A method, device, equipment and storage medium for establishing a multi-person session

Abstract

The embodiment of the present invention discloses a method, device, equipment and storage medium for establishing a multi-person session, the method comprising: determining a behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area; inputting the behavior data sequence into a dynamic vector representation model, and determining a target vector corresponding to each target user according to the output of the dynamic vector representation model; clustering each target user according to a preset clustering algorithm and each target vector, obtaining each target user set, and establishing a multi-person session corresponding to each target user set. The technical solution of the embodiment of the present invention can improve the accuracy of session establishment.

Description

Technical Field

[0001] The embodiments of the present invention relate to computer technology, and in particular to a method, apparatus, device and storage medium for establishing a multi-person session. Background Art

[0002] In the Internet age, the social characteristics of people's shopping are becoming more and more obvious. Most buyers have their own interest circles, that is, groups of users with the same interests and hobbies, and interest circles often affect their purchasing decisions.

[0003] In the prior art, statistical preference features of hundreds of dimensions are usually developed manually to characterize users, and the interests of each user are analyzed based on the characterization results, so that a multi-person session including users with the same interests can be established for communication.

[0004] However, in the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art:

[0005] Since each preference feature in the prior art is manually designed based on business experience and it is impossible to dig out abstract features more deeply, the characterization results cannot accurately reflect the interests and hobbies of users, thereby reducing the accuracy of session establishment. Summary of the invention

[0006] The embodiments of the present invention provide a method, apparatus, device and storage medium for establishing a multi-person session, so as to improve the accuracy of session establishment.

[0007] In a first aspect, an embodiment of the present invention provides a method for establishing a multi-person session, including:

[0008] Determine a behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area;

[0009] Inputting the behavior data sequence into a dynamic vector representation model, and determining a target vector corresponding to each target user according to an output of the dynamic vector representation model;

[0010] According to a preset clustering algorithm and each of the target vectors, each of the target users is clustered to obtain each target user set, and a multi-person session corresponding to each of the target user sets is established.

[0011] In a second aspect, an embodiment of the present invention further provides a multi-person session establishment device, including:

[0012] A behavior data sequence determination module, used to determine the behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area;

[0013] A target vector determination module, used to input the behavior data sequence into a dynamic vector representation model, and determine a target vector corresponding to each target user according to an output of the dynamic vector representation model;

[0014] The multi-person session establishing module is used to cluster the target users according to a preset clustering algorithm and the target vectors, obtain target user sets, and establish a multi-person session corresponding to each target user set.

[0015] In a third aspect, an embodiment of the present invention further provides a device, the device comprising:

[0016] one or more processors;

[0017] A memory for storing one or more programs;

[0018] When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the multi-person session establishment method provided by any embodiment of the present invention.

[0019] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for establishing a multi-person session as provided in any embodiment of the present invention.

[0020] The embodiments of the above invention have the following advantages or beneficial effects:

[0021] By determining the behavior data sequence of each target user in a preset area within a preset time, and inputting each behavior data sequence into a dynamic vector characterization model, a dynamic target vector corresponding to each target user can be obtained according to the output of the dynamic vector characterization model, so that the obtained target vector can more accurately reflect the interests and hobbies of the user within the preset time, and all target users in the preset area are clustered based on the target vector, so that a set of target users with the same interests and hobbies can be accurately obtained, and a corresponding multi-person session can be established based on each target user set, thereby improving the accuracy of session establishment. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a flowchart of a method for establishing a multi-person session provided in Embodiment 1 of the present invention;

[0023] Figure 2 This is an example of an information sending system involved in Embodiment 1 of the present invention;

[0024] Figure 3 is a flow chart of a method for establishing a multi-person session provided in Embodiment 2 of the present invention;

[0025] Figure 4 It is a structural diagram of a multi-person session establishment device provided in Embodiment 3 of the present invention;

[0026] Figure 5 It is a structural schematic diagram of a device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0027] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0028] Embodiment 1

[0029] Figure 1 This is a flowchart of a method for establishing a multi-person session provided in Embodiment 1 of the present invention. This embodiment can be applied to the case where the hobbies of users in a preset area are analyzed to establish a multi-person session consisting of users with the same hobbies. The method can be executed by a multi-person session establishment device, which can be implemented in software and / or hardware and integrated into a device for monitoring user behavior, such as a server. The method specifically includes the following steps:

[0030] S110: Determine a behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area.

[0031] Among them, the preset area can be a geographical area preset based on business scenarios and population distribution. Exemplarily, the preset area can refer to a geographical area where a cell or multiple adjacent cells are located to improve the trust between users. The target user can refer to an online user located in the preset area. The online user can refer to a user who has interacted with an online e-commerce company through the network. Among them, the interactive behavior can refer to but is not limited to browsing, adding to a shopping cart, placing an order, searching, and clicking on items. The preset time can be a historical time period preset based on business needs, such as within the past month. The behavior data sequence can refer to a sequence composed of the information of items with which the target user has interacted within the preset time. Exemplarily, the behavior data sequence can be a sequence composed of the names of items with which the target user has interacted within the preset time in chronological order.

[0032] Specifically, each user in a preset area can be selected from all online users according to their geographical location as the target user. This embodiment can mine the behavior log of each target user within a preset time, determine all items with which the target user has interactive behavior, and determine the corresponding behavior data sequence based on the item information.

[0033] S120: Input the behavior data sequence into a dynamic vector representation model, and determine a target vector corresponding to each target user according to an output of the dynamic vector representation model.

[0034] Among them, the dynamic vector representation model can refer to a model that dynamically represents word vectors based on the context of the word, which can more accurately reflect the semantics. For example, when the context of a word is different, the dynamic vector corresponding to the word is also different, thereby solving the problem that static vector representation models in the prior art, such as Word2Vector and TF-IDF (TermFrequency-Inverse Document Frequency), cannot accurately represent semantics. Exemplarily, the dynamic vector representation model can be but is not limited to the Bert network model. The Bert network model is pre-trained using two unsupervised prediction tasks, namely: (1) Masked LM task: Bert randomly masks some input words and only predicts those masked words. The final hidden vector corresponding to the masked words is input into the output softmax on the vocabulary to train a deep bidirectional network representation. (2) Next sentence prediction task: Many important NLP (Natural Language Processing) downstream tasks are based on understanding the relationship between two sentences. In order to train a model that understands sentence relationships, a binary next sentence prediction model is pre-trained. For example, when sentences A and B are selected as pre-training samples, there is a 50% chance that B is the next sentence of A, and there is also a 50% chance that it is a random sentence from the corpus. After the Bert network model is pre-trained, the Bert network model is Transformer micro-debugged based on the business corpus, so that the debugged Bert network model can be used to determine the target vector corresponding to each target user.

[0035] Specifically, the behavior data sequence corresponding to each target user is input into the dynamic vector representation model. Based on the context of each word in the behavior data sequence, the dynamic vector representation model can be used to dynamically represent each word vector, and the word vectors are integrated to output the target vector corresponding to each behavior data sequence, so that the target vector corresponding to each target user can be obtained.

[0036] S130 , clustering target users according to a preset clustering algorithm and target vectors to obtain target user sets, and establishing a multi-person session corresponding to each target user set.

[0037] The preset clustering algorithm may refer to a method for classifying users, which may be, but is not limited to, an LDA (Latent Dirichlet Allocation, document topic generation model) method and a K-Means (K-means) clustering method. The target user set may be composed of target users with the same interests and hobbies. The target user set corresponds to a multi-person session one-to-one. A multi-person session may refer to a session composed of all target users in the target user set, so as to facilitate information sharing and communication between users.

[0038] Specifically, this embodiment can use the target vector corresponding to each target user to calculate the similarity between target users. The higher the similarity, the higher the possibility that the users have the same interests and hobbies, so that the target users with higher similarity can be clustered together to generate each target user set. Exemplarily, the process of clustering target users using the LDA method is: randomly specify a topic as the initial topic according to the target vector corresponding to each target user, and resample the topic to which each target user belongs by using the Bayesian statistical Gibbs Sampling formula in LDA until the Gibbs Sampling formula converges. After convergence, the target users under each topic are users with the same interests and hobbies, so that the target users under each topic can be grouped into a target user set, thereby more conveniently obtaining each target user set. After obtaining each target user set, a corresponding multi-person session can be created more accurately for each target user set.

[0039] The technical solution of this embodiment determines the behavior data sequence of each target user in a preset area within a preset time, and inputs each behavior data sequence into a dynamic vector characterization model, so that a dynamic target vector corresponding to each target user can be obtained according to the output of the dynamic vector characterization model, so that the obtained target vector can more accurately reflect the interests and hobbies of the user within the preset time, and cluster all target users in the preset area based on the target vector, so that a set of target users with the same interests and hobbies can be accurately obtained, and a corresponding multi-person session can be established based on each target user set, thereby improving the accuracy of session establishment.

[0040] Based on the above technical solution, S110 may include: determining the behavior log text of each target user within a preset time; based on a preset mapping relationship, mapping each item name in each behavior log text to a corresponding item identifier; based on the order of items in the behavior log text, sorting each item identifier to obtain a behavior data sequence corresponding to each target user.

[0041] Among them, the behavior log text may refer to the recorded text of the interactive behavior generated within a preset time. The preset mapping relationship may refer to the correspondence between the item name and the item identifier, which can be pre-set according to the business scenario and requirements. The item identifier may refer to but is not limited to at least one of numbers, letters and symbols. For example, the item name is: Spring Lace Dress, and its corresponding item identifier is 1.

[0042] Specifically, the present embodiment can identify the names of various items in the behavior log text, and determine the item identifier corresponding to each item name based on a preset mapping relationship, and sort the item identifiers according to the arrangement order of the item names in the behavior log text, while maintaining the original arrangement order of the items, thereby obtaining a behavior data sequence composed of the item identifiers. Compared with the behavior data sequence composed of the item names, the behavior data sequence can be greatly simplified, thereby improving the processing efficiency of the dynamic vector representation model.

[0043] On the basis of the above technical solution, before "determining the behavior log text of each target user within the preset time", it can also include: determining the user delivery address corresponding to each online user; determining each online user located in the preset area as the first user according to the delivery address of each user; determining the number of interaction behaviors of each first user within the preset time; and determining the first user whose number of interaction behaviors is greater than or equal to the preset number threshold as the target user.

[0044] The user delivery address may refer to a delivery address commonly used by the user, which is used to represent the user's geographical location. The first user may refer to an online user located in a preset area. The preset number threshold may refer to the minimum number of interactive behaviors within a preset time.

[0045] Specifically, this embodiment can mine the behavior logs of online users to determine the user delivery address corresponding to each online user. Each user delivery address can be reverse geocoded to obtain the longitude and latitude information of each user delivery address, and based on the longitude and latitude range corresponding to the preset area, it is detected whether each online user is located in the preset area, so that the first user located in the preset area can be obtained. This embodiment can count the interactive behavior of each first user based on the behavior log of each first user within the preset time, obtain the corresponding number of interactive behaviors, and determine the first user whose number of interactive behaviors is greater than or equal to the preset number threshold as the target user, so as to obtain the target user with higher activity, avoid the situation where the data item is 0 due to low activity and the target vector cannot be accurately determined, and further improve the accuracy of establishing multi-person sessions.

[0046] On the basis of the above technical solution, when the preset clustering algorithm is the K-Mean mean clustering algorithm, S130 may include: taking each preset initial centroid vector in each initial user set as the first centroid vector, and taking the initial user set as the current user set, wherein the initial user set is an empty user set; calculating the similarity between each target user and each current user set according to each target vector and each first centroid vector; dividing each target user into the current user set with the greatest similarity; determining the second centroid vector corresponding to each current user set according to the target vector corresponding to each target user in each current user set; if the first centroid vector is different from the corresponding second centroid vector, updating the first centroid vector to the second centroid vector, and returning to execute the operation of calculating the similarity between each target user and each current user set based on the Euclidean distance formula, each target vector and each first centroid vector; if the first centroid vector is the same as the corresponding second centroid vector, determining each current user set as the target user set.

[0047] Among them, the initial user set is a preset empty user set, and the number K of the initial user sets can be preset. Each initial user set corresponds to a different preset initial centroid vector. The preset initial centroid vector may refer to the center vector of the initial user set, which can be preset based on business data. The current user set may refer to each user set during the clustering, such as during the first clustering, the current user set refers to each initial user set. The first centroid vector may refer to the centroid vector of each current user set during the clustering, such as during the first clustering, the first centroid vector refers to each preset initial centroid vector.

[0048] Specifically, this embodiment can calculate the Euclidean distance between the target vector corresponding to each target user and the first centroid vector of each current user set based on the Euclidean distance formula, and the Euclidean distance can reflect the size of the similarity, that is, the larger the Euclidean distance, the smaller the similarity. The target user can be divided into the current user set with the greatest similarity to the target user, that is, the current user set with the smallest Euclidean distance, so that all target users can be divided into the corresponding current user set to update the current user set. After the clustering is completed, the target vector corresponding to each target user included in the current user set can be averaged, and the obtained average target vector is used as the second centroid vector of the current user set, so that the second centroid vector corresponding to each current user set can be determined, and by comparing whether the first centroid vector and the second centroid vector corresponding to the same current user set are the same, it is determined whether the clustering of the target user is completed. If there is a situation where the first centroid vector is different from the corresponding second centroid vector, it indicates that the centroid of the current user set has changed. At this time, each first centroid vector can be updated to the corresponding second centroid vector, and the next clustering is performed based on the updated first centroid vectors. If each first centroid vector is the same as the corresponding second centroid vector, it means that the centroid of the current user set has not changed after clustering, that is, clustering is completed. At this time, the current user set can be directly used as the target user set, so that each target user set can be determined more conveniently based on the K-Mean clustering algorithm.

[0049] On the basis of the above technical solution, after S130, it also includes: determining the first session to which each target user belongs according to the target users included in each session; sending the information of the first session to the user terminal corresponding to the corresponding target user, so as to display the information of the first session on the display interface of the user terminal.

[0050] The information of the first session may refer to, but is not limited to, a session number, a session name, and session profile information. The user terminal may refer to a terminal device used by a target user, such as a smart phone.

[0051] Specifically, after each session is created, a corresponding session number may be assigned to each session, and a corresponding session name, session brief information, etc. may be set based on the target user information contained in the session. Figure 2 An example of an information sending system is given. Figure 2 The server in the context may refer to a device used to establish a multi-person session. Figure 2 The network in the text may refer to the medium that provides the communication link. The network may include various connection types, such as wired, wireless or optical cables. The server may interact with each user terminal device through the network. Figure 2As shown, the server can send the information of the first session to the user terminal corresponding to each target user, so that the information of the corresponding first session is displayed on the display interface of each user terminal, thereby reminding the user to join the corresponding session for communication. Exemplarily, the user terminal can display a message in the form of a poster "Your exclusive session has found you" so that users with the same interests can be added to the same session, so that they can share shopping experiences, recommend items, and group orders with each other in the session. It should be noted that since the target users in this embodiment are all users in the preset area, the degree of trust between users can be increased when communicating in the session, thereby improving the recommendation effect and purchase rate of items.

[0052] Embodiment 2

[0053] Figure 3 This is a flow chart of a method for establishing a multi-person session provided in Embodiment 2 of the present invention. Based on the above embodiment, this embodiment further includes determining a session manager for each session after establishing a multi-person session corresponding to each target user set, so as to better manage the session. Explanations of terms that are the same or corresponding to those in the above embodiment are not repeated here.

[0054] See also Figure 3 The method for establishing a multi-person session provided in this embodiment specifically includes the following steps:

[0055] S210: Determine a behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area.

[0056] S220: Input the behavior data sequence into the dynamic vector representation model, and determine the target vector corresponding to each target user according to the output of the dynamic vector representation model.

[0057] S230 , clustering target users according to a preset clustering algorithm and target vectors to obtain target user sets, and establishing a multi-person session corresponding to each target user set.

[0058] S240: Determine the activity level corresponding to each target user according to the behavior data corresponding to each target user.

[0059] The behavior data may refer to the interactive behavior data generated by the target user in the online e-commerce, such as shopping behavior data. The activity level may refer to the degree of interaction of the user in the online e-commerce.

[0060] Specifically, the number of interactive behaviors in the behavioral data of each target user may be counted to determine the total number of interactions of each target user, and the total number of interactions may be directly used as the activity level corresponding to the target user.

[0061] S250: In each session, the target user with the highest activity is used as a pending session manager in the session, and a session management request message is sent to the pending user terminal of the pending session manager.

[0062] The session manager may refer to a user who manages the session so as to organize users to conduct group purchases and item distribution, etc. The session management request message may refer to a message for requesting whether to agree to become a session manager.

[0063] Specifically, this embodiment may use the target user with the highest activity in the session as a pending session manager, and seek the consent of the pending session manager by sending a session management request message.

[0064] S260: After receiving the session management consent message sent by the pending user terminal, use the pending session manager as the session manager of the corresponding session.

[0065] Specifically, if the pending session manager agrees to become a session manager, that is, after receiving the session management consent message, the pending session manager can be used as the session manager of the session. In the session, the session manager can first collect the purchase needs of offline users, and then generate and match better and more accurate items, and deliver the items directly to the session manager, who will distribute the items. In this way, the pre-sale model can be adopted to effectively reduce warehousing and operating costs, while also saving labor costs. It should be noted that if a session management rejection message is received from a pending user terminal, a session management request message can be sent to the target user with the second highest activity based on the order of activity from large to small; a target user outside the session who agrees to become a session manager can also be added to the session to become the session manager of the session.

[0066] The technical solution of this embodiment uses the target user with the highest activity in each session and who agrees to become a session manager as the session manager of the session, so that the session manager can be used to manage the session and distribute items, thereby improving the item promotion effect and purchase rate.

[0067] On the basis of the above scheme, after establishing a multi-person session corresponding to each target user set, it may also include: when a newly added first target user is detected in a preset area, obtaining a first behavior data sequence of the first target user; inputting the first behavior data sequence into a dynamic vector representation model, and determining a first target vector corresponding to the first target user according to an output of the dynamic vector representation model; inputting the first target vector into a pre-trained preset network model, and determining the session to which the first target user belongs according to the output of the preset network model.

[0068] Among them, the preset network model may refer to but is not limited to a decision tree model, a random forest model, a neural network model, etc. Specifically, before using the preset network model, the present embodiment may use the target vector corresponding to each target user and the corresponding session as training samples to train the preset network model until the training error is less than the preset error, indicating that the training of the preset network model is completed. When a new first target user is detected in the preset area, the dynamic vector characterization model may be used to determine the first target vector corresponding to the first target user, and the first target vector may be used as the input of the preset network model. According to the output of the preset network model, the session to which the first target user belongs is determined, so that the first target user can be added to the corresponding session, realizing real-time addition and update of users in the session.

[0069] The following is an embodiment of a multi-person session establishment device provided in an embodiment of the present invention. The device and the multi-person session establishment method of the above-mentioned embodiments belong to the same inventive concept. For details not described in detail in the embodiment of the multi-person session establishment device, reference can be made to the embodiment of the above-mentioned multi-person session establishment method.

[0070] Embodiment 3

[0071] Figure 4 This is a structural diagram of a multi-person conversation establishment device provided in Example 3 of the present invention. This embodiment is applicable to the situation where a hobby analysis is performed on users in a preset area to establish a multi-person conversation composed of users with the same interests and hobbies. The device specifically includes: a behavior data sequence determination module 310, a target vector determination module 320 and a multi-person conversation establishment module 330.

[0072] Among them, the behavior data sequence determination module 310 is used to determine the behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area; the target vector determination module 320 is used to input the behavior data sequence into the dynamic vector characterization model, and determine the target vector corresponding to each target user according to the output of the dynamic vector characterization model; the multi-person session establishment module 330 is used to cluster each target user according to a preset clustering algorithm and each target vector, obtain each target user set, and establish a multi-person session corresponding to each of the target user sets.

[0073] Optionally, the behavior data sequence determination module 310 is specifically configured to:

[0074] Determine the behavior log text of each target user within a preset time; based on a preset mapping relationship, map each item name in each behavior log text to a corresponding item identifier; based on the order of items in the behavior log text, sort the item identifiers to obtain a behavior data sequence corresponding to each target user.

[0075] Optionally, the device also includes: a target user determination module, which is used to: determine the user delivery address corresponding to each online user before determining the behavior log text of each target user within a preset time; determine each online user located in a preset area as a first user based on the delivery address of each user; determine the number of interaction behaviors of each first user within the preset time; and determine the first user whose number of interaction behaviors is greater than or equal to a preset number threshold as a target user.

[0076] Optionally, the dynamic vector representation model is a Bert network model.

[0077] Optionally, the preset clustering algorithm is: K-Mean mean clustering algorithm; accordingly, the target multi-person conversation establishing module 330 is specifically used for:

[0078] Each preset initial centroid vector in each initial user set is used as the first centroid vector, and the initial user set is used as the current user set, wherein the initial user set is an empty user set; based on each target vector and each first centroid vector, the similarity between each target user and each current user set is calculated; each target user is divided into the current user set with the largest similarity; based on the target vector corresponding to each target user in each current user set, the second centroid vector corresponding to each current user set is determined; if the first centroid vector is different from the corresponding second centroid vector, the first centroid vector is updated to the second centroid vector, and the operation of calculating the similarity between each target user and each current user set based on the Euclidean distance formula, each target vector and each first centroid vector is returned; if the first centroid vector is the same as the corresponding second centroid vector, each current user set is determined as the target user set.

[0079] Optionally, the device further comprises:

[0080] A first session determination module, configured to determine, after establishing a multi-person session corresponding to each target user set, a first session to which each target user belongs based on the target users included in each session;

[0081] The information sending module of the first conversation is used to send the information of the first conversation to the user terminal corresponding to the corresponding target user, so as to display the information of the first conversation on the display interface of the user terminal.

[0082] Optionally, the device further comprises:

[0083] An activity determination module, used to determine the activity of each target user according to the behavior data of each target user after establishing a multi-person session corresponding to each target user set;

[0084] A session management request message sending module is used to, in each session, use the target user with the highest activity as the pending session manager in the session, and send a session management request message to the pending user terminal of the pending session manager;

[0085] The session manager determining module is used to use the pending session manager as the session manager of the corresponding session after receiving the session management consent message sent by the pending user terminal.

[0086] The multi-person session establishment device provided in the embodiment of the present invention can execute the multi-person session establishment method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the multi-person session establishment method.

[0087] Embodiment 4

[0088] Figure 5 A schematic diagram of the structure of a device provided in Embodiment 4 of the present invention. Figure 5 A block diagram of an exemplary device 12 suitable for use in implementing embodiments of the present invention is shown. Figure 5 The device 12 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0089] like Figure 5 As shown, device 12 is in the form of a general purpose computing device. Components of device 12 may include, but are not limited to, one or more processors or processing units 16, system memory 28, and bus 18 that connects various system components (including system memory 28 and processing unit 16).

[0090] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. By way of example, these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0091] Device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by device 12, including volatile and non-volatile media, removable and non-removable media.

[0092] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. Device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Figure 5 not shown, usually called a "hard drive"). Although Figure 5 Not shown in the figure, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 18 via one or more data medium interfaces. The system memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.

[0093] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. Program modules 42 generally perform the functions and / or methods of the embodiments described herein.

[0094] The device 12 may also communicate with one or more external devices 14 (e.g., keyboards, pointing devices, displays 24, etc.), may communicate with one or more devices that enable a user to interact with the device 12, and / or may communicate with any device that enables the device 12 to communicate with one or more other computing devices (e.g., network cards, modems, etc.). Such communication may be performed via an input / output (I / O) interface 22. Furthermore, the device 12 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 20. As shown, the network adapter 20 communicates with other modules of the device 12 via a bus 18. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the device 12, including, but not limited to, microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0095] The processing unit 16 executes various functional applications and data processing by running the program stored in the system memory 28, for example, implementing a multi-person session establishment method provided in an embodiment of the present invention, the method comprising:

[0096] Determine a behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area;

[0097] Input the behavior data sequence into the dynamic vector representation model, and determine the target vector corresponding to each target user according to the output of the dynamic vector representation model;

[0098] According to the preset clustering algorithm and each target vector, each target user is clustered to obtain each target user set, and a multi-person session corresponding to each target user set is established.

[0099] Of course, those skilled in the art can understand that the processor can also implement the technical solution of the multi-person session establishment method provided by any embodiment of the present invention.

[0100] Embodiment 5

[0101] This embodiment 5 provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the steps of a method for establishing a multi-person session provided in any embodiment of the present invention are implemented. The method includes:

[0102] Determine a behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area;

[0103] Input the behavior data sequence into the dynamic vector representation model, and determine the target vector corresponding to each target user according to the output of the dynamic vector representation model;

[0104] According to the preset clustering algorithm and each target vector, each target user is clustered to obtain each target user set, and a multi-person session corresponding to each target user set is established.

[0105] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, 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 above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0106] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0107] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0108] Computer program code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0109] It should be understood by those skilled in the art that the modules or steps of the present invention described above can be implemented by a general-purpose computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, optionally, they can be implemented by a program code executable by a computer device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0110] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for establishing a multi-person session, characterized in that: include: Determine a behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area; the behavior data sequence is a sequence consisting of item information with which the target user generates interactive behavior within the preset time; Input the behavior data sequence into a dynamic vector representation model, and determine the target vector corresponding to each target user according to the output of the dynamic vector representation model; the dynamic vector representation model is a model that dynamically represents word vectors based on the context of words; According to a preset clustering algorithm and each of the target vectors, each of the target users is clustered according to vector similarity to obtain each target user set, and a multi-person session corresponding to each of the target user sets is established; the target vector is used to reflect the interests and hobbies of the target user within a preset time; the target user set is a user set composed of target users with the same interests and hobbies; the target user set has a one-to-one correspondence with the multi-person session; the multi-person session is a session composed of all the target users in the target user set, so that the target users can share information and communicate through the multi-person session; When a newly added first target user is detected in a preset area, a first behavior data sequence of the first target user is obtained; the first behavior data sequence is input into a dynamic vector characterization model, and a first target vector corresponding to the first target user is determined according to an output of the dynamic vector characterization model; the first target vector is input into a pre-trained preset network model, and a session to which the first target user belongs is determined according to an output of the preset network model.

2. The method according to claim 1, characterized in that Determine the behavioral data sequence of each target user within a preset time, including: Determine the behavior log text of each target user within a preset time; Based on a preset mapping relationship, each item name in each of the behavior log texts is mapped to a corresponding item identifier; Based on the order of items in the behavior log text, the item identifiers are sorted to obtain a behavior data sequence corresponding to each target user.

3. The method according to claim 2, characterized in that Before determining the behavior log text of each target user within the preset time, it also includes: Determine the user delivery address corresponding to each online user; According to the delivery addresses of the users, each online user in the preset area is determined as a first user; Determining the number of interaction behaviors of each of the first users within a preset time; The first user whose number of interaction behaviors is greater than or equal to a preset number threshold is determined as a target user.

4. The method according to claim 1, characterized in that The dynamic vector representation model is a Bert network model.

5. The method according to claim 1, characterized in that The preset clustering algorithm is: K-Mean clustering algorithm; Accordingly, clustering the target users according to the preset clustering algorithm and the target vectors to obtain target user sets includes: Taking each preset initial centroid vector in each initial user set as a first centroid vector, and taking the initial user set as a current user set, wherein the initial user set is an empty user set; Calculating the similarity between each of the target users and each current user set according to each of the target vectors and each of the first centroid vectors; Dividing each of the target users into the current user set with the greatest similarity; Determine a second centroid vector corresponding to each current user set according to the target vector corresponding to each target user in each current user set; If the first centroid vector is different from the corresponding second centroid vector, the first centroid vector is updated to the second centroid vector, and the operation of calculating the similarity between each target user and each current user set based on the Euclidean distance formula, each target vector and each first centroid vector is returned; If the first centroid vector is the same as the corresponding second centroid vector, each current user set is determined as a target user set.

6. The method according to claim 1, characterized in that After establishing a multi-person session corresponding to each target user set, the method further includes: According to the target users included in each of the sessions, determining the first session to which each of the target users belongs; The information of the first conversation is sent to a user terminal corresponding to the corresponding target user, so that the information of the first conversation is displayed on a display interface of the user terminal.

7. The method according to any one of claims 1 to 6, characterized in that: After establishing a multi-person session corresponding to each target user set, the method further includes: Determine the activity level of each target user according to the behavior data of each target user; In each of the sessions, the target user with the highest activity is used as the pending session manager of the session, and a session management request message is sent to the pending user terminal of the pending session manager; After receiving the session management consent message sent by the pending user terminal, the pending session manager is used as the session manager of the corresponding session.

8. A device for establishing a multi-person conversation, characterized in that: include: A behavior data sequence determination module is used to determine the behavior data sequence of each target user within a preset time, wherein the target user is a user located in a preset area; the behavior data sequence is a sequence composed of item information with which the target user generates interactive behavior within the preset time; A target vector determination module, used to input the behavior data sequence into a dynamic vector representation model, and determine the target vector corresponding to each target user according to the output of the dynamic vector representation model; the dynamic vector representation model is a model that dynamically represents word vectors based on the context of words; A multi-person session establishment module is used to cluster the vector similarity of each target user according to a preset clustering algorithm and each target vector, obtain each target user set, and establish a multi-person session corresponding to each target user set; the target vector is used to reflect the interests and hobbies of the target user within a preset time; the target user set is a user set composed of target users with the same interests and hobbies; the target user set has a one-to-one correspondence with the multi-person session; the multi-person session is a session composed of all target users in the target user set, so that the target users can share information and communicate through the multi-person session; The session determination module is used to obtain a first behavior data sequence of the first target user when a newly added first target user is detected in a preset area; input the first behavior data sequence into a dynamic vector representation model, and determine a first target vector corresponding to the first target user according to the output of the dynamic vector representation model; input the first target vector into a pre-trained preset network model, and determine the session to which the first target user belongs according to the output of the preset network model.

9. A device, characterized in that: The device comprises: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the multi-person session establishment method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for establishing a multi-person session as described in any one of claims 1 to 7 are implemented.

Images