network set attribute information of the bilateral network set.

[0094] As an optional embodiment of this specification, the device also includes: a first determination module, used to obtain current decision information and determine a first network range corresponding to the current decision information; a second determination module, used to determine a target bilateral network set that matches the current decision information, and determine a second network range of the target bilateral network set; a generation module, used to generate decision recommendation information based on a non-intersection range of the first network range and the second network range.

[0095] Those skilled in the art will clearly understand that the technical solutions of the embodiments of the present application can be implemented with the help of software and / or hardware. "Unit" and "module" in this specification refer to software and / or hardware that can independently perform or cooperate with other components to perform specific functions, where the hardware can be, for example, a field programmable gate array (FPGA), an integrated circuit (IC), etc.

[0096] Each processing unit and / or module in the embodiments of the present application may be implemented by an analog circuit that implements the functions described in the embodiments of the present application, or may be implemented by software that executes the functions described in the embodiments of the present application.

[0097] Next, please refer to FIG6 , which shows a schematic structural diagram of an electronic device provided in an embodiment of this specification.

[0098] As shown in FIG. 6 , the electronic device 600 may include: at least one processor 601 , at least one network interface 604 , a user interface 603 , a memory 605 , and at least one communication bus 602 .

[0099] The communication bus 602 may be used to implement connection and communication among the above components.

[0100] The user interface 603 may include buttons, and the optional user interface may also include a standard wired interface or a wireless interface.

[0101] The network interface 604 may include, but is not limited to, a Bluetooth module, an NFC module, a Wi-Fi module, and the like.

[0102] Among them, the processor 601 may include one or more processing cores. The processor 601 uses various interfaces and lines to connect the various parts of the entire electronic device 600, and executes various functions of the electronic device 600 and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 605, and calling data stored in the memory 605. Optionally, the processor 601 can be implemented in at least one hardware form of DSP, FPGA, PLA. The processor 601 can integrate one or a combination of CPU, GPU and modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to handle wireless communications. It is understandable that the above-mentioned modem may not be integrated into the processor 601, but may be implemented separately through a chip.

[0103] Among them, the memory 605 may include RAM and may also include ROM. Optionally, the memory 605 includes a non-transitory computer-readable medium. The memory 605 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 605 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 605 may also be at least one storage device located away from the aforementioned processor 601. As shown in Figure 6, the memory 605 as a computer storage medium may include an operating system, a network communication module, a user interface module and program instructions.

[0104] Specifically, the processor 601 can be used to call the bilateral network partitioning application stored in the memory 605, and specifically perform the following operations: obtain historical interaction data of the first object group, determine the bilateral object group based on the historical interaction data, the bilateral object group includes at least one first object in the first object group and at least one second object in the second object group, and the second object has an interaction history with at least one first object; use each individual object in the bilateral object group as a node and the data attribute information of the target interaction data of the individual object as the feature vector connecting the nodes, construct a probability model, and assign node attribute information to each node; determine the optimization method of the probability model based on the target classification rule, and process and expand the nodes and / or feature vectors in the probability model based on the optimization method; perform node clustering on each feature vector in the processed probability model, and divide each node into at least one bilateral network set.

[0105] As an optional embodiment of this specification, obtaining the historical interaction data of the first object group includes: determining the interaction type of the target second object, and setting data acquisition conditions based on the interaction type, the data acquisition conditions including time conditions and space conditions; and obtaining the historical interaction data of the first object group based on the data acquisition conditions.

[0106] As an optional embodiment of this specification, determining the bilateral object group based on the historical interaction data includes: respectively calculating the correlation value between each data type of each historical interaction data and the interaction type, filtering out invalid data whose correlation value is less than a preset correlation value, and obtaining filtered data; and constructing the bilateral object group based on the filtered data.

[0107] As an optional embodiment of this specification, the probability model is constructed with each individual object in the bilateral object group as a node and the data attribute information of the target interaction data of the individual object as the feature vector connecting the nodes, including: determining the individual interaction data of each individual object in the bilateral object group, clustering each individual interaction data based on the data type to obtain the target interaction data; and constructing the probability model with each individual object as a node and the data attribute information of the target interaction data as the feature vector connecting the nodes.

[0108] As an optional embodiment of this specification, after constructing the probability model, the method further includes: performing data distribution analysis on the probability model and deleting outlier samples in the probability model.

[0109] As an optional embodiment of this specification, the optimization method of determining the probability model based on the target classification rule includes: determining the target classification rule in a preset classification rule library, and determining the optimization method of the probability model based on the target classification rule and the number of bilateral groups, the classification rule library including a first classification rule based on the classification of the data attribute information and a second classification rule based on the classification of the node attribute information and the data attribute information.

[0110] As an optional embodiment of this specification, the method further includes: setting a label for the bilateral network set based on network set attribute information of the bilateral network set.

[0111] As an optional embodiment of this specification, the method also includes: obtaining current decision information, determining a first network range corresponding to the current decision information; determining a target bilateral network set that matches the current decision information, and determining a second network range of the target bilateral network set; and generating decision recommendation information based on a non-intersecting range of the first network range and the second network range.

[0112] The embodiments of this specification also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above method. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a microdrive, a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic or optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.

[0113] It should be noted that for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0114] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0115] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interface, and the indirect coupling or communication connection of the device or unit can be electrical or other forms.

[0116] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0117] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0118] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0119] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0120] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A bilateral network partitioning method, comprising: Obtaining historical interaction data of a first object group, and determining a bilateral object group based on the historical interaction data, where the bilateral object group includes at least one first object in the first object group and at least one second object in a second object group, and the second object has an interaction history with at least one of the first objects; Constructing a probability model with each individual object in the bilateral object group as a node and the data attribute information of the target interaction data of the individual object as the feature vector connecting the nodes, and assigning node attribute information to each node; Determining an optimization method for the probability model based on a target classification rule, and processing and expanding the nodes and / or feature vectors in the probability model based on the optimization method; Performing node clustering on each feature vector in the processed probability model, and dividing each node into at least one bilateral network set.

2. The method according to claim 1, wherein the obtaining historical interaction data of a first object group comprises: Determining the interaction type of a target second object, and setting data acquisition conditions based on the interaction type, where the data acquisition conditions include a time condition and a space condition; Obtaining historical interaction data of a first object group based on the data acquisition conditions.

3. The method according to claim 2, wherein the determining a bilateral object group based on the historical interaction data comprises: Calculating the correlation value between each data type of each historical interaction data and the interaction type respectively, screening out invalid data with a correlation value less than a preset correlation value, and obtaining screened data; Constructing a bilateral object group based on the screened data.

4. The method according to claim 1, wherein the constructing a probability model with each individual object in the bilateral object group as a node and the data attribute information of the target interaction data of the individual object as the feature vector connecting the nodes comprises: Determining the individual interaction data of each individual object in the bilateral object group, and clustering each individual interaction data based on the data type to obtain target interaction data; Constructing a probability model with each individual object as a node and the data attribute information of the target interaction data as the feature vector connecting the nodes.

5. The method according to claim 1, after constructing the probability model, further comprising: Performing data distribution analysis on the probability model, and deleting outlier samples in the probability model.

6. The method according to claim 1, wherein the determining an optimization method for the probability model based on a target classification rule comprises: Determining a target classification rule in a preset classification rule library, and determining an optimization method for the probability model based on the target classification rule and the number of bilateral groups, where the classification rule library includes a first classification rule classified based on the data attribute information and a second classification rule classified based on the node attribute information and the data attribute information.

7. The method according to claim 1, the method further comprising: Setting labels for the bilateral network sets based on the network set attribute information of the bilateral network sets.

8. The method according to claim 1, the method further comprising: Obtain the current decision-making information and determine the first network range corresponding to the current decision-making information; Determine the target bilateral network set that matches the current decision-making information, and determine the second network range of the target bilateral network set; Generate decision-making recommendation information based on the non-intersecting range of the first network range and the second network range.

9. A bilateral network partitioning device, comprising: An acquisition module, configured to acquire historical interaction data of a first object group, and determine a bilateral object group based on the historical interaction data, where the bilateral object group includes at least one first object in the first object group and at least one second object in a second object group, and the second object has an interaction history with at least one of the first objects; A construction module, configured to construct a probability model with each individual object in the bilateral object group as a node and the data attribute information of the target interaction data of the individual object as a feature vector connecting the nodes, and assign node attribute information to each of the nodes; An expansion module, configured to determine an optimization method for the probability model based on a target classification rule, and process and expand the nodes and / or feature vectors in the probability model based on the optimization method; A partitioning module, configured to perform node clustering on each of the feature vectors in the processed probability model, and partition each of the nodes into at least one bilateral network set.

10. The device according to claim 9, wherein the acquisition module is specifically configured to: Determine the interaction type of the target second object, and set data acquisition conditions based on the interaction type, where the data acquisition conditions include a time condition and a space condition; Acquire the historical interaction data of the first object group based on the data acquisition conditions.

11. The device according to claim 10, wherein the acquisition module is specifically further configured to: Calculate the correlation value between each data type of each of the historical interaction data and the interaction type respectively, screen out invalid data with a correlation value less than a preset correlation value, and obtain screened data; Construct a bilateral object group based on the screened data.

12. The device according to claim 9, wherein the construction module is specifically configured to: Determine the individual interaction data of each individual object in the bilateral object group, perform clustering on each of the individual interaction data based on the data type, and obtain target interaction data; Construct a probability model with each individual object as a node and the data attribute information of the target interaction data as a feature vector connecting the nodes.

13. The device according to claim 9, wherein the construction module is specifically further configured to: Perform data distribution analysis on the probability model, and delete the outlier samples in the probability model.

14. The device according to claim 9, wherein the expansion module is specifically configured to: Determine a target classification rule in a preset classification rule library, and determine an optimization method for the probability model based on the target classification rule and the number of bilateral groups, where the classification rule library includes a first classification rule classified based on the data attribute information and a second classification rule classified based on the node attribute information and the data attribute information.

15. The device according to claim 9, the device further includes: A setting module, configured to set a label for the bilateral network set based on the network set attribute information of the bilateral network set.

16. The apparatus according to claim 9, wherein the apparatus further comprises: A first determination module, configured to obtain current decision information and determine a first network range corresponding to the current decision information; A second determination module, configured to determine a target bilateral network set that matches the current decision information and determine a second network range of the target bilateral network set; A generation module, configured to generate decision recommendation information based on a non-intersecting range between the first network range and the second network range.

17. An electronic device, comprising a processor and a memory; The processor is connected to the memory; The memory is configured to store executable program code; The processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the method according to any one of claims 1-8.

18. A computer-readable storage medium, on which a computer program is stored, wherein instructions are stored in the computer-readable storage medium, and when the instructions are run on a computer or a processor, the computer or the processor is caused to execute the steps of the method according to any one of claims 1-8.

Images