Virtual room processing method, device, electronic device and storage medium
By using environmental prediction models and room creation decision models, the problem of limited interaction scenarios for multiple people has been solved, enabling intelligent matching and creation of virtual rooms, thus improving user experience and the scale of interaction.
Patent Information
- Application Number
- CN202310149834.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-02-14
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2043-02-14
AI Technical Summary
In existing technologies, multi-person communication and interaction scenarios are limited. Live streaming mode and two-person matching mode cannot be applied to equal interaction between more than two people, and the content and atmosphere of user interaction are difficult to control.
By using an environmental prediction model and a room creation decision model, based on user behavior data, historical room data, and environmental data, the probability of creating a new room or joining an existing room is predicted, thus enabling the matching and creation of virtual rooms.
It has expanded the pool of available rooms, improved user experience and user scale, and supports equal interaction among two or more people, making it widely applicable.
Smart Images

Figure CN116155636B_ABST
Abstract
Description
Background Art
[0002] This section is intended to provide a background or context to embodiments of the invention that are recited in the claims. No description herein is admitted to be prior art by inclusion in this section.
[0003] Providing independent virtual network space for multiple online users to communicate and interact is a popular online social model today.
[0004] In the related technologies, multi-person communication and interaction are usually achieved through a live broadcast room mode or a two-person matching mode. In the live broadcast room mode, the host of the live broadcast room needs to provide exquisite content for the audience of the live broadcast room to communicate and interact. The host of the live broadcast room is required to have a higher level of professionalism than the audience of the live broadcast room. This method of relying on professional content creators to provide social space is not essentially an equal social model. It not only limits the scale of content for communication and interaction, but also is not suitable for equal communication and interaction between users. In the two-person matching mode, there is a lack of consideration of factors such as room atmosphere and user changes, and it is not suitable for interactive communication scenarios with more than two people.
[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention
[0006] The present application provides a virtual room processing method, a virtual room processing device, a computer-readable storage medium and an electronic device, thereby at least to a certain extent solving the problem of limited applicable scenarios for multi-person communication and interaction in related technologies.
[0007] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.
[0008] According to the first aspect of the present application, a virtual room processing method is provided, the method comprising: responding to a virtual room matching request of a requesting user, based on first environmental data corresponding to multiple historical sampling moments, using an environmental prediction model to perform environmental prediction, and obtaining second environmental data corresponding to future sampling moments, wherein any environmental data includes the number of matchable rooms and the number of users to be matched; based on the second environmental data, using a room creation decision model to determine a room creation decision for the requesting user; and matching a virtual room for the requesting user based on the room creation decision.
[0009] In an exemplary embodiment of the present application, the room creation decision model is used based on the second environment data to determine the room creation decision for the requesting user, including: inputting the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user and the second environment data into the room creation decision model, so that the room creation decision model predicts the probability of creating a new room and the probability of joining an already created room through feature extraction and feature intersection, and outputs the room creation decision by comparing the probability of creating a new room and the probability of joining an already created room.
[0010] In an exemplary embodiment of the present application, the prediction of the probability of creating a new room and the probability of joining an already created room through feature extraction and feature cross-linking includes: extracting features from the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data, respectively, to obtain user-side feature data, room-side feature data, and environment-side feature data; performing feature cross-linking on the user-side feature data, the room-side feature data, and the environment-side feature data to obtain feature association data; and predicting the probability of creating a new room and the probability of joining an already created room based on the feature association data.
[0011] In an exemplary embodiment of the present application, the feature cross-linking of the user-side feature data, the room-side feature data and the environment-side feature data to obtain feature association data includes: feature cross-linking the user behavior feature in the user-side feature data with the room identification feature in the room-side feature data to obtain first feature association data; feature cross-linking the matching user feature in the room-side feature data with the user identification feature in the user-side feature data to obtain second feature association data; feature cross-linking the to-be-matched user feature in the environment-side feature data with the user identification feature in the user-side feature data to obtain third feature association data; feature cross-linking the matchable room feature in the environment-side feature data with the room identification feature in the room-side feature data to obtain fourth feature association data; and determining feature association data between the user side, the room side and the environment side based on the first feature association data, the second feature association data, the third feature association data and the fourth feature association data.
[0012] In an exemplary embodiment of the present application, the method also includes: generating an environmental time series based on sample environmental data corresponding to multiple sample sampling moments; performing a stationarity test on the environmental time series, and if the environmental time series fails the stationarity test, performing a differential operation on the environmental time series until the environmental time series passes the stationarity test; using an initialized environmental prediction model to fit the environmental time series until the environmental prediction model meets a preset fitting stop condition, thereby obtaining the environmental prediction model finally used.
[0013] In an exemplary embodiment of the present application, the preset fitting stop condition includes the ring
[0014] The environmental prediction model passed the white noise test.
[0015] In an exemplary embodiment of the present application, the method further includes: training the initialized room creation decision model based on user behavior data corresponding to the sample user, historical room data corresponding to the sample user, and matching environment data corresponding to the sample user, until the room creation decision model meets a preset training stop condition to obtain the room creation decision model finally used.
[0016] In an exemplary embodiment of the present application, the initialized room creation decision model is trained based on the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user, including: inputting the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user into the room creation decision model to obtain a decision result output by the room decision model; and optimizing the model parameters of the room creation decision model based on the decision result and the positive and negative sample types corresponding to the sample user.
[0017] In an exemplary embodiment of the present application, the sample users are users who have created virtual rooms. The positive and negative sample types corresponding to the sample users who stay in the created virtual room for more than a preset time are positive samples, and the positive and negative sample types corresponding to the sample users who stay in the created virtual room for no more than a preset time are negative samples.
[0018] In an exemplary embodiment of the present application, the matching environment data corresponding to the sample user includes environment data estimated by the environment estimation model when the sample user initiates a virtual room matching request.
[0019] In an exemplary embodiment of the present application, the preset training stop condition includes that the loss function of the room creation decision model is less than a preset threshold.
[0020] According to a second aspect of the present application, a virtual room processing device is disclosed, which includes: an environment prediction module, which is used to respond to a virtual room matching request of a requesting user, and based on first environment data corresponding to multiple historical sampling moments, use an environment prediction model to perform environment prediction to obtain second environment data corresponding to future sampling moments, where any environment data includes the number of matchable rooms and the number of users to be matched; a decision determination module, which is used to determine a room creation decision for the requesting user based on the second environment data and use a room creation decision model; and a room matching module, which is used to match a virtual room for the requesting user based on the room creation decision.
[0021] In an exemplary embodiment of the present application, the decision determination module includes: a room creation decision model processing module, which is used to input user behavior data corresponding to the requesting user, historical room data corresponding to the requesting user, and the second environment data into the room creation decision model, so that the room creation decision model predicts the probability of creating a new room and the probability of joining an already created room through feature extraction and feature intersection, and outputs the room creation decision by comparing the probability of creating a new room and the probability of joining an already created room.
[0022] In an exemplary embodiment of the present application, the room creation decision model processing module includes: a feature extraction module, which is used to perform feature extraction on the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data, respectively, to obtain user-side feature data, room-side feature data, and environment-side feature data; a feature intersection module, which is used to perform feature intersection on the user-side feature data, the room-side feature data, and the environment-side feature data to obtain feature association data; a probability prediction module, which is used to predict the probability of creating a new room and the probability of joining an already created room based on the feature association data.
[0023] In an exemplary embodiment of the present application, the feature intersection module is configured to: perform feature intersection on the user behavior features in the user-side feature data and the room identification features in the room-side feature data to obtain first feature association data; perform feature intersection on the matching user features in the room-side feature data and the user identification features in the user-side feature data to obtain second feature association data; perform feature intersection on the to-be-matched user features in the environment-side feature data and the user identification features in the user-side feature data to obtain third feature association data; perform feature intersection on the matchable room features in the environment-side feature data and the room identification features in the room-side feature data to obtain fourth feature association data; determine feature association data between the user side, the room side, and the environment side based on the first feature association data, the second feature association data, the third feature association data, and the third feature association data.
[0024] In an exemplary embodiment of the present application, the device also includes: a sequence generation module, which is used to generate an environmental time series based on sample environmental data corresponding to multiple sample sampling moments; a stationarity test module, which is used to perform a stationarity test on the environmental time series. If the environmental time series fails the stationarity test, a differential operation is performed on the environmental time series until the environmental time series passes the stationarity test; and an environmental prediction model generation module, which is used to fit the environmental time series using an initialized environmental prediction model until the environmental prediction model meets a preset fitting stop condition to obtain the environmental prediction model finally used.
[0025] In an exemplary embodiment of the present application, in the environment prediction model generation module, the preset fitting stop condition includes that the environment prediction model passes the white noise test.
[0026] In an exemplary embodiment of the present application, the device also includes: a room creation decision model generation module, which is used to train the initialized room creation decision model based on user behavior data corresponding to the sample user, historical room data corresponding to the sample user, and matching environment data corresponding to the sample user, until the room creation decision model meets a preset training stop condition to obtain the room creation decision model finally used.
[0027] In an exemplary embodiment of the present application, the room creation decision model generation module includes: a decision result determination module, which is used to input the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user into the room creation decision model to obtain the decision result output by the room decision model; a model parameter optimization module, which is used to optimize the model parameters of the room creation decision model based on the decision result and the positive and negative sample types corresponding to the sample user.
[0028] In an exemplary embodiment of the present application, in the virtual room processing device, the sample users are users who have created virtual rooms, and the positive and negative sample types corresponding to the sample users who have stayed in the created virtual room for more than a preset time are positive samples, and the positive and negative sample types corresponding to the sample users who have stayed in the created virtual room for no more than a preset time are negative samples.
[0029] In an exemplary embodiment of the present application, in the room creation decision model generation module, the matching environment data corresponding to the sample user includes the environment data estimated by the environment estimation model when the sample user initiates a virtual room matching request.
[0030] In an exemplary embodiment of the present application, in the room creation decision model generation module, the preset training stop condition includes that the loss function of the room creation decision model is less than a preset threshold.
[0031] According to a third aspect of an embodiment of the present application, an electronic device is disclosed, including: a processor; and a memory, wherein computer-readable instructions are stored in the memory, and when the computer-readable instructions are executed by the processor, the virtual room processing method disclosed in the first aspect is implemented.
[0032] According to a fourth aspect of an embodiment of the present application, a computer program medium is disclosed, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a computer, the computer executes the virtual room processing method disclosed in the first aspect of the present application.
[0033] The technical solution of this application has the following beneficial effects:
[0034] In the above virtual room processing process, in response to the virtual room matching request of the requesting user, based on the first environmental data corresponding to multiple historical sampling moments, an environmental prediction model is used to perform environmental prediction to obtain the second environmental data corresponding to the future sampling moment, and any environmental data includes the number of matching rooms and the number of users to be matched; based on the second environmental data, a room creation decision model is used to determine the room creation decision for the requesting user; based on the room creation decision, a virtual room is matched for the requesting user. On the one hand, the present disclosure combines environmental data to solve the virtual room creation problem, and can enrich the available room pool while ensuring the user interaction experience, which is conducive to forming a virtuous cycle of "enrichment of the available room pool improves the user experience, improvement of the user experience leads to an increase in the user scale, and an increase in the user scale makes the available room pool more enriched". On the other hand, the present disclosure can be applied to multi-person equal interaction scenarios of two or more people. The number of people that a virtual room can accommodate is not limited to two people, and the user who creates the virtual room can be any user in the virtual room, and the applicability is relatively wide.
[0035] It should be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0037] Figure 1 A schematic diagram showing a flow chart of a method for implementing a virtual room processing in this exemplary embodiment;
[0038] Figure 2 A schematic diagram of a process for predicting the probability of creating a new room and the probability of adding an already created room in this exemplary embodiment is shown;
[0039] Figure 3 A diagram showing an example of execution logic for determining a virtual room creation decision in this exemplary embodiment is shown;
[0040] Figure 4 A schematic diagram of a process of generating an environment prediction model in this exemplary embodiment is shown;
[0041] Figure 5 A schematic diagram of a room creation decision model generation process in this exemplary embodiment is shown;
[0042] Figure 6A structural block diagram of a virtual room processing device in this exemplary embodiment is shown;
[0043] Figure 7 An electronic device for implementing the above-mentioned virtual room processing method in this exemplary embodiment is shown.
[0044] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts. DETAILED DESCRIPTION
[0045] The principles and spirit of the present application will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present application, and are not intended to limit the scope of the present application in any way. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.
[0046] Those skilled in the art know that the embodiments of the present application can be implemented as a device, apparatus, method or computer program product. Therefore, the present application can be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.
[0047] According to an embodiment of the present application, a virtual room processing method, a virtual room processing device, an electronic device, and a computer-readable storage medium are proposed.
[0048] It should be noted that the data involved in the present disclosure may be data authorized by the user or fully authorized by all parties, and the collection, dissemination, and use of the data shall comply with the requirements of relevant national laws and regulations. The implementation methods / embodiments of the present disclosure may be combined with each other. The number of any element in the drawings is for illustration rather than limitation, and any naming is only for distinction and does not have any limiting meaning.
[0049] The principles and spirit of the present application are explained in detail below with reference to several representative implementations of the present application. SUMMARY OF THE INVENTION
[0051] In the related art of this application, multi-person communication and interaction are achieved through a live broadcast room mode or a two-person matching mode. The methods adopted in the related art either require professionalism of the user who creates the room, or are limited to interaction between two people, and are not suitable for scenarios where multiple people interact equally.
[0052] Based on the above problems, this application proposes a virtual room processing method, which solves the problem of virtual room creation in multi-person equal interaction scenarios by combining environmental data. It can enrich the available room pool while ensuring the user's interactive experience, and has a wide range of applicability.
[0053] Application Scenario Overview
[0054] It should be noted that the following application scenarios are only shown to facilitate understanding of the spirit and principle of the present application, and the implementation of the present application is not limited in this respect. On the contrary, the implementation of the present application can be applied to any applicable scenario.
[0055] Exemplarily, it can be applied to a virtual room matching scenario in a multi-person music social setting, in which users in the same virtual room can enjoy songs together, communicate and interact, share listening experiences, and so on.
[0056] Exemplary Methods
[0057] In combination with the above application scenarios, refer to Figure 1 and Figure 5 To describe a virtual room processing method according to an exemplary embodiment of the present application.
[0058] See also Figure 1 , Figure 1 FIG. 1 is a flow chart of a virtual room processing method according to an exemplary embodiment of the present application. Figure 1 As shown, the virtual room processing method may include:
[0059] Step S110, in response to the virtual room matching request of the requesting user, based on the first environment data corresponding to multiple historical sampling moments, an environment prediction model is used to perform environment prediction to obtain second environment data corresponding to a future sampling moment, where any environment data includes the number of matching rooms and the number of users to be matched;
[0060] Step S120, based on the second environment data, using the room creation decision model to determine a room creation decision for the requesting user;
[0061] Step S130: matching a virtual room for the requesting user based on the room creation decision.
[0062] Implementation Figure 1The virtual room processing method shown, on the one hand, combines environmental data to solve the virtual room creation problem, and can enrich the available room pool while ensuring the user interaction experience, which is conducive to forming a virtuous cycle of "the enrichment of the available room pool improves the user experience, the improvement of the user experience leads to the growth of the user scale, and the growth of the user scale makes the available room pool richer". On the other hand, it can be applied to multi-person equal interaction scenarios of two or more people. The number of people that a virtual room can accommodate is not limited to two people, and the user who creates the virtual room can be any user in the virtual room, which has a wide range of applicability.
[0063] These steps are described in detail below.
[0064] In step S110, in response to the virtual room matching request of the requesting user, based on the first environment data corresponding to multiple historical sampling moments, an environment prediction model is used to perform environment prediction to obtain second environment data corresponding to future sampling moments. Any environment data includes the number of matchable rooms and the number of users to be matched.
[0065] The requesting user refers to the user who initiates the virtual room matching request.
[0066] The virtual room matching request may be a virtual room matching request generated by the terminal device in response to a room matching operation performed by a user in its graphical interface.
[0067] The historical sampling time refers to the sampling time obtained by sampling the historical period, and the future sampling time refers to the sampling time obtained by sampling the future period.
[0068] For example, sampling can be performed every 15 seconds in the historical period to form the historical sampling moment; sampling can be performed every 15 seconds in the future period to form the future sampling moment. In addition, a scheduled task can be pre-set to sample every 15 seconds to form the historical sampling moment, and continue to sample every 15 seconds based on the latest historical sampling moment to form the future sampling moment.
[0069] It should be noted that the historical period and the future period may be distinguished by the time when the requesting user initiates the virtual room matching request, or may be distinguished by the time when the virtual room matching request of the requesting user is received.
[0070] The first environment data refers to the environment data collected at the historical sampling time, and the second environment data refers to the environment data at the future sampling time estimated by the sampling environment estimation model. It should be noted that in the present disclosure, any environment data may include but is not limited to parameter data such as the number of matching rooms and the number of users to be matched.
[0071] Among them, the environment prediction model can perform environment prediction. When the model performs online prediction, the input may include first environment data, and the output may include second environment data.
[0072] In step S120, based on the second environment data, a room creation decision model is used to determine a room creation decision for the requesting user.
[0073] Among them, the room creation decision model can make room creation decisions. When the model makes online decisions, the input of the model may include second environment data, and the output may include creating a new room or joining an already created room.
[0074] In order to ensure the decision reliability of the room creation decision model, the user behavior data corresponding to the user and the historical room data corresponding to the user may also be requested as inputs of the room creation decision model.
[0075] In an optional embodiment, in step S120, based on the second environment data, a room creation decision model is used to determine the room creation decision for the requesting user, which can be specifically achieved through the following steps: the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data are input into the room creation decision model, so that the room creation decision model predicts the probability of creating a new room and the probability of joining an already created room through feature extraction and feature intersection, and outputs the room creation decision by comparing the probability of creating a new room and the probability of joining an already created room.
[0076] Exemplarily, if the probability of creating a new room exceeds the probability of joining an already created room, the output room creation decision is to create a new room; if the probability of creating a new room does not exceed the probability of joining an already created room, the output room creation decision is to join an already created room.
[0077] The expected experience of users creating new rooms and the expected experience of users joining already created rooms can be reflected by the probability of creating new rooms and the probability of joining already created rooms. If the probability of creating new rooms is low, it means that the user has a poor experience in creating new rooms, otherwise, the experience is good; if the probability of joining already created rooms is low, it means that the user has a poor experience in joining already created rooms, otherwise, the experience is good.
[0078] In the above steps, the user's expected experience is quantified by probability value, which provides data support for whether to create a new room, so that the two strategies of creating a new room and joining an already created room can be compared, so that the room creation decision model can make a decision.
[0079] In an optional implementation, the above prediction of the probability of creating a new room and the probability of adding an already created room is performed through feature extraction and feature intersection, such as Figure 2As shown, this can be achieved by the following steps:
[0080] Step S210, performing feature extraction on the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data, respectively, to obtain user-side feature data, room-side feature data, and environment-side feature data;
[0081] Step S220, performing feature cross-talk on the user-side feature data, the room-side feature data, and the environment-side feature data to obtain feature association data;
[0082] Step S230, based on the feature association data, predict the probability of creating a new room and the probability of joining an already created room.
[0083] Figure 2 In the steps shown, the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data are used to characterize the state and attributes of the user side, the state and attributes of the room side, and the state and attributes of the environment side, respectively. Probabilistic prediction is performed from three perspectives: the user side, the room side, and the environment side. The factors considered are more comprehensive, which can improve the reliability of the decision-making of the room creation decision model to a certain extent.
[0084] The user behavior data corresponding to the requesting user may be data related to the content usage preferences of the requesting user. Taking the music virtual room scene as an example, the user behavior data may include user song request data, user comment data, user like data, etc. It should be noted that the user behavior data here, in addition to the user behavior data of the user in the music virtual room scene, may also include the user behavior data of the user in non-music virtual room scenes (for example, the user's interaction data in the music community), the user's demographic data (such as age, gender, etc.), etc.
[0085] No specific limitation.
[0086] The historical room data corresponding to the requesting user may be the room data that the requesting user has historically joined or created. The historical room data may include a list of matching users in the room, the type of room, and other data.
[0087] Among them, user-side feature data refers to feature data extracted from the user side perspective, which may include user behavior features, user identification features, etc.; room-side feature data refers to feature data extracted from the room side perspective, which may include matching user features, room identification features, etc.; environment-side feature data refers to feature data extracted from the environment side perspective, which may include user features to be matched, matching room features, etc.
[0088] Among them, feature association data refers to the feature data obtained after feature cross-talk, which can reflect the degree of association between the user side, room side, and environment side.
[0089] In an optional embodiment, the above-mentioned feature cross-linking of the user-side feature data, the room-side feature data and the environment-side feature data to obtain feature association data can be achieved through the following steps: feature cross-linking the user behavior features in the user-side feature data with the room identification features in the room-side feature data to obtain first feature association data; feature cross-linking the matching user features in the room-side feature data with the user identification features in the user-side feature data to obtain second feature association data; feature cross-linking the to-be-matched user features in the environment-side feature data with the user identification features in the user-side feature data to obtain third feature association data; feature cross-linking the matchable room features in the environment-side feature data with the room identification features in the room-side feature data to obtain fourth feature association data; based on the first feature association data, the second feature association data, the third feature association data and the fourth feature association data, determine the feature association data between the user side, the room side and the environment side.
[0090] Among them, the first feature association data characterizes the matching degree between the virtual room and the user through the relationship between the virtual room and the user behavior sequence; the second feature association data indirectly characterizes the matching degree between the user and the virtual room matching user sequence; the third feature association data characterizes the matching degree between the user and the environment through the relationship between the user and the user sequence to be matched; the fourth feature association data characterizes the matching degree between the virtual room and the environment through the relationship between the virtual room and the matchable virtual room sequence.
[0091] Optionally, feature cross processing can be performed through DIN (Deep Interest Network).
[0092] Optionally, the first feature association data, the second feature association data, the third feature association data and the third feature association data can be directly used as feature association data between the user side, the room side and the environment side; the first feature association data, the second feature association data, the third feature association data and the third feature association data can also be further fused to obtain feature association data between the user side, the room side and the environment side.
[0093] In the above steps, feature cross-processing is performed from multiple dimensions to deeply characterize the degree of correlation between the user side, room side, and environment side, so as to further improve the reliability of probability assessment.
[0094] In step S130 , a virtual room is matched for the requesting user based on the room creation decision.
[0095] Exemplarily, if the room creation decision is "new room", a new room is created for the requesting user, and the requesting user is allowed to join the created new virtual room. If the room creation decision is "join an already created room", a new room is not created, and the requesting user is allowed to join the already created virtual room.
[0096] In addition, if Figure 3 As shown, an execution logic instance diagram for determining a virtual room creation decision is also provided, which includes an asynchronous estimation logic 310 for a matching environment and a synchronous decision logic 320 for a user request environment.
[0097] The matching environment asynchronous estimation logic 310 includes the following steps:
[0098] Step S311, obtaining sampling environment data through online asynchronous sampling;
[0099] Step S312, inputting the sampled environmental data into the environmental prediction model to perform environmental prediction to obtain environmental prediction data;
[0100] Step S313, storing the environment prediction data, and providing second environment data for the synchronization decision logic 320 under the user request environment.
[0101] The synchronization decision logic 320 under the user request environment includes the following steps:
[0102] Step S321, obtaining a matching request initiated by a requesting user;
[0103] Step S322, obtaining second environment data from the matching environment asynchronous estimation logic 310;
[0104] Step S323, based on the acquired second environment data, estimating the experience of the requesting user in creating a new room and the experience of joining an already created room;
[0105] Step S324, obtaining the room creation decision of the requesting user according to the experience of the requesting user in creating a new room and the experience of joining an already created room.
[0106] It should be noted that Figure 3 In the example, the experience of creating a new room and the experience of joining an already created room can be represented by the probability of creating a new room and the probability of joining an already created room, respectively. The larger the probability value, the better the experience.
[0107] The following is an explanation of the modeling and training process of the environment prediction model and the room creation decision model used above.
[0108] In an optional embodiment, the environmental prediction model finally used can be obtained through the following steps: generating an environmental time series based on sample environmental data corresponding to multiple sample sampling moments; performing a stationarity test on the environmental time series, and if the environmental time series fails the stationarity test, performing a differential operation on the environmental time series until the environmental time series passes the stationarity test; using the initialized environmental prediction model to fit the environmental time series until the environmental prediction model meets the preset fitting stop condition, thereby obtaining the environmental prediction model finally used.
[0109] The sample environment data refers to the environment data collected at the sample sampling time, and the sample sampling time refers to the time when the sample is sampled at a preset sampling frequency. For example, the sample sampling time within the last week can be obtained with a sampling frequency of 15 seconds / time, and the number of matchable rooms, the number of users to be matched, etc. at the sample sampling time are collected as sample environment data.
[0110] Specifically, the collected sample environmental data may be arranged in chronological order to obtain an environmental time series.
[0111] After obtaining the environmental time series, a stationarity test may be performed on the environmental time series to ensure that the environmental time series is stable. Optionally, in order to ensure processing efficiency, when the environmental time series satisfies wide stationarity, it may be allowed to pass the stationarity test.
[0112] For example, the environmental prediction model can be constructed using ARIMA (Autoregressive Integrated Moving Average model).
[0113] The preset fitting stop condition may include that the environment prediction model passes the white noise test. Specifically, the loss function of the environment prediction model may be tested using white noise to ensure that most of the feature information of the environment time series has been fully learned by the environment prediction model.
[0114] like Figure 4 As shown, a schematic diagram of the process of generating an environmental prediction model is provided:
[0115] Step S410, generating an environmental time series based on sample environmental data corresponding to a plurality of sample sampling moments;
[0116] Step S420, performing a stationarity test on the environmental time series to determine whether the environmental time series passes the stationarity test, if not, executing step S430, if yes, executing step S440;
[0117] Step S430, performing a differential operation on the environmental time series, and jumping to step S420;
[0118] Step S440, performing a white noise test on the environment prediction model to determine whether the environment prediction model passes the white noise test, if not, executing step S450, if yes, executing step S460;
[0119] Step S450, fitting the environmental time series, and jumping to step S440;
[0120] Step S460, obtaining the final environment prediction model.
[0121] Figure 4 The final environmental prediction model can be used to predict environmental data at future times.
[0122] In an optional embodiment, the room creation decision model finally used can be obtained through the following steps: based on the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user, the initialized room creation decision model is trained until the room creation decision model meets the preset training stop condition to obtain the room creation decision model finally used.
[0123] Among them, the sample users can be users who have created virtual rooms. The positive and negative sample types corresponding to the sample users who have stayed in the created virtual room for more than a preset time can be set as positive samples, and the positive and negative sample types corresponding to the sample users who have stayed in the created virtual room for less than a preset time can be set as negative samples. Here, the users who created the room are used as the sample set of the room creation decision model, so that when the room creation decision model is used later, the waiting cost of users entering the room can be reduced, more users can be attracted to enter the room, and they can stay stably in the created room.
[0124] Among them, the user behavior data corresponding to the sample user may be data related to the sample user's content usage preferences; the historical room data corresponding to the sample user may be room data that the sample user has historically joined or created; the matching environment data corresponding to the sample user may include the environment data estimated by the environment estimation model when the sample user initiates a virtual room matching request. Here, the environment data estimated by the environment estimation model is applied to the room creation decision model, which can make the modeling process of the room creation decision model more robust.
[0125] Exemplarily, the room creation decision model can be constructed using a DeepFM (Deep Factorization Machine) model.
[0126] Among them, the preset training stop condition may include that the loss function of the room creation decision model is less than a preset threshold.
[0127] In an optional implementation, the above-mentioned training of the initialized room creation decision model based on the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user can be specifically achieved through the following steps: inputting the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user into the room creation decision model to obtain the decision result output by the room decision model; based on the decision result and the positive and negative sample types corresponding to the sample user, optimizing the model parameters of the room creation decision model.
[0128] Specifically, after the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user are input into the room creation decision model, the room creation decision model can predict the probability of the sample user creating a new room and the probability of the sample user joining an already created room through feature extraction and feature intersection, and output the decision result by comparing the probability of the sample user creating a new room and the probability of the sample user joining an already created room, and optimize the model parameters of the room creation decision model based on the decision result and the positive and negative sample types corresponding to the sample user.
[0129] like Figure 5 As shown, a schematic diagram of the room creation decision model generation process is provided:
[0130] Step S510, inputting user behavior data corresponding to the sample user, historical room data corresponding to the sample user, and matching environment data corresponding to the sample user into a room creation decision model to obtain a decision result output by the room decision model;
[0131] Step S520, optimizing the model parameters of the room creation decision model based on the decision result and the positive and negative sample types corresponding to the sample users;
[0132] Step S530, determining whether the room creation decision model meets the preset training stop condition, if not, jump to step S510, if yes, execute step S540;
[0133] Step S540, obtaining the room creation decision model finally used.
[0134] Figure 5 The resulting room creation decision model can be used to determine the room creation decision for the requesting user.
[0135] The embodiments of the present application, on the one hand, solve the problem of virtual room creation by combining environmental data, and can enrich the available room pool while ensuring the user interaction experience, which is conducive to forming a virtuous cycle of "the enrichment of the available room pool improves the user experience, the improvement of the user experience leads to the growth of the user scale, and the growth of the user scale makes the available room pool more enriched". On the other hand, it can be applied to multi-person equal interaction scenarios of two or more people. The number of people that a virtual room can accommodate is not limited to two people, and the user who creates the virtual room can be any user in the virtual room, which has a wide range of applicability.
[0136] Exemplary Media
[0137] After introducing the method according to the exemplary embodiment of the present application, next, the medium according to the exemplary embodiment of the present application will be described.
[0138] In some possible implementations, various aspects of the present application may also be implemented as a medium on which a program product capable of implementing the virtual room processing method described above in this specification is stored. In some possible implementations, various aspects of the present application may also be implemented in the form of a program product on which a program code is included. When the program product is run on an electronic device, the program code is used to enable the electronic device to execute the steps according to various exemplary embodiments of the present disclosure described in the above “Exemplary Method” section of this specification.
[0139] Specifically, when the processor of the electronic device executes the program code, it can be used to implement the following steps:
[0140] In response to a virtual room matching request from a requesting user, an environment prediction model is used to perform environment prediction based on first environment data corresponding to a plurality of historical sampling moments, to obtain second environment data corresponding to a future sampling moment, wherein any environment data includes the number of matching rooms and the number of users to be matched;
[0141] Based on the second environment data, using the room creation decision model, determine a room creation decision for the requesting user;
[0142] Based on the room creation decision, a virtual room is matched to the requesting user.
[0143] In an optional embodiment, the above-mentioned use of a room creation decision model based on the second environment data to determine the room creation decision for the requesting user can be achieved through the following steps: inputting the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data into the room creation decision model, so that the room creation decision model predicts the probability of creating a new room and the probability of joining an already created room through feature extraction and feature intersection, and outputs the room creation decision by comparing the probability of creating a new room and the probability of joining an already created room.
[0144] In an optional embodiment, the above-mentioned prediction of the probability of creating a new room and the probability of adding an already created room through feature extraction and feature cross-linking can be achieved through the following steps: feature extraction is performed on the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data, respectively, to obtain user-side feature data, room-side feature data, and environment-side feature data; feature cross-linking is performed on the user-side feature data, the room-side feature data, and the environment-side feature data to obtain feature association data; based on the feature association data, the probability of creating a new room and the probability of adding an already created room is predicted.
[0145] In an optional embodiment, the above-mentioned feature cross-linking of the user-side feature data, the room-side feature data and the environment-side feature data to obtain feature association data can be achieved through the following steps: feature cross-linking the user behavior features in the user-side feature data with the room identification features in the room-side feature data to obtain first feature association data; feature cross-linking the matching user features in the room-side feature data with the user identification features in the user-side feature data to obtain second feature association data; feature cross-linking the to-be-matched user features in the environment-side feature data with the user identification features in the user-side feature data to obtain third feature association data; feature cross-linking the matchable room features in the environment-side feature data with the room identification features in the room-side feature data to obtain fourth feature association data; based on the first feature association data, the second feature association data, the third feature association data and the fourth feature association data, determine the feature association data between the user side, the room side and the environment side.
[0146] In an optional embodiment, the processor of the electronic device can also be used to implement the following steps when executing the program code: generating an environmental time series based on sample environmental data corresponding to multiple sample sampling moments; performing a stationarity test on the environmental time series, and if the environmental time series fails the stationarity test, performing a differential operation on the environmental time series until the environmental time series passes the stationarity test; using the initialized environmental prediction model to fit the environmental time series until the environmental prediction model meets the preset fitting stop condition, thereby obtaining the environmental prediction model finally used.
[0147] In an optional implementation, the preset fitting stop condition includes that the environment prediction model passes the white noise test.
[0148] In an optional embodiment, the processor of the electronic device can also be used to implement the following steps when executing the program code: based on the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user, the initialized room creation decision model is trained until the room creation decision model meets the preset training stop condition to obtain the room creation decision model finally used.
[0149] In an optional implementation, the above-mentioned training of the initialized room creation decision model based on the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user can be achieved through the following steps: inputting the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user into the room creation decision model to obtain the decision result output by the room decision model; based on the decision result and the positive and negative sample types corresponding to the sample user, optimizing the model parameters of the room creation decision model.
[0150] In an optional embodiment, the above-mentioned sample users are users who have created virtual rooms. The positive and negative sample types corresponding to the sample users who stay in the created virtual room for more than a preset time are positive samples, and the positive and negative sample types corresponding to the sample users who stay in the created virtual room for no more than the preset time are negative samples.
[0151] In an optional implementation, the matching environment data corresponding to the sample user includes environment data estimated by an environment estimation model when the sample user initiates a virtual room matching request.
[0152] In an optional implementation, the preset training stop condition includes that the loss function of the room creation decision model is less than a preset threshold.
[0153] In the above virtual room processing, on the one hand, the problem of virtual room creation is solved by combining environmental data, which can enrich the available room pool while ensuring the user interaction experience, which is conducive to forming a virtuous cycle of "the enrichment of the available room pool improves the user experience, the improvement of the user experience leads to the growth of the user scale, and the growth of the user scale makes the available room pool richer". On the other hand, it can be applied to multi-person equal interaction scenarios of two or more people. The number of people that a virtual room can accommodate is not limited to two people, and the user who creates the virtual room can be any user in the virtual room, which has a wide range of applicability.
[0154] It should be noted that the above-mentioned medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable 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 thereof.
[0155] A readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, wherein readable program code is carried. 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. A readable signal medium may also be any readable medium other than a 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.
[0156] The program code contained in the readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, optical cable, RF (Radio Frequency), etc., or any suitable combination of the above.
[0157] Program code for performing the operations of the present application may be written in any combination of one or more programming languages, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user computing device, partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).
[0158] Exemplary Devices
[0159] After introducing the medium of the exemplary embodiment of the present application, next, reference is made to Figure 6 A virtual room processing device according to an exemplary embodiment of the present application is described.
[0160] See also Figure 6 , Figure 6FIG. 1 is a block diagram of a virtual room processing device according to an exemplary embodiment of the present application. Figure 6 As shown, a virtual room processing device 600 according to an exemplary embodiment of the present application includes: an environment prediction module 610, a decision determination module 620, and a room matching module 630, wherein:
[0161] The environment prediction module 610 is used to respond to the virtual room matching request of the requesting user, and to use the environment prediction model to perform environment prediction based on the first environment data corresponding to multiple historical sampling moments, and obtain the second environment data corresponding to the future sampling moment, wherein any environment data includes the number of matching rooms and the number of users to be matched;
[0162] A decision determination module 620, configured to determine a room creation decision for a requesting user based on the second environment data and using a room creation decision model;
[0163] The room matching module 630 is used to match a virtual room for the requesting user based on the room creation decision.
[0164] In an optional embodiment, based on the aforementioned scheme, the decision determination module 620 includes: a room creation decision model processing module, which is used to input user behavior data corresponding to the requesting user, historical room data corresponding to the requesting user, and second environment data into the room creation decision model, so that the room creation decision model predicts the probability of creating a new room and the probability of joining an already created room through feature extraction and feature intersection, and outputs a room creation decision by comparing the probability of creating a new room and the probability of joining an already created room.
[0165] In an optional embodiment, based on the aforementioned scheme, the room creation decision model processing module includes: a feature extraction module, which is used to extract features of the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data, respectively, to obtain user-side feature data, room-side feature data, and environment-side feature data; a feature cross-module, which is used to perform feature cross-section on the user-side feature data, the room-side feature data, and the environment-side feature data to obtain feature association data; a probability prediction module, which is used to predict the probability of creating a new room and the probability of joining an already created room based on the feature association data.
[0166] In an optional embodiment, based on the aforementioned scheme, the feature intersection module is configured to: perform feature intersection on the user behavior features in the user-side feature data and the room identification features in the room-side feature data to obtain first feature association data; perform feature intersection on the matching user features in the room-side feature data and the user identification features in the user-side feature data to obtain second feature association data; perform feature intersection on the to-be-matched user features in the environment-side feature data and the user identification features in the user-side feature data to obtain third feature association data; perform feature intersection on the matchable room features in the environment-side feature data and the room identification features in the room-side feature data to obtain fourth feature association data; determine the feature association data between the user side, the room side, and the environment side based on the first feature association data, the second feature association data, the third feature association data, and the fourth feature association data.
[0167] In an optional embodiment, based on the aforementioned scheme, the virtual room processing device 600 also includes: a sequence generation module, which is used to generate an environmental time series based on sample environmental data corresponding to multiple sample sampling moments; a stationarity test module, which is used to perform a stationarity test on the environmental time series. If the environmental time series fails the stationarity test, a differential operation is performed on the environmental time series until the environmental time series passes the stationarity test; an environmental prediction model generation module, which is used to fit the environmental time series with an initialized environmental prediction model until the environmental prediction model meets a preset fitting stop condition to obtain the environmental prediction model finally used.
[0168] In an optional implementation, based on the above solution, in the environment prediction model generation module, the preset fitting stop condition includes that the environment prediction model passes the white noise test.
[0169] In an optional embodiment, based on the aforementioned scheme, the virtual room processing device 600 also includes: a room creation decision model generation module, which is used to train the initialized room creation decision model based on user behavior data corresponding to the sample user, historical room data corresponding to the sample user, and matching environment data corresponding to the sample user, until the room creation decision model meets the preset training stop condition, thereby obtaining the room creation decision model finally used.
[0170] In an optional embodiment, based on the aforementioned scheme, the room creation decision model generation module includes: a decision result determination module, which is used to input user behavior data corresponding to the sample user, historical room data corresponding to the sample user, and matching environment data corresponding to the sample user into the room creation decision model to obtain the decision result output by the room decision model; a model parameter optimization module, which is used to optimize the model parameters of the room creation decision model based on the decision result and the positive and negative sample types corresponding to the sample user.
[0171] In an optional embodiment, based on the aforementioned scheme, in the virtual room processing device 600, the sample users are users who have created virtual rooms, and the positive and negative sample types corresponding to the sample users who have stayed in the created virtual room for more than a preset time are positive samples, and the positive and negative sample types corresponding to the sample users who have stayed in the created virtual room for no more than a preset time are negative samples.
[0172] In an optional implementation, based on the above solution, in the room creation decision model generation module, the matching environment data corresponding to the sample user includes the environment data estimated by the environment estimation model when the sample user initiates a virtual room matching request.
[0173] In an optional implementation, based on the aforementioned solution, in the room creation decision model generation module, the preset training stop condition includes that the loss function of the room creation decision model is less than a preset threshold.
[0174] In the above virtual room processing, on the one hand, the problem of virtual room creation is solved by combining environmental data, which can enrich the available room pool while ensuring the user interaction experience, which is conducive to forming a virtuous cycle of "the enrichment of the available room pool improves the user experience, the improvement of the user experience leads to the growth of the user scale, and the growth of the user scale makes the available room pool richer". On the other hand, it can be applied to multi-person equal interaction scenarios of two or more people. The number of people that a virtual room can accommodate is not limited to two people, and the user who creates the virtual room can be any user in the virtual room, which has a wide range of applicability.
[0175] It should be noted that although several modules or units of the virtual room processing device are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiment of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units for embodiment.
[0176] Exemplary Electronic Devices
[0177] After introducing the virtual room processing method, medium and device according to the exemplary embodiment of the present application, next, an electronic device according to another exemplary embodiment of the present application is introduced.
[0178] Those skilled in the art will appreciate that various aspects of the present application may be implemented as a system, method or program product. Therefore, various aspects of the present application may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to as "circuit", "module" or "system" herein.
[0179] Refer to the following Figure 7 An electronic device 700 according to such an embodiment of the present invention is described. Figure 7 The electronic device 700 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.
[0180] like Figure 7 As shown, the electronic device 700 is in the form of a general computing device. The components of the electronic device 700 may include, but are not limited to: the at least one processing unit 710, the at least one storage unit 720, a bus 730 connecting different system components (including the storage unit 720 and the processing unit 710), and a display unit 740.
[0181] The storage unit stores program codes, which can be executed by the processing unit 710, so that the processing unit 710 executes the steps according to various exemplary embodiments of the present invention described in the above “Exemplary Method” section of this specification.
[0182] The storage unit 720 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 721 and / or a cache memory unit 722 , and may further include a read-only memory unit (ROM) 723 .
[0183] The storage unit 720 may also include a program / utility 724 having a set (at least one) of program modules 725, such program modules 725 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 the reality of a network environment.
[0184] The bus 730 may include a data bus, an address bus, and a control bus.
[0185] The electronic device 700 may also communicate with one or more external devices 770 (e.g., a keyboard, a pointing device, a Bluetooth device, etc.), and such communication may be performed through an input / output (I / O) interface 750. Furthermore, the electronic device 700 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 760. Figure 7As shown, the network adapter 760 communicates with other modules of the electronic device 700 via the bus 730. It should be understood that although Figure 7 Not shown, other hardware and / or software modules may be used in conjunction with the electronic device 700, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID (Redundant Arrays of Independent Disks) systems, tape drives, and data backup storage systems.
[0186] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the embodiment of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiment of the present disclosure.
[0187] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.
[0188] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims
1. A virtual room processing method, characterized in that: The method comprises: In response to a virtual room matching request from a requesting user, an environment prediction model is used to perform environment prediction based on first environment data corresponding to a plurality of historical sampling moments, to obtain second environment data corresponding to a future sampling moment, wherein any environment data includes the number of matching rooms and the number of users to be matched; Based on the second environment data, using a room creation decision model, determining a room creation decision for the requesting user; Based on the room creation decision, matching a virtual room for the requesting user; The step of using a room creation decision model based on the second environment data to determine a room creation decision for the requesting user includes: The user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user and the second environment data are input into the room creation decision model, so that the room creation decision model predicts the probability of creating a new room and the probability of joining an already created room through feature extraction and feature intersection, and outputs the room creation decision by comparing the probability of creating a new room and the probability of joining an already created room.
2. The virtual room processing method according to claim 1, characterized in that: The prediction of the probability of creating a new room and the probability of adding a created room by feature extraction and feature intersection includes: Perform feature extraction on the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data, respectively, to obtain user-side feature data, room-side feature data, and environment-side feature data; Perform feature cross-talk on the user-side feature data, the room-side feature data, and the environment-side feature data to obtain feature association data; Based on the feature association data, the probability of creating a new room and the probability of joining an already created room are predicted.
3. The virtual room processing method according to claim 2, characterized in that: The performing feature cross-talk on the user-side feature data, the room-side feature data, and the environment-side feature data to obtain feature association data includes: Perform feature cross-talk on the user behavior feature in the user-side feature data and the room identification feature in the room-side feature data to obtain first feature association data; Perform feature cross-talk on the matching user features in the room-side feature data and the user identification features in the user-side feature data to obtain second feature association data; Perform feature cross-talk on the to-be-matched user features in the environment-side feature data and the user identification features in the user-side feature data to obtain third feature association data; Perform feature intersection on the matchable room features in the environment-side feature data and the room identification features in the room-side feature data to obtain fourth feature association data; Based on the first feature association data, the second feature association data, the third feature association data, and the third feature association data, feature association data between the user side, the room side, and the environment side are determined.
4. The virtual room processing method according to claim 1, characterized in that: The method further comprises: Generate an environmental time series based on sample environmental data corresponding to multiple sample sampling moments; Performing a stationary test on the environmental time series, and if the environmental time series fails the stationary test, performing a differential operation on the environmental time series until the environmental time series passes the stationary test; The initialized environment prediction model is used to fit the environment time series until the environment prediction model meets a preset fitting stop condition, thereby obtaining the environment prediction model finally used.
5. The virtual room processing method according to claim 4, characterized in that: The preset fitting stop condition includes that the environment prediction model passes the white noise test.
6. The virtual room processing method according to claim 1, characterized in that: The method further comprises: Based on the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user, the initialized room creation decision model is trained until the room creation decision model meets the preset training stop condition to obtain the room creation decision model finally used.
7. The virtual room processing method according to claim 6, characterized in that: The training of the initialized room creation decision model based on the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user includes: Inputting user behavior data corresponding to the sample user, historical room data corresponding to the sample user, and matching environment data corresponding to the sample user into the room creation decision model to obtain a decision result output by the room creation decision model; Based on the decision result and the positive and negative sample types corresponding to the sample users, the model parameters of the room creation decision model are optimized.
8. The virtual room processing method according to claim 7, characterized in that: The sample users are users who have created virtual rooms. The positive and negative sample types corresponding to the sample users who stay in the created virtual room for more than the preset time are positive samples, and the positive and negative sample types corresponding to the sample users who stay in the created virtual room for no more than the preset time are negative samples.
9. The virtual room processing method according to claim 6, characterized in that: The matching environment data corresponding to the sample user includes the environment data estimated by the environment estimation model when the sample user initiates a virtual room matching request.
10. The virtual room processing method according to claim 6, characterized in that: The preset training stop condition includes that the loss function of the room creation decision model is less than a preset threshold.
11. A virtual room processing device, characterized in that: The device comprises: An environment prediction module is used to respond to a virtual room matching request from a requesting user, and to use an environment prediction model to perform environment prediction based on first environment data corresponding to multiple historical sampling moments, and obtain second environment data corresponding to future sampling moments, wherein any environment data includes the number of matching rooms and the number of users to be matched; A decision determination module, configured to determine a room creation decision for the requesting user based on the second environment data and using a room creation decision model; A room matching module, configured to match a virtual room for the requesting user based on the room creation decision; Wherein, the decision determination module includes: The room creation decision model processing module is used to input the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user and the second environment data into the room creation decision model, so that the room creation decision model predicts the probability of creating a new room and the probability of joining an already created room through feature extraction and feature intersection, and outputs the room creation decision by comparing the probability of creating a new room and the probability of joining an already created room.
12. The virtual room processing device according to claim 11, characterized in that: The room creation decision model processing module includes: A feature extraction module, used to extract features from the user behavior data corresponding to the requesting user, the historical room data corresponding to the requesting user, and the second environment data, respectively, to obtain user-side feature data, room-side feature data, and environment-side feature data; A feature crossover module, used for performing feature crossover on the user-side feature data, the room-side feature data and the environment-side feature data to obtain feature association data; The probability prediction module is used to predict the probability of creating a new room and the probability of adding a created room based on the feature association data.
13. The virtual room processing device according to claim 12, characterized in that: The feature cross-connect module is configured as follows: Perform feature cross-talk on the user behavior feature in the user-side feature data and the room identification feature in the room-side feature data to obtain first feature association data; Perform feature cross-talk on the matching user features in the room-side feature data and the user identification features in the user-side feature data to obtain second feature association data; Perform feature cross-talk on the to-be-matched user features in the environment-side feature data and the user identification features in the user-side feature data to obtain third feature association data; Perform feature intersection on the matchable room features in the environment-side feature data and the room identification features in the room-side feature data to obtain fourth feature association data; Based on the first feature association data, the second feature association data, the third feature association data, and the third feature association data, feature association data between the user side, the room side, and the environment side are determined.
14. The virtual room processing device according to claim 11, characterized in that: The device comprises: A sequence generation module, used to generate an environmental time series based on sample environmental data corresponding to a plurality of sample sampling moments; A stationarity test module is used to perform a stationarity test on the environmental time series. If the environmental time series fails the stationarity test, a differential operation is performed on the environmental time series until the environmental time series passes the stationarity test. The environment prediction model generation module is used to use the initialized environment prediction model to fit the environment time series until the environment prediction model meets the preset fitting stop condition to obtain the environment prediction model finally used.
15. The virtual room processing device according to claim 14, characterized in that: In the environment prediction model generation module, the preset fitting stop condition includes that the environment prediction model passes the white noise test.
16. The virtual room processing device according to claim 11, characterized in that: The device also includes: The room creation decision model generation module is used to train the initialized room creation decision model based on the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user, until the room creation decision model meets the preset training stop condition, thereby obtaining the room creation decision model finally used.
17. The virtual room processing device according to claim 16, characterized in that: The room creation decision model generation module includes: A decision result determination module, used to input the user behavior data corresponding to the sample user, the historical room data corresponding to the sample user, and the matching environment data corresponding to the sample user into the room creation decision model to obtain the decision result output by the room creation decision model; A model parameter optimization module is used to optimize the model parameters of the room creation decision model based on the decision result and the positive and negative sample types corresponding to the sample users.
18. The virtual room processing device according to claim 17, characterized in that: In the virtual room processing device, the sample users are users who have created virtual rooms. The positive and negative sample types corresponding to the sample users who have stayed in the created virtual room for more than a preset time are positive samples, and the positive and negative sample types corresponding to the sample users who have stayed in the created virtual room for no more than a preset time are negative samples.
19. The virtual room processing device according to claim 16, characterized in that: The matching environment data corresponding to the sample user includes the environment data estimated by the environment estimation model when the sample user initiates a virtual room matching request.
20. The virtual room processing device according to claim 16, characterized in that: The preset training stop condition includes that the loss function of the room creation decision model is less than a preset threshold.
21. An electronic device, characterized in that: include: processor; A memory, configured to store executable instructions of the processor; The processor is configured to execute the virtual room processing method according to any one of claims 1 to 10 by executing the executable instructions.
22. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the virtual room processing method according to any one of claims 1 to 10 is implemented.
Citation Information
Patent Citations
Virtual room distribution method and device, storage medium and processor
CN108876081A
User classification method and device, computer readable storage medium and electronic equipment
CN111738331A
Prediction type intelligent automobile decision control method and device, vehicle and storage medium
CN114084155A
Distributed photovoltaic power generation electric quantity prediction method and system
CN115036922A
