Message recommendation method and apparatus, electronic device, and medium

Abstract

The application discloses a message recommendation method and device, electronic equipment and medium, after a user logs in an application program by using login identification information, an application program server acquires the login identification information and device identification of the current login of the user, then acquires application characteristic data corresponding to the terminal device, determines target identification information corresponding to the login identification information according to the login identification information, the device identification and the application characteristic data, acquires a recommended message corresponding to the target identification information according to the target identification information, and sends the recommended message to the terminal device, so that the application program can perform message recommendation according to the real identity of the actually logged-in user, effectively improves the accuracy of message recommendation, and further improves the recommendation conversion rate.

Description

Technical Field

[0001] This disclosure generally relates to data recommendation, specifically to advertising data recommendation, and more particularly to a message recommendation method, apparatus, electronic device, and medium. Background Technology

[0002] Sending recommendations to users when they use application features is a common technique used by modern apps. Typically, recommendations are determined directly based on the currently logged-in user's account information. However, in apps with high VIP privileges, multiple users often share a single account. This makes recommendations based on account information inaccurate and unable to accurately recommend information that might be of interest to the actual user. Summary of the Invention

[0003] In view of the above-mentioned defects or deficiencies in the prior art, it is desirable to provide a message recommendation method, apparatus, electronic device and medium that can accurately identify the real user identity, thereby making targeted message recommendations, improving the accuracy of message recommendations, and thus improving the conversion rate of message recommendations.

[0004] In a first aspect, embodiments of this application provide a message recommendation method, including:

[0005] Obtain the login identifier information and the device identifier of the currently logged-in terminal device;

[0006] Obtain application feature data corresponding to the terminal device, wherein the application feature data is operation data received on the terminal device;

[0007] Based on the login identifier information, the device identifier, and the application feature data, determine the target identifier information corresponding to the login identifier information;

[0008] The recommended message corresponding to the target identifier information is obtained based on the target identifier information, and the recommended message is sent to the terminal device.

[0009] Secondly, embodiments of this application provide a message recommendation device, including:

[0010] The first acquisition module is used to acquire the login identifier information of the currently logged-in user and the device identifier of the terminal device.

[0011] The second acquisition module is used to acquire application feature data corresponding to the terminal device, wherein the application feature data is operation data received by the terminal device.

[0012] The determination module is used to determine the target identification information corresponding to the login identification information based on the login identification information, the device identification, and the application feature data;

[0013] The recommendation module is used to obtain the recommendation message corresponding to the target identification information based on the target identification information, and send the recommendation message to the terminal device.

[0014] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in embodiments of this application.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in embodiments of this application.

[0016] The message recommendation method proposed in this application involves the application server obtaining the user's current login identifier and device identifier after the user logs in to the application. Then, it obtains the application feature data corresponding to the terminal device. Based on the login identifier, device identifier, and application feature data, it determines the target identifier corresponding to the login identifier and obtains the recommendation message corresponding to the target identifier. The recommendation message is then sent to the terminal device, enabling the application to recommend messages based on the actual identity of the logged-in user, effectively improving the accuracy of message recommendations and thus increasing the recommendation conversion rate.

[0017] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0018] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0019] Figure 1 This is a schematic diagram of the implementation environment involved in this application;

[0020] Figure 2 This is a flowchart of a message recommendation method according to an embodiment of this application;

[0021] Figure 3 This is a schematic diagram illustrating an application scenario of a message recommendation method according to a specific embodiment of this application;

[0022] Figure 4 This is a flowchart illustrating a message recommendation method according to a specific embodiment of this application;

[0023] Figure 5 This is a schematic diagram illustrating the principle of a message recommendation method according to an embodiment of this application;

[0024] Figure 6 This is a schematic diagram of the structure of a message recommendation device according to an embodiment of this application;

[0025] Figure 7 This is a schematic diagram of the structure of a computer system suitable for implementing electronic devices or servers according to the embodiments of this application. Detailed Implementation

[0026] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0027] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0028] On the pages of an application app, there are usually functional controls that recommend content to users. These controls typically display recommended information that corresponds to the user's account profile. The recommended information is usually relevant information that the currently logged-in user may be interested in, including but not limited to videos, music, novels, shopping recommendations, etc.

[0029] However, as VIP benefits within applications have increased, many users have shared a single account, such as multiple roommates in a university dormitory or a couple sharing an account. In this situation, the preference information received by the application for user profiling becomes confused due to the large number of users who have used the same account. This leads to contradictory or inaccurate user profile tags, making recommendations based on these tags inconsistent with the actual user's interests and failing to achieve the intended purpose of information recommendation.

[0030] Please see Figure 1 , Figure 1 This is a schematic diagram of the implementation environment involved in this application. The implementation environment is a message recommendation system, which includes terminal 1 and server 2.

[0031] In this system, terminal 1 and server 2 are pre-connected via a network for data transmission, such as the transmission required during message recommendation. This network can be a wired or wireless network; this application does not impose any restrictions. Message types include, but are not limited to, videos, music, novels, and shopping information.

[0032] Terminal 1 has a message recommendation client (Application App) installed. The message recommendation client can display several recommended messages on its homepage. For example, when the Application App is a video client, it can display video tags or a list of recommended videos on the homepage; when it is a music client, it can display music tags or a list of recommended music; and when it is a novel reader, it can display tags or a list of recommended novels. The tags include cover information and / or summary information of the recommended information, and the list includes, but is not limited to, the name of the recommended information. When the user clicks on a tag of a recommended information on the homepage, the Application App plays and / or displays the specific content of the recommended information. When the user clicks on any recommended message in the list, the Application App displays, but is not limited to, the cover information and / or summary information of the recommended message, or directly plays and / or displays the specific content of the recommended information.

[0033] Optionally, the application (App) can also call a message recommendation control to recommend messages, and the message control can be displayed in front of the application's homepage.

[0034] Optionally, the type of recommended messages can be different from that of the application app. For example, when the application app is a video client, recommended messages may include, but are not limited to, shopping information, novels, etc.

[0035] Terminal 1 can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, in-vehicle terminal, smart TV, etc., but is not limited to these.

[0036] Terminals and servers can be connected directly or indirectly via wired or wireless communication, and this application does not impose any restrictions on this.

[0037] Server 2 is used to store recommendation messages, receive user operation records, browsing records, and on-demand records on the application App, and create user profiles based on these records. It then sends recommendation messages corresponding to the user profiles to the application App on terminal 1.

[0038] Server 2 can be a single server, a server cluster consisting of several servers, or it can contain one or more virtualization platforms, or it can be a cloud computing service center.

[0039] Figure 2 This is a flowchart illustrating a message recommendation method according to an embodiment of this application.

[0040] It should be noted that the execution subject of the message recommendation method in this embodiment is a message recommendation device. The message recommendation device can be implemented by software and / or hardware. In this embodiment, the message recommendation device can be configured in an electronic device or in a server used to control the electronic device. The server communicates with the electronic device and then controls it.

[0041] like Figure 2 As shown, message recommendation methods include:

[0042] Step 101: Obtain the login identifier information and the device identifier of the terminal device currently logged in.

[0043] When a user logs into the application's client on their terminal, the application server obtains the user's current login identifier information. This login identifier information is the user's registered and / or paid account within the application; that is, the user's username within the application. This username can be a combination of characters, letters, arrays, and symbols that the user sets according to their personal preferences.

[0044] Device identifiers generally refer to unique and independent identification information that can be used to identify a device, such as IMEI / IDFA / TAID / OAID, etc. Among them, IMEI is the unique device identifier for Android system devices, IDFA is the unique device identifier for iOS system devices, TAID is the Turing Shield identifier, which can be used across Android and iOS systems, and OAID is a device number developed by the Android system manufacturer alliance and is only used on the Android platform.

[0045] Step 102: Obtain the application feature data corresponding to the terminal device. The application feature data is the operation data received by the terminal device.

[0046] Among them, application feature data refers to the user's operation data on the terminal device. For example, the operation data generated on the user's terminal before and after logging into the application, including but not limited to operation information such as the selection of each tab in the current application, or the selection and / or closing of other applications.

[0047] Step 103: Determine the target identifier information corresponding to the login identifier information based on the login identifier information, device identifier, and application feature data.

[0048] Among them, the target identification information is the account information used to represent the identity of the currently logged-in real user. Specifically, it can be the account that the user registered in a specific application, which is usually the user's account in a personal information interaction platform.

[0049] Step 104: Obtain the recommendation message corresponding to the target identifier information based on the target identifier information, and send the recommendation message to the terminal device.

[0050] Therefore, after a user logs in to the application using their login identifier, the application server obtains the user's current login identifier and device identifier, then obtains the application feature data corresponding to the terminal device. Based on the login identifier, device identifier, and application feature data, it determines the target identifier corresponding to the login identifier, obtains the recommendation message corresponding to the target identifier, and sends the recommendation message to the terminal device. This enables the application to recommend messages based on the actual identity of the logged-in user, effectively improving the accuracy of message recommendations and thus increasing the recommendation conversion rate.

[0051] In some embodiments, determining the target identifier information corresponding to the login identifier information based on the login identifier information, the device identifier, and the application feature data includes: determining at least one candidate identifier information based on the login identifier information and the device identifier; and determining the target identifier information based on the candidate identifier information and the application feature data.

[0052] The candidate identifier information refers to identifier information that is associated with the login identifier information, including but not limited to account information from platforms such as content interaction platforms and shopping platforms. Further, the associated identifier information may include account information logged in on the terminal device corresponding to the device identifier, as well as account information logged in on other terminal devices that have logged in with the login identifier information. It should be understood that the terminal device routinely records the account information used by the user when opening and / or logging into each application. Then, when the user logs into an application containing this application, the recorded account information and device identifier are sent to the server so that the server can record and store the account information associated with the login identifier information and the device identifier corresponding to each account.

[0053] The target identifier is an account used to represent the user's real identity, determined from the candidate identifiers according to preset rules. For example, the candidate identifiers can be scored to obtain the matching degree between each candidate identifier and the user's real identity. Then, the candidate identifier with the highest matching degree is used as the target identifier. When multiple candidate identifiers have the same matching degree, the target identifier can also be determined based on auxiliary conditions such as the application type corresponding to the candidate identifier.

[0054] It's important to note that this application targets a scenario where multiple users share an account. Therefore, simply obtaining the login identifier information from an application cannot determine which specific user among the multiple users is currently using that identifier. When multiple users share a single login identifier, they typically operate on their own private devices. Therefore, using the login identifier information and device identifier can identify, for example, account information from other applications frequently logged into on that device. These accounts are more private and better reflect the user's true identity.

[0055] It should be noted that the mapping relationship between login identifier information, device identifier, and candidate identifier information can be established based on historical information collected on the terminal device. In other words, after a user logs into the application (App), the application server uses the terminal device identifier to obtain the user's account information from other applications on that device as candidate identifier information. That is, it obtains the mapping relationship between the device identifier and candidate identifier information, thereby obtaining at least one candidate identifier.

[0056] For example, when the application is a content interaction platform, users typically log in using only their personal terminals. Therefore, the user's login identifier and device identifier usually have a stable mapping relationship. In this case, there is a mapping relationship between the login identifier, device identifier, and target identifier. Based on the mapping relationship between the login identifier, device identifier, and real identity, a unique candidate identifier can be obtained. Based on this unique candidate identifier, it can be determined that the candidate identifier is the target identifier. When, for example, a couple shares a user account, they may use a shared terminal device, such as a home computer. If the two are relatively close and can share the content interaction platform account, there is a mapping relationship between the login identifier, device identifier, and other login identifiers and the couple's target identifier. That is, based on the mapping relationship between the login identifier, device identifier, and other accounts and real identity, two candidate identifiers can be mapped. It is necessary to further determine the target identifier based on the two candidate identifiers.

[0057] Therefore, this application can identify candidate identifier information through device identifiers that are not easily shared, effectively reducing the scope of obtaining user target identifier information, further improving the speed and accuracy of determining real login identifier information, and thus effectively improving the accuracy of message recommendation.

[0058] Furthermore, based on the candidate identifier information and the application feature data, the target identifier information is determined, including: inputting the candidate identifier information and feature data into the feature learning model to obtain the target identifier information.

[0059] The feature learning model is trained using historical information from the terminal device.

[0060] It should be noted that historical information includes, but is not limited to, historical information generated by the user using the current application, and may also include historical information generated by the user using other applications. Historical information includes login identification information, application characteristic data, and the mapping relationship between login identification information and application characteristic data.

[0061] In other words, a large amount of login identification information from various applications (Apps), user history operation information, terminal device identification information, and the relationships between these information can be obtained. This large amount of data is then used to create training and / or validation sets based on the type of feature learning model used. For example, when the feature learning model is supervised, the collected data can be used to create both training and validation sets; when the feature learning model is unsupervised, only the collected data can be used to create a training set. The acquired data is then used to train the feature learning model until it can effectively determine the target identifier information based on the input candidate identifier information and applied feature data. At this point, the feature learning model training is complete. When it is necessary to determine the target identifier information, the determined candidate identifier information and applied feature data are input into the feature learning model to obtain the user's target identifier information.

[0062] In one or more embodiments, before the acquired large amount of data is made into a training set and / or a validation set according to the type of feature learning model used, the data can be cleaned. For example, if a device identifier corresponds to multiple login identifiers for a certain application, and the application feature data corresponding to the multiple login identifiers represents different users, then the multiple user identity identifiers and device identifiers corresponding to the device identifier are deleted, thereby effectively improving the accuracy of the training set, reducing the training difficulty, and increasing the model training speed.

[0063] In one or more embodiments, the feature learning model may focus on the source application of the candidate identifier information and / or application feature data, as well as the user's usage duration and / or frequency. That is, the feature learning model can determine a first weight value for the candidate identifier information and / or application feature data based on the type of the source application. For example, when the application is a content interaction platform, the first weight value of the candidate identifier information and / or application feature data is higher than that of applications such as video playback applications and novel reading applications. Furthermore, the user's usage duration and / or frequency of the application corresponds to a second weight value; that is, the higher the user's usage duration and / or frequency, the greater the corresponding weight value. The feature learning model can combine the first and second weight values ​​to score the authenticity of the candidate identifier information, and then, based on the descending order of the scores of multiple candidate identifier information, select the candidate identifier information with the highest score as the user's target identifier information. The authenticity score can be a weighted sum of the application feature data corresponding to each candidate identifier information, the source application of the application feature data, and the user's usage duration and / or frequency, along with the weights corresponding to each attribute.

[0064] Optionally, the storage of acquired data and the training of the feature learning model can be performed on the application's backend server or in other services, such as a vendor server. For example, after the vendor server completes the training of the feature learning model, it updates the parameters of the feature learning model in the application server. Alternatively, the application server can communicate with the vendor server, sending the acquired candidate identifier information and feature data to the vendor server, and then the vendor server returns the target identifier information predicted by the feature learning model to the application server.

[0065] Therefore, this application first uses the mapping relationship to determine the candidate identifier information of the user, and then uses the latest application data generated on the terminal device to determine the target identifier information of the user, which further improves the accuracy of judging the user's real identity and effectively improves the accuracy of message recommendation.

[0066] In some embodiments, obtaining the recommendation message corresponding to the target identifier information based on the target identifier information includes: obtaining the profile tag corresponding to the target identifier information based on the target identifier information, and obtaining the recommendation message corresponding to the target identifier information based on the target identifier information and the profile tag.

[0067] User profile tags describe the user's characteristic type. The number of user profile tags can be determined based on the actual situation; that is, different users may have different numbers of user profile tags. For example, the obtained user profile tags may include, but are not limited to, basic information such as age, gender, and online activity; user information such as education level, marital status, lifestyle, business interests, and keywords; user behavior such as paying users, application behavior, and app users; weather-related information such as clothing index, UV index, makeup index, weather conditions, and temperature; and spending power such as residential community prices.

[0068] Optionally, a data service platform can be established on the application server or vendor server to perform user feature modeling, profiling, and user classification based on user data. Specifically, the data service platform can use an underlying deep learning model to obtain user features corresponding to the target identifier information based on the user's target identifier information, and then use the mapping relationship between user features and profile tags to obtain the corresponding profile tags for the user.

[0069] After obtaining the profile tags, the recommendation message engine can be used to further obtain the recommendation message corresponding to the target identifier information based on the target identifier information and the profile tags.

[0070] The recommended messages can be mapped to user profile tags. For example, candidate recommended messages corresponding to user profile tags can be obtained, and then the recommended message to be used for recommendation can be selected from the candidate recommended messages based on the target identifier information. Furthermore, the target identifier information may include user preference information, such as tags that the user has previously selected as uninteresting. After determining the candidate recommended messages based on the profile tags, recommended messages that match tags that the user is not interested in are excluded based on the target identifier information, thereby obtaining the final recommended message recommended by the user.

[0071] It should be understood that when the recommended messages are provided by the same type of application, the message recommendation engine can be set up in the application's backend server. When the recommended messages are provided by a third-party application, the message recommendation engine can be set up in the third-party's server. Specifically, when the application is a video player, if the recommended message is video information, the message recommendation engine can be set up in the application's backend server. If the recommended message is shopping information, the message recommendation engine can be set up in the shopping platform's backend server. After obtaining the recommended message based on the user profile tags and target identifier information, the shopping platform's backend server sends the recommended message to the video player's backend server, so that the video player's backend server can send the recommended message to the terminal device. Alternatively, the shopping platform's backend server obtains candidate recommended messages based on the user profile tags and sends them to the video player's backend server. The video player's backend server then selects the final recommended message from the candidate recommended messages based on the user's target identifier information and sends the recommended message to the terminal device.

[0072] In some embodiments, after sending the recommendation message to the terminal device, the method further includes: obtaining operation information on the recommendation message, using the operation information as historical information of the terminal device, and using the historical information and target identification information to train the feature learning model.

[0073] In other words, in order to ensure that the target identification information obtained each time can accurately describe the user's current preference characteristics, it is also necessary to use the user's operation information on the application each time to train the feature learning model.

[0074] Specifically, the system acquires user interaction information for recommended messages. Clicks and views are recorded as positive samples, while no clicks or views are recorded as negative samples. The feature learning model is then updated and trained using target identification information, positive samples, and negative samples. This allows the model to understand the latest trends in user preferences, enabling accurate identification of user target information during subsequent real-name authentication and effectively improving the accuracy of user target information acquisition.

[0075] As a specific example, such as Figures 3-5 As shown, taking a video playback application as an example, other applications are content interaction platforms produced by the video playback application company, and the recommendation message is a video advertisement message as an example.

[0076] The message recommendation system comprises a client, a backend server, and a real user inference system. The backend server, corresponding to the application, includes an access layer (Display), a logic service layer, a backend service data layer, and an advertising recommendation engine. The access layer interacts with the client and the real user inference system. When interacting with the client, the access layer receives the account information corresponding to the currently logged-in video playback application and the advertising recommendation request, and returns the recommended advertising information to the client. When interacting with the real user inference system, it sends device identification and application characteristic data to the system. The logic service layer interacts with the real user inference system, the backend service data layer, and the advertising recommendation engine. Specifically, the logic service layer receives target identification information from the real user inference system, sends it to the backend service data layer, receives a precise user profile from the backend service data layer, sends the precise user profile to the advertising recommendation engine, receives the recommended advertising information from the advertising recommendation engine, and finally sends the recommended advertising information to the client through the access layer. The real-user recommendation system comprises an access layer, an ID mapping system, and a user feature learning module. The access layer receives device identifiers and application feature data sent by the backend server's access layer. The ID mapping system determines candidate identifiers for each user based on the device identifiers sent by the access layer and sends this candidate identifier information to the user feature learning module. The user feature learning module determines the user's target identifier information based on the application feature data and candidate identifier information from the access layer and feeds this target identifier information back to the backend server's logical service layer through the access layer. The real-user recommendation system also receives operation logs from the client regarding the recommended advertisements, facilitating the training of the user feature learning module based on these logs.

[0077] Optionally, the real user recommendation system can also be set up in the backend server, and this application does not limit this.

[0078] The corresponding message recommendation method includes the following steps:

[0079] Step 201: The client sends its current login identifier information and initiates a request to the advertising backend access layer (Display) of the backend server.

[0080] The requests sent by the client to the backend server may include, but are not limited to, requests to authenticate login credentials and requests to retrieve recommendation messages. The login credentials are the account information corresponding to the video playback application.

[0081] Step 202: The advertising backend access layer in the backend server extracts the device identifier and application characteristic data from the request.

[0082] Step 203: The advertising backend access layer in the application server sends the login identification information, device identification, and application feature data to the access layer of the real user recommendation system.

[0083] Step 204: The access layer of the real user recommendation system sends the login identifier information and device identifier to the ID mapping system.

[0084] The Id mapping system is a mapping system between device identifiers and candidate identifier information. It can store the mapping relationship between device identifiers and candidate identities using a data list.

[0085] Step 205: The ID mapping system sends a list of candidate identifier information to the access layer of the real user recommendation system.

[0086] Candidate identifier information is determined through the mapping relationship between login identifier information and device identifier and accounts of other applications. The candidate identifier information list includes at least one candidate identifier information. The candidate identifier information includes at least the login identifier information of a specific application that has been logged into the terminal device. For example, the candidate identifier information includes account 1 of the shopping application, account 1, account 2 and account 3 corresponding to the content interaction platform, and account 1 of the novel reading platform that have been logged into the same terminal device as the currently received login identifier information.

[0087] Step 206: The access layer of the real user recommendation system sends the candidate identifier information list and application feature data to the feature learning model.

[0088] Step 207: The feature learning model performs online inference based on the candidate identifier information list and applied feature data.

[0089] Step 208: The feature learning model sends the inferred user target identification information to the access layer of the real user recommendation system.

[0090] In other words, the feature learning model scores the authenticity of the input shopping application account 1, content interaction platform accounts 1, 2 and 3, and novel reading platform account 1, and obtains the authenticity score ranking of content interaction platform account 2 > content interaction platform account 3 > content interaction platform account 1 > shopping application account 1 > novel reading platform account 1. The content interaction platform account 2 with the highest score is used as the inferred target identifier information.

[0091] Step 209: The access layer of the real user recommendation system feeds back the user's target identification information to the advertising backend access layer of the backend server.

[0092] Step 210: The advertising backend access layer of the backend server sends the user's target identification information to the logical service layer (Mixer) of the backend server.

[0093] Step 211: The logical service layer of the backend server sends the user's target identification information to the backend service data layer (Datahub) of the backend server.

[0094] Step 212: The background service data layer of the background server returns user profile tags to the logical service layer of the background server.

[0095] The backend service data layer of the backend server uses a large amount of user data collected by multiple applications to build a mapping relationship between user target identification information and user characteristics, thereby enabling the acquisition of corresponding user characteristics based on user target identification information, and then obtaining user profile tags based on the mapping relationship between user characteristics and user profile tags.

[0096] Step 213: The backend server's logical service layer sends the user's target identification information and user profile tags to the advertising recommendation engine.

[0097] The ad recommendation engine can be set up in the backend server. Advertisers who cooperate with manufacturers store ad information in the backend server and select recommended ads through the ad recommendation engine in the backend server.

[0098] Step 214: The advertising recommendation engine feeds back the advertising information to be recommended to the logical service layer of the backend server.

[0099] Step 215: The logic service layer of the backend server feeds back the advertising information to be recommended to the advertising backend access layer of the application server.

[0100] Step 216: The advertising backend access layer of the backend server sends the advertising information to be recommended to the client.

[0101] Step 217: The logic service layer of the backend server sends the ad recommendation log information to the database. This database is the one used to store data in the real user recommendation system.

[0102] Step 218: The client collects user interaction information regarding the recommended advertising information.

[0103] Step 219: The client sends the operation information log to the data training library.

[0104] Step 220: The database converts the operation information in the operation information log into positive and negative data samples.

[0105] Step 221: The database sends positive and negative data samples to the feature learning model.

[0106] Step 222: The feature learning model is trained using positive and negative data samples.

[0107] It should be understood that the message recommendation method proposed in this application can process message recommendation requests from multiple client terminals simultaneously, for example... Figure 4 The client 1, client 2 and client 3 are shown, and the target identifier information of the user can be inferred from the three login identifier information provided by client 1, client 2 and client 3 respectively, as well as the user profile tags inferred from the target identifier information respectively.

[0108] In summary, after a user logs in to the application using their login identifier, the application server obtains the user's current login identifier and device identifier. It then monitors the terminal device of the logged-in account, acquires application feature data, and determines the user's target identifier based on the login identifier, device identifier, and application feature data. Furthermore, it retrieves the corresponding recommendation message based on the target identifier and sends the recommendation message to the terminal device. This allows the application to recommend messages based on the actual identity of the logged-in user, effectively improving the accuracy of message recommendations and thus increasing the recommendation conversion rate.

[0109] Figure 6 This is a schematic diagram of the message recommendation device according to an embodiment of this application.

[0110] like Figure 6 As shown in the embodiment of this application, a message recommendation device 10 includes:

[0111] The first acquisition module 11 is used to acquire the login identifier information of the currently logged-in device and the device identifier of the terminal device;

[0112] The second acquisition module 12 is used to acquire application feature data corresponding to the terminal device, where the application feature data is the operation data received on the terminal device.

[0113] The determination module 13 is used to determine the target identification information corresponding to the login identification information based on the login identification information, device identification, and application feature data;

[0114] The recommendation module 14 is used to obtain the recommendation message corresponding to the target identifier information based on the target identifier information, and send the recommendation message to the terminal device.

[0115] In some embodiments, the determining module 13 is further configured to:

[0116] Based on the login identifier and the device identifier, at least one candidate identifier is determined;

[0117] The target identifier information is determined based on the candidate identifier information and application feature data.

[0118] In some embodiments, the determining module 13 is further configured to:

[0119] The candidate identifier information and feature data are input into the feature learning model to obtain the target identifier information;

[0120] The feature learning model was trained using historical information from the terminal device.

[0121] In some embodiments, the recommendation module 14 is further configured to:

[0122] Based on the target identification information, obtain the profile tag corresponding to the target identification information;

[0123] Based on the target identifier information and profile tags, obtain the recommendation message corresponding to the target identifier information.

[0124] In some embodiments, the recommendation module 14 is further configured to:

[0125] Get information about the actions taken with the recommended messages;

[0126] The operation information is used as the historical information of the terminal device, and the feature learning model is trained using the historical information and target identification information.

[0127] In summary, after a user logs in to the application using their login identifier, the application server obtains the user's current login identifier and device identifier. It then acquires the application feature data corresponding to the terminal device. Based on the login identifier, device identifier, and application feature data, the server determines the target identifier corresponding to the login identifier and retrieves the corresponding recommendation message. This recommendation message is then sent to the terminal device, enabling the application to recommend messages based on the actual identity of the logged-in user, effectively improving the accuracy of message recommendations and thus increasing the recommendation conversion rate.

[0128] It should be understood that the units or modules described in the message recommendation device 10 are related to the reference. Figure 4 The steps in the described method correspond to each other. Therefore, the operations and features described above for the method also apply to the message recommendation device 10 and the units contained therein, and will not be repeated here. The message recommendation device 10 can be pre-implemented in the browser or other security applications of an electronic device, or it can be loaded into the browser or other security applications of an electronic device by means of downloading. The corresponding units in the message recommendation device 10 can cooperate with the units in the electronic device to implement the solution of the embodiments of this application.

[0129] The division of modules or units mentioned in the detailed description above is not mandatory. In fact, according to the embodiments of this disclosure, 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 and embodied by multiple modules or units.

[0130] It should be noted that for details not disclosed in the message recommendation device of this application embodiment, please refer to the details disclosed in the above embodiments of this application, which will not be repeated here.

[0131] The following is for reference. Figure 7 , Figure 7 A schematic diagram of the structure of a computer system suitable for implementing the embodiments of this application is shown.

[0132] like Figure 7 As shown, the computer system includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 502 or programs loaded from storage section 508 into random access memory (RAM) 503. RAM 503 also stores various programs and data required for the system's operating instructions. CPU 501, ROM 502, and RAM 503 are interconnected via bus 504. Input / output (I / O) interface 505 is also connected to bus 504.

[0133] The following components are connected to I / O interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to I / O interface 505 as needed. A removable medium 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 510 as needed so that computer programs read from it can be installed into storage section 508 as needed.

[0134] Specifically, according to embodiments of this application, the flowchart above refers to... Figure 2The described process can be implemented as a computer software program. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. In such an embodiment, the computer program contains program code for performing the methods shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by central processing unit (CPU) 501, it performs the functions defined in the system of this application.

[0135] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0136] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operational instructions of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two connected blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified functions or operational instructions, or using a combination of dedicated hardware and computer instructions.

[0137] The units or modules described in the embodiments of this application can be implemented in software or hardware. The described units or modules can also be housed in a processor; for example, a processor can be described as including a first acquisition module, a second acquisition module, a determination module, and a recommendation module. The names of these units or modules do not necessarily limit the specific unit or module itself; for example, the first acquisition module can also be described as "acquiring the user's currently logged-in login identifier information and the device identifier of the terminal device."

[0138] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments, or may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium stores one or more programs that, when used by one or more processors, execute the message recommendation method described in this application.

[0139] In summary, after a user logs in to the application using their login identifier, the application server obtains the user's current login identifier and device identifier. It then acquires the application feature data corresponding to the terminal device. Based on the login identifier, device identifier, and application feature data, the server determines the target identifier corresponding to the login identifier and retrieves the corresponding recommendation message. This recommendation message is then sent to the terminal device, enabling the application to recommend messages based on the actual identity of the logged-in user, effectively improving the accuracy of message recommendations and thus increasing the recommendation conversion rate.

[0140] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A message recommendation method, characterized in that, include: Obtain the login identifier information of the currently logged-in user and the device identity identifier of the terminal device, wherein the login identifier information is account information shared by multiple objects in the target application; Obtain application feature data corresponding to the terminal device, the application feature data including operation data received by the terminal device on other applications; Based on the login identifier information, the device identity identifier, and the application feature data, the target identifier information corresponding to the login identifier information is determined. The target identifier information is account information in a specific application other than the target application. The target identifier information is used to represent the identity of the currently logged-in real object among the multiple objects. The recommended message corresponding to the target identifier information is obtained based on the target identifier information, and the recommended message is sent to the terminal device.

2. The method according to claim 1, characterized in that, The step of determining the target identifier information corresponding to the login identifier information based on the login identifier information, the device identity identifier, and the application feature data includes: Based on the login identifier information and the device identity identifier, at least one candidate identifier information is determined; The target identifier information is determined based on the candidate identifier information and the application feature data.

3. The method according to claim 2, characterized in that, The step of determining the target identifier information based on the candidate identifier information and the application feature data includes: The candidate identifier information and the feature data are input into the feature learning model to obtain the target identifier information; The feature learning model is trained using historical information from the terminal device.

4. The method according to claim 1, characterized in that, The step of obtaining the recommendation message corresponding to the target identifier information based on the target identifier information includes: Based on the target identification information, obtain the image tag corresponding to the target identification information; Based on the target identifier information and the profile tag, obtain the recommendation message corresponding to the target identifier information.

5. The method according to claim 1, characterized in that, Also includes: Obtain operation information for the recommended message; The operation information is used as the historical information of the terminal device, and the feature learning model is trained using the historical information and the target identification information.

6. A message recommendation device, characterized in that, include: The first acquisition module is used to acquire the login identifier information of the currently logged-in user and the device identity identifier of the terminal device. The login identifier information is account information shared by multiple objects in the target application. The second acquisition module is used to acquire application feature data corresponding to the terminal device, the application feature data including operation data on other applications received by the terminal device. The determination module is used to determine the target identification information corresponding to the login identification information based on the login identification information, the device identity identifier, and the application feature data. The target identification information is account information in a specific application other than the target application. The target identification information is used to represent the identity of the currently logged-in real object among the multiple objects. The recommendation module is used to obtain the recommendation message corresponding to the target identification information based on the target identification information, and send the recommendation message to the terminal device.

7. The apparatus according to claim 6, characterized in that, The determining module is further configured to: Based on the login identifier information and the device identity identifier, at least one candidate identifier information is determined; The target identifier information is determined based on the candidate identifier information and the application feature data.

8. The apparatus according to claim 7, characterized in that, The determining module is further configured to: The candidate identifier information and the feature data are input into the feature learning model to obtain the target identifier information; The feature learning model is trained using historical information from the terminal device.

9. The apparatus according to claim 6, characterized in that, The recommendation module is also used for: Based on the target identification information, obtain the image tag corresponding to the target identification information; Based on the target identifier information and the profile tag, obtain the recommendation message corresponding to the target identifier information.

10. The apparatus according to claim 6, characterized in that, The recommendation module is also used for: Obtain operation information for the recommended message; The operation information is used as the historical information of the terminal device, and the feature learning model is trained using the historical information and the target identification information.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the message recommendation method as described in any one of claims 1-5.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the message recommendation method as described in any one of claims 1-5.

Images