Group chat message recommendation method and device, equipment, medium and program product

Abstract

The invention discloses a group chat message recommendation method and device, equipment, a medium and a program product, and belongs to the field of data processing. The method comprises: converting chat messages in a chat group into phrase data, the phrase data comprising phrases, each phrase comprising a keyword of a corresponding chat message, and the chat messages comprising a target user chat message and a group friend user chat message; obtaining recommendation parameters of the to-be-determined chat messages according to the phrases of the to-be-determined chat messages, the phrases of the chat messages of the target users and the phrases of the chat messages of the group friends forming reply pairs with the chat messages of the target users; under the condition that the recommendation parameter is greater than a preset recommendation triggering threshold value, determining the to-be-determined chat message as a recommendation message; and displaying a prompt identifier on the user interface, wherein the prompt identifier is used for prompting the target user to view the recommended message. According to the embodiment of the invention, the user can be prevented from missing important messages in a large number of chat messages of group chat.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method, apparatus, device, medium, and program product for recommending group chat messages. Background Art

[0002] With the continuous development of information technology, the use of instant messaging applications has become increasingly common. Through instant messaging applications, users can chat individually or form a communication group for group chats. Group chats often contain numerous messages, and if a user has not participated in a group chat for a period of time, a large number of unread messages will appear. Users may not be able to check these unread messages in a timely manner, and may miss some important messages that they are concerned about.

[0003] While some unread messages can be used to remind users to check their messages promptly by @ing a user or by adding them to a special follower list, both @ing a user and adding them to a special follower list require explicit identification of the sender or recipient. For example, @ing a user specifies the recipient, while adding them to a special follower list specifies the sender. Without explicit identification of the sender or recipient, users are still unable to check their messages promptly, potentially missing out on important messages. Summary of the Invention

[0004] Embodiments of the present application provide a method, apparatus, device, medium, and program product for recommending group chat messages, which can prevent users from missing important messages among a large number of group chat messages.

[0005] In a first aspect, an embodiment of the present application provides a method for recommending group chat messages, comprising: converting chat messages within a chat group into phrase data, the phrase data including phrases, a phrase including keywords of a corresponding chat message, and the chat messages including target user chat messages and group friend user chat messages; obtaining recommendation parameters for the chat message to be determined based on the phrase of the chat message to be determined, the phrase of the target user chat message, and the phrase of the group friend user chat message that forms a reply pair with the target user chat message; determining that the chat message to be determined is a recommended message when the recommendation parameter is greater than a preset recommendation trigger threshold; and displaying a prompt mark on the user interface, the prompt mark being used to prompt the target user to view the recommended message.

[0006] In a second aspect, an embodiment of the present application provides a method for recommending group chat messages, including: a preprocessing module for converting chat messages in a chat group into phrase data, the phrase data including phrases, a phrase including keywords of a corresponding chat message, and chat messages including target user chat messages and group friend user chat messages; a recommendation parameter determination module for obtaining recommendation parameters of the chat message to be determined based on the phrase of the chat message to be determined, the phrase of the target user chat message, and the phrase of the group friend user chat message that forms a reply pair with the target user chat message; a recommendation module for determining that the chat message to be determined is a recommended message when the recommendation parameter is greater than a preset recommendation trigger threshold; a display module for displaying a prompt mark on the user interface, the prompt mark being used to prompt the target user to view the recommended message.

[0007] In a third aspect, an embodiment of the present application provides a terminal device comprising: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the method for recommending group chat messages of the first aspect is implemented.

[0008] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the method for recommending group chat messages in the first aspect is implemented.

[0009] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method for recommending group chat messages in the first aspect.

[0010] The embodiments of the present application provide a method, apparatus, device, medium and program product for recommending group chat messages, which can convert chat messages in a chat group into phrases, and learn the target user's reply tendency to the group user's chat information based on the phrases of the target user's chat message and the group user's chat message that form a reply pair. The method can obtain recommendation parameters of the chat message to be determined that can characterize the target user's interest level based on the phrases of the chat message to be determined by forming the reply pair of the target user's chat message and the group user's chat message. The recommended message is determined based on the recommendation parameters, and a prompt mark for prompting the user to view the recommended message is displayed on the user interface. The method can automatically identify the recommended messages that the target user is interested in from a large number of chat messages in the chat group without explicitly indicating the sender or receiver, prompt the target user to browse, and enable the target user to view the messages he or she is interested in in a timely manner to avoid missing important messages in a large number of chat messages. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0012] Figure 1 A flowchart of a method for recommending group chat messages provided in one embodiment of the present application;

[0013] Figure 2 A schematic diagram of an example of a display interface of an instant messaging application provided in an embodiment of the present application;

[0014] Figure 3 A schematic diagram of an example of a message list of an instant messaging application provided in an embodiment of the present application;

[0015] Figure 4 A schematic diagram of an example of a feedback control in a message list provided in an embodiment of the present application;

[0016] Figure 5 A schematic diagram of an example of a setting page for recommendation-related parameters provided in an embodiment of the present application;

[0017] Figure 6 A schematic diagram of an example of the logic of the group chat message recommendation process provided in an embodiment of the present application;

[0018] Figure 7 A schematic diagram of the structure of a device for recommending group chat messages provided in one embodiment of the present application;

[0019] Figure 8 A schematic diagram of the structure of a terminal device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0020] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is merely to provide a better understanding of the present application by illustrating examples of the present application. It should be noted that the acquisition, storage, use, processing, etc. of information and data in the embodiments of the present application are authorized by the user or relevant agencies and comply with the relevant provisions of national laws and regulations.

[0021] With the continuous development of information technology, the use of instant messaging applications is becoming increasingly common. Instant messaging applications allow users to chat individually or form a communication group for group chats. Group chats often contain numerous messages, and if a user has not participated in a group chat for a period of time, a large number of unread messages will appear. Users may not be able to check these unread messages promptly, potentially missing out on important messages they care about. While some unread messages can be used to remind users to check their messages promptly by @ing a user or by adding a user as a special follower, both @ing a user and adding a user as a special follower require explicit indication of the sender or the recipient. For example, @ing a user specifies the recipient, while adding a user as a special follower specifies the sender. Without explicit indication of the sender or recipient, these two methods are ineffective. For example, user A1 sends a chat message in the "Work Group" chat group, "I need the installation package for XXX. Which colleague can help me?" In this scenario, user A1 doesn't @ anyone, and the user who owns the installation package might not be particularly interested in it. If the user who owns the installation package doesn't see this chat message among the numerous chat logs, they will miss it. Therefore, users in the chat group currently cannot view the messages they are interested in in a timely manner, and thus miss important messages.

[0022] The present application provides a group chat message recommendation method, apparatus, device, medium and program product, which can convert chat messages in a chat group into phrases, analyze the phrases obtained by converting the chat messages, and recommend chat messages that the target user is very likely to pay attention to in the chat messages, so that the target user can check the status of the messages he or she is interested in in a timely manner to avoid missing important messages.

[0023] The following describes the recommended method, device, equipment, medium and program product for group chat messages provided by this application.

[0024] The present application provides a method for recommending group chat messages, which can be applied to scenarios where users use instant messaging applications. The method for recommending group chat messages can be executed by a group chat message recommendation device, a user terminal, or other terminal device, and is not limited here. Figure 1 This is a flow chart of a method for recommending group chat messages provided in one embodiment of the present application, such as Figure 1 As shown, the method for recommending group chat messages may include steps S101 to S104.

[0025] In step S101, chat messages in a chat group are converted into phrase data.

[0026] A chat group may include multiple users, and the multiple users may include a target user and group friend users. The target user may include the currently logged-in user of the instant messaging application, and may also be considered as a user of a user terminal that performs the group chat message recommendation method in the embodiment of the present application. Group friend users include other users in the chat group other than the target user. Chat messages include target user chat messages and group friend user chat messages. Target user chat messages include chat messages sent by the target user. Group friend user chat messages include chat messages sent by group friend users.

[0027] During the initialization phase of enabling the group chat message recommendation feature, all historical chat messages within the chat group can be converted into phrase data. After the group chat message recommendation feature is enabled, each new chat message that appears in the chat group is converted into phrase data. Each phrase data item corresponds to a chat message, and phrase data includes phrases, each of which includes keywords from the corresponding chat message. The chat message can be segmented to obtain multiple words, and keywords can be selected from the multiple words obtained through segmentation, and the keywords can be added to the phrases. The segmentation method used for chat message segmentation is not limited herein; for example, the segmentation method may include a segmentation method based on a large language model or a segmentation algorithm.

[0028] In some examples, chat messages can be input into a large language model, which then converts the chat information into phrase data. A chat group can contain a variety of chat messages, including, but not limited to, sentences, voice messages, text files, images, and videos. The large language model's file summarization and image recognition functions can be used to convert these various chat messages into phrases. For example, the large language model can be used to summarize keywords from files in a chat message, and phrases can be derived from these keywords. The large language model can also be used to identify the content of an image and summarize keywords from the image content, and phrases can be derived from these keywords. A chat message knowledge base can be constructed based on the phrase data in the chat messages, and this knowledge base can be connected to the large language model. When a target user receives a chat message from a group member that requires a reply, the target user can input the chat message from the group member into the large language model, and the large language model can generate a reply chat message based on the chat message knowledge base.

[0029] In some examples, a word segmentation algorithm can be used to segment chat messages, and phrase data can be obtained based on the words obtained from the word segmentation. If the chat message includes sentences, voice, text files, pictures, videos, etc., the sentences, voice, files, pictures, videos, etc. can be converted into text through various means, and then the text can be segmented using a word segmentation algorithm. For shorter text in chat messages, word segmentation libraries such as Jieba and THULAC can be used to decompose the text into words, and then keywords can be filtered from the decomposed words. For example, the text "I need the latest installation package of XXX" can be decomposed into words such as "I", "need", "one", "XXX", "of", "latest", and "installation package". For longer text in chat messages, such as long text converted from text files in formats such as doc and txt, the term frequency-inverse document frequency (TF-IDF) algorithm can be used to extract keywords from the text. Specifically, the frequency of a word in a text can be represented by word frequency. For example, the total number of words in text d is N, and the number of times word t appears in text d is n, then the word frequency TF(t, d) of word t in text d = n / N; the inverse document frequency can represent the prevalence of documents containing a word. The fewer documents containing a word, the greater the inverse document frequency corresponding to the word, indicating that the word has better category distinction ability. For example, a message database has D texts, of which d texts contain word t, then the inverse document frequency IDF(t) of word t = log[D / (d+1)]; the word frequency and the inverse document frequency are multiplied to obtain the word frequency-inverse document frequency, for example, word frequency-inverse document frequency TF-IDF(t, d) = TF(t, d)×IDF(t); the greater the word frequency-inverse document frequency corresponding to a word, the greater the importance of the word in the text; words whose word frequency-inverse document frequency is greater than a preset threshold can be determined as keywords and added to the phrase.

[0030] In some examples, when a chat message includes a file, the phrase also includes file keywords derived from the file's file information. Files in this context may include text files, images, videos, audio files, and other file formats, though this is not a limitation. To enrich the content of the phrase in the chat message corresponding to the file, file keywords may be derived from the file information and added to the phrase. File keywords include keywords extracted from the file information. File information may include one or more of the following information, though this is not a limitation.

[0031] The phrase data of the chat message can be stored in the local database of the user terminal client so that the phrase data can be quickly obtained in the subsequent process. By converting the chat message into phrase data, it is easy to obtain similar chat messages in the chat message and carry out the subsequent process.

[0032] In step S102, the recommended parameters of the chat message to be determined are obtained based on the phrases of the chat message to be determined, the phrases of the target user's chat message, and the phrases of the chat messages of group friends that form reply pairs with the target user's chat message.

[0033] The chat messages to be determined are chat messages whose identification is yet to be determined as recommended messages for the user. The chat messages to be determined are chat messages sent by group members, that is, chat messages sent by group members in the chat group. In some examples, each latest message sent by a group member in the chat group can be considered a chat message to be determined.

[0034] Among the multiple chat messages in a chat group, there are message pairs that form a reply conversation. For ease of explanation, the message pairs forming a reply conversation, consisting of a friend user chat message and a target user chat message, are referred to herein as reply pairs. Each reply pair includes a friend user chat message and a target user chat message. The target user chat message in each reply pair is the reply message to the friend user chat message. Whether the target user chat message and the friend user chat message can form a reply pair can be determined based on the reply method between the target user chat message and the friend user chat message. Reply methods can include explicit reply methods, implicit reply methods, and no reply methods. If the reply method between the target user chat message and the friend user chat message is explicit reply method or implicit reply method, the target user chat message and the friend user chat message can form a reply pair. If the reply method between the target user chat message and the friend user chat message is no reply method, the target user chat message and the friend user chat message cannot form a reply pair. Explicit reply methods can include reference reply methods, quick reply methods, and other reply methods that are directly associated with the friend user chat message. Implicit reply methods may include privately chatting with a group friend after receiving a chat message from a group friend, or sending a chat message related to but not directly related to the group friend's chat message within the chat group after receiving a chat message from a group friend. The group friend chat message that the target user's chat message replies to using the implicit reply method can be determined by the word repetition rate of the phrases in the target user's chat message and the phrases in the target user's chat message. The word repetition rate of the phrases in the m group friend chat messages before the target user's chat message and the phrases in the target user's chat message can be calculated, and the group friend chat message with the highest word repetition rate is used as the group friend chat message that the target user's chat message replies to using the implicit reply method. No reply method indicates that the target user's chat message is unrelated to the group friend chat message.

[0035] The target user's reply tendency to the group friend user's chat message can be obtained based on the phrases in the target user's chat message in the reply pair and the phrases in the group friend user's chat message in the reply pair, and the reply tendency can characterize the target user's interest in the group friend user's chat message. According to the target user's reply tendency to the group friend's chat message, the target user's reply tendency to the determined chat message can be determined. In an embodiment of the present application, the target user's reply tendency to the determined chat message is characterized by a recommendation parameter. Similarly, the recommendation parameter can characterize the target user's interest in the determined chat message. The recommendation parameter of the chat message to be determined can be positively correlated with the target user's reply tendency to the determined chat message, that is, the larger the recommendation parameter, the higher the target user's reply tendency to the determined chat message, the higher the target user's interest in the determined chat message, and the higher the possibility of determining the chat message to be determined as a recommended message.

[0036] In step S103, when the recommendation parameter is greater than a preset recommendation trigger threshold, the chat message to be determined is determined to be a recommended message.

[0037] The recommendation trigger threshold is the threshold for determining whether a chat message is a recommended message. It can be determined based on the scenario, needs, experience, etc. and is not limited here. It should be noted that the recommendation trigger threshold is adjustable. The recommendation trigger threshold can be adjusted based on user input, user feedback on recommended messages, etc., thereby continuously optimizing the accuracy of recommended messages and the relevance of recommended messages to target users.

[0038] If the recommendation parameter of the pending chat message is greater than the recommendation trigger threshold, it indicates that the target user is inclined to reply to the pending chat message. That is, the target user is interested in the pending chat message, and the confirmed chat message can be recommended to the user as a recommended message. In some examples, the phrases of the chat messages that the target user is interested in can be stored in a database. After receiving the recommended message, the phrases of the recommended message can also be stored in the database. The phrases of the recommended message can be used as phrases for replying to chat messages from group users in the subsequent message recommendation process.

[0039] In step S104, a prompt icon is displayed on the user interface.

[0040] The user interface may include, but is not limited to, the display interface of an instant messaging application, the operating system of a user terminal, the notification bar interface of the operating system of a user terminal, the lock screen interface of the user terminal, etc. The prompt indicator is used to prompt the target user to view the recommended message. The prompt indicator can be implemented in the form of, but not limited to, text, static graphics, dynamic graphics, voice, etc., and is not limited here. After the target user observes the prompt indicator, they can further input the prompt indicator on the user interface to jump to the recommended message, so that the user can directly and quickly browse the recommended message without being affected by other irrelevant messages in the chat group.

[0041] In some examples, in response to the target user's input of a chat group with a prompt identifier, a message list with a recommended message jump identifier can be displayed; in response to the user's input of the recommended message jump identifier, the recommended message can be jumped to the location of the recommended message to display the recommended message. If there are multiple chat groups on the display interface of the instant messaging application, the prompt identifier can be displayed at the corresponding location of the chat group where the recommended message is located, and the user inputs the chat group with the prompt identifier, and the input may include but is not limited to single-click, double-click, long press and other operations. In response to the input, the user terminal can display a message list of the chat group. The message list generally displays a preset number of chat messages that have appeared recently in the chat group, but the preset number of chat messages displayed in the message list may not include recommended messages. The recommended message jump identifier can be input to jump to the location of the recommended message, so that the target user can directly browse the recommended message without having to search for the recommended message one by one in the numerous chat messages. The configuration information of the recommended message can be changed so that the display method of the recommended message is different from that of other chat messages. For example, the color and font of the recommended message can be controlled to be different from the color and font of other chat messages, or the chat bubble box of the recommended message can be controlled to be different from the chat bubble box of other chat messages. This is not limited here.

[0042] For example, Figure 2 This is a schematic diagram of an example of a display interface of an instant messaging application provided in an embodiment of the present application. Figure 3 A schematic diagram of an example of a message list of an instant messaging application provided in an embodiment of the present application, such as Figure 2 As shown, the display interface 21 includes five chat groups, wherein chat group 1 has a prompt mark 22, and the content of the prompt mark 22 is "[There is a recommended message]". The user can click on the prompt mark 22 to enter the chat group 1. Figure 3The message list 23 shown shows the five most recent chat messages 24, but these five chat messages 24 do not contain recommended messages. The recommended messages are in earlier chat messages. The message list has a recommended message jump mark 25. The content of the recommended message jump mark 25 is "︽There is a recommended message". The user can click the recommended message jump mark 25 in the message list 23 to jump to the location of the recommended message.

[0043] In an embodiment of the present application, chat messages within a chat group can be converted into phrases. Based on the phrases of the target user's chat message and the group friend's chat message that form a reply pair, the target user's reply tendency to the group friend's chat message can be learned. The target user's chat message and the group friend's chat message that form a reply pair can be used to obtain recommendation parameters of the chat message to be determined that can characterize the target user's interest level based on the phrases of the chat message to be determined. The recommended message can be determined based on the recommendation parameters, and a prompt mark for prompting the user to view the recommended message can be displayed on the user interface. The system can automatically identify the recommended messages that the target user is interested in from a large number of chat messages in the chat group without explicitly indicating the sender or receiver, prompt the target user to browse, and enable the target user to view the messages he or she is interested in in a timely manner to avoid missing important messages in a large number of chat messages.

[0044] In some embodiments, the recommended parameters for the chat message to be determined can be obtained by using the word repetition rate. The word repetition rate of the phrase in the chat message to be determined and the phrases in the chat messages of each target user, as well as the word repetition rate of the phrase in the chat message to be determined and the phrases in the chat messages of the group members in the reply pair, can be obtained. The recommended parameters for the chat message to be determined are calculated based on the word repetition rate, the weight coefficient corresponding to the phrase in the chat message to be determined, and the weight coefficient corresponding to the phrase in the chat messages of the group members in the reply pair.

[0045] The word repetition rate of two phrases can reflect the similarity between the chat messages corresponding to the two words. The greater the word repetition rate of two phrases, the higher the similarity between the chat messages corresponding to the two words; the smaller the word repetition rate of two phrases, the lower the similarity between the chat messages corresponding to the two words. The word repetition rate of two phrases can be determined by the ratio of the number of identical words in the two phrases to the number of words in the phrases. For example, the word repetition rate of two phrases can be obtained according to the following formula (1):

[0046] r = N A / N (1)

[0047] Among them, r is the word repetition rate; N Ais the number of words in the intersection of the two phrases, that is, the number of identical words that appear in both phrases; N is the number of words in the phrase with fewer words in the two phrases.

[0048] The number of chat messages from the target user and the number of chat messages from group members in the reply pair may be relatively large. Therefore, the phrases from the first n chat messages, ranked from highest to lowest word repetition rate with the phrase in the chat message to be determined, can be selected from the chat messages from the target user and the chat messages from group members in the reply pair. Based on the word repetition rate of the phrases from these first n chat messages with the phrase in the chat message to be determined, as well as the weight coefficients of the phrases from these first n chat messages, a weighted algorithm is used to obtain recommended parameters for the chat message to be determined. In some examples, the weight coefficients of the phrases from the first n chat messages may be equal, namely, 1 / n, meaning that the recommended parameters for the chat message to be determined are the average of the word repetition rates of the phrases from the first n chat messages with the phrase in the chat message to be determined. In other examples, the weight coefficients of the phrases from the first n chat messages may differ at least in part. The weight coefficients may be determined based on the reply method between the chat messages from group members and the chat message from the target user. For example, reply methods may include explicit reply, implicit reply, and no reply. For details, please refer to the relevant descriptions in the above embodiments and will not be repeated here. Since implicit reply methods are uncertain, correspondingly, the weight coefficient corresponding to the explicit reply method is higher than the weight coefficient corresponding to the implicit reply method, and the weight coefficient corresponding to the implicit reply method is higher than the weight coefficient corresponding to the no reply method. In some examples, the phrase data may also include a phrase reply matching score. The phrase reply matching score is obtained based on the reply method between the chat message of the group user and the chat message of the target user. The phrase reply matching score corresponding to the explicit reply method is higher than the phrase reply matching score corresponding to the implicit reply method, and the phrase reply matching score corresponding to the implicit reply method is higher than the phrase reply matching score corresponding to the no reply method. For example, the phrase reply matching score corresponding to the no reply method can be set to 0. The phrase reply matching score is positively correlated with the weight coefficient; the higher the phrase reply matching score, the higher the corresponding weight coefficient; the lower the phrase reply matching score, the lower the corresponding weight coefficient.

[0049] In some embodiments, a word vector model can be used to obtain recommended parameters for the chat message to be determined. A pre-trained word vector model can be used to convert the target user's chat message phrase and the phrases in the group user's chat message in the reply pair into a first word vector, and the phrase in the chat message to be determined into a second word vector. Based on the first word vector and the weight coefficient corresponding to the first word vector, a user interest vector of the target user is obtained. Based on the similarity between the second word vector and the user interest vector, recommended parameters for the chat message to be determined are obtained.

[0050] The word vector model may include a Word2Vec model. The word vector model is trained based on phrases and a vocabulary obtained by converting chat messages within a chat group. The vocabulary is constructed based on phrases converted from chat messages within the chat group. Low-frequency words and stop words can be removed from the phrases converted from the chat messages, and a vocabulary is constructed based on the words in the phrases after removing the low-frequency words and stop words. The Word2Vec model is trained using the phrases and vocabulary obtained from the chat messages. For example, a Skip-gram model or a Continuous Bag of Words (CBOW) model can be selected. Model parameters such as the vector dimension, context window size, and number of iterations can be set. The Word2Vec model is trained using the phrases and vocabulary obtained from the chat messages. The first word vector includes the word vector obtained by converting the target user's chat phrase and the word vector obtained by converting the phrases of the group users in the reply pair. The second word vector includes the word vector obtained by converting the phrases of the chat message to be determined. The user interest vector of the target user can be considered as the vector of the phrases of interest to the target user. In some examples, the weight coefficients corresponding to each first word vector are equal, and the user interest vector of the target user is the average of the multiple first word vectors. In other examples, the weight coefficients corresponding to each first word vector are at least partially different, and the user interest vector of the target user may be the weighted sum of the multiple first word vectors based on the weight coefficients. The similarity between the second word vector and the user interest vector can represent the similarity between the phrases in the chat message to be determined and the phrases of interest to the target user. The similarity between the second word vector and the user interest vector can be obtained using, but not limited to, metrics such as Euclidean distance and cosine similarity. The similarity between the second word vector and the user interest vector is positively correlated with the recommendation parameter. The higher the similarity between the second word vector and the user interest vector, the higher the recommendation parameter of the chat message to be determined, and the greater the likelihood that the chat message to be determined will be determined as a recommended message. The lower the similarity between the second word vector and the user interest vector, the lower the recommendation parameter of the chat message to be determined, and the lower the likelihood that the chat message to be determined will be determined as a recommended message.

[0051] In some embodiments, the recommendation trigger threshold can be adjusted based on the degree of match between the recommended message and the chat message of the friend user identified as a reply pair to optimize the accuracy of recommended messages in group chat messages. If the number of times a friend user's chat message in a reply pair is not identified as a recommended message is greater than or equal to a first preset threshold, the recommendation trigger threshold can be lowered; if the number of times a recommended message is not identified as a friend user's chat message in a reply pair is greater than or equal to a second preset threshold, the recommendation trigger threshold can be increased. The first and second preset thresholds can be set based on scenarios, needs, experience, and other factors, and are not limited here. The first threshold can be an integer of 1 or greater, and the second threshold can be an integer of 1 or greater. The target user's chat message in the reply pair is a reply to the friend user's chat message, meaning that the friend user's chat message in the reply pair is a chat message of interest and interest to the target user. If a friend user's chat message in a reply pair is not identified as a recommended message, this indicates that the recommendation trigger threshold is set too high, and the friend user's chat message, which should have been identified as a recommended message, was not. Accordingly, the recommendation trigger threshold can be lowered to improve the accuracy of identifying recommended messages. If the recommended message is not determined as a reply to the chat message of the group user, it means that the recommendation trigger threshold is set too low, and the chat message of the group user that should not be determined as a recommended message is determined as a recommended message. Accordingly, the recommendation trigger threshold should be increased to improve the accuracy of determining the recommended message.

[0052] In some embodiments, the accuracy of recommended messages in group chat messages can be optimized based on the target user's feedback on the recommended messages. The recommendation trigger threshold can be adjusted based on the target user's feedback operation on the recommended message. Feedback operations can include explicit operations and implicit operations. Display operations can include but are not limited to input operations on feedback controls, quote reply operations, quick reply operations, etc. For example, Figure 4 This is a schematic diagram of an example of a feedback control in a message list provided in an embodiment of the present application, such as Figure 4As shown, message list 23 includes a recommended message 26. A feedback control 27 is provided in the location corresponding to recommended message 26. Feedback control 27 can be specifically implemented as an "Interested" button and a "Not Interested" button. If a user clicks the "Interested" button, the number of times the user enters the "Interested" button and the number of times the user enters the "Not Interested" button can be counted. If the number of times the user enters the "Interested" button is greater than or equal to a third preset threshold, the recommendation trigger threshold can be lowered; if the number of times the user enters the "Not Interested" button is greater than or equal to a fourth preset threshold, the recommendation trigger threshold can be raised. If the number of reference reply operations and / or quick reply operations exceeds or equal to a fifth preset threshold, the recommendation trigger threshold can be lowered. Implicit operations can include privately chatting with a group member after receiving a chat message from the group member, sending a chat message related to the group member's chat message but not directly related to the chat message from the group member after receiving the chat message from the group member, and so on. The recommendation trigger threshold can be lowered if the number of implicit operations is greater than or equal to a sixth preset threshold.

[0053] In some embodiments, an independent group chat recommendation function may be used or not, and the independent group chat recommendation function may determine recommended messages for each group chat separately. If the number of chat groups is two or more and it is determined to use the independent group chat recommendation function, the recommended messages in one chat group are determined by the chat messages in one chat group, that is, in each chat group, based on the phrases of the chat messages to be determined in the chat group, the phrases of the target user's chat messages, and the phrases of the chat messages of the group friends that form a reply pair with the target user's chat messages, the recommendation parameters of the chat messages to be determined are obtained, and the recommended messages for the chat group are determined based on the recommended parameters. If the number of chat groups is two or more and it is determined not to use the independent group chat recommendation function, the recommended messages in each chat group are comprehensively determined by the chat messages in all chat groups, that is, based on the phrases of the chat messages to be determined in all chat groups, the phrases of the target user's chat messages, and the phrases of the chat messages of the group friends that form a reply pair with the target user's chat messages, the recommendation parameters of the chat messages to be determined are obtained, and the recommended messages for each chat group are determined based on the recommended parameters. The content in different chat groups is different. The same target user has different roles in different chat groups and is interested in different messages. The independent group chat recommendation function makes the determination of recommended messages more targeted and more in line with the role of the target user in the independent chat group, further improving the accuracy of recommended messages.

[0054] In some embodiments, before step S102, a method for obtaining recommended parameters may be selected based on the performance of the user terminal, so that the method for obtaining recommended parameters matches the performance of the user terminal, thereby improving the efficiency of the user terminal in executing the group chat message recommendation method. Configuration parameters of the target user's user terminal may be obtained. If the user terminal performance represented by the configuration parameters is lower than a preset performance standard, a first determination method may be used as the method for determining recommended parameters. If the user terminal performance represented by the configuration parameters is higher than or equal to the preset performance standard, a second determination method may be used as the method for determining recommended parameters. Configuration parameters may represent the performance of the target user's user terminal. For example, configuration parameters may include, but are not limited to, the CPU model, GPU model, memory size, and hard disk size. The preset performance standard may be used to determine the relative performance of the user terminal. If the user terminal performance represented by the configuration parameters is lower than the preset performance standard, it indicates relatively low performance of the user terminal; if the user terminal performance represented by the configuration parameters is higher than or equal to the preset performance standard, it indicates relatively high performance of the user terminal. The first determination method includes a method for obtaining recommended parameters based on word repetition rate, and the second determination method includes a method for obtaining recommended parameters based on a word vector model. The details of the methods for obtaining recommended parameters based on word repetition rate and the methods for obtaining recommended parameters based on a word vector model can be found in the relevant descriptions of the above embodiments and will not be repeated here. The method of recommending parameters based on word repetition rate has lower requirements for user terminal performance and can be used when the user terminal performance represented by the configuration parameters is lower than the preset performance standard. The method of recommending parameters based on word vector models has higher requirements for user terminal performance and can be used when the user terminal performance represented by the configuration parameters is higher than or equal to the preset performance standard.

[0055] In some embodiments, based on the user's input of recommendation-related parameters, processing logic corresponding to the inputted recommendation-related parameters may be executed to adjust the processing logic in the group chat message recommendation method. Recommendation-related parameters may include, but are not limited to, one or more of the following: recommendation sensitivity, recommended message sending time, recommended message sender, independent group chat recommendation parameters, and recommendation parameter determination method. Figure 5 A schematic diagram of an example of a setting page for recommendation-related parameters provided in an embodiment of the present application, such as Figure 5 As shown, recommendation sensitivity can be set using a slider, while the recommended message sending time, recommended message sender, recommended parameters for independent group chats, and the method for determining recommended parameters can be set using checkboxes. By allowing users to set recommendation-related parameters, chat message recommendations can be more tailored to user needs and provide greater flexibility.

[0056] When the recommendation-related parameters include recommendation sensitivity, the processing logic corresponding to the recommendation sensitivity includes: adjusting the recommendation trigger threshold according to the recommendation sensitivity parameter. Recommendation sensitivity affects the number of recommended messages and the probability that the target user is interested in the recommended messages. As the recommendation sensitivity increases, the recommendation trigger threshold is lowered, resulting in an increase in the number of recommended messages and a relatively lower probability that the target user is interested in the recommended messages. As the recommendation sensitivity decreases, the recommendation trigger threshold is raised, resulting in a decrease in the number of recommended messages and a relatively higher probability that the target user is interested in the recommended messages.

[0057] When the recommendation-related parameters include a recommended message sending time, the processing logic corresponding to the recommended message sending time includes: obtaining the chat message to be determined from the chat messages of group members that match the recommended message sending time. The recommended message sending time can be set by the user, and chat messages sent within the set recommended message sending time can be filtered from the chat messages of group members as the chat messages to be determined. In other words, the sending time of the recommended message obtained from the chat message to be determined is also within the recommended message sending time.

[0058] When the recommendation-related parameters include a recommended message sender, the processing logic for the recommended message sender includes obtaining chat messages to be determined from chat messages of group members whose sender is the recommended message sender. The recommended message sender can be set by the user, and chat messages from group members whose sender is the set recommended message sender can be filtered as chat messages to be determined. In other words, the sender of the recommended message obtained from the chat messages to be determined is also the recommended message sender.

[0059] The recommendation-related parameters include independent group chat recommendation parameters, and the processing logic corresponding to the independent group chat recommendation parameters includes: if the independent group chat recommendation parameters represent the use of the independent group chat recommendation function, the recommended messages in a chat group are determined through the chat messages in a chat group; if the independent group chat recommendation parameters represent that the independent group chat recommendation function is not used, the recommended messages in each chat group are comprehensively determined through the chat messages in all chat groups. Different chat groups have different natures. For example, some chat groups are related to news recommendations, some chat groups are related to work content, and some chat groups are related to product recommendations. The same target user has different roles in chat groups of different natures, and the types of messages that tend to be replied to are also different. Message recommendations are made separately according to the chat groups, which is more in line with the roles of the target users in different groups. For the specific content of using the independent group chat recommendation function and not using the independent group chat recommendation function, please refer to the relevant descriptions in the above embodiments, which will not be repeated here.

[0060] The recommendation-related parameters include a recommendation parameter determination method, and the processing logic corresponding to the recommendation parameter determination method includes: if the recommendation parameter determination method represents the first determination method, then the method of obtaining the recommendation parameter based on the word repetition rate is used; if the recommendation parameter determination method represents the second determination method, then the method of obtaining the recommendation parameter based on the word vector model is used. For example, Figure 5 As shown, if the check box for the recommended parameter determination method is checked, it represents the second determination method; if the check box for the recommended parameter determination method is not checked, it represents the first determination method. The details of the first and second determination methods can be found in the relevant descriptions of the above embodiments and will not be repeated here.

[0061] For ease of understanding, the following example illustrates the process of recommending group chat messages. Figure 6 A schematic diagram of an example of the logic of the group chat message recommendation process provided in the embodiment of the present application is as follows: Figure 6 As shown, the group chat message recommendation process may include a preprocessing process 31, a model optimization process 32, and a message recommendation process 33. The preprocessing process 31 includes steps a1 to a4, the model optimization process 32 includes steps a5 to a7, and the message recommendation process 33 includes steps a8 to a12.

[0062] In step a1, a chat message is obtained.

[0063] In step a2, the chat message is parsed and converted into phrases.

[0064] In step a3, the phrases in the chat message are scored to obtain a phrase reply matching score for the phrases in the chat message.

[0065] In step a4, the phrases of the chat message are stored in the database as processed messages.

[0066] In step a5, a feedback operation of the target user with respect to the recommended message is received.

[0067] In step a6, the recommendation trigger threshold is adjusted according to the feedback operation.

[0068] In step a7, recommendation-related parameters are set according to the user's input.

[0069] In step a8, a new chat message is obtained.

[0070] In step a9, the new chat message is parsed and converted into phrases.

[0071] In step a10, the recommended parameters of the new chat message are obtained according to the phrase of the new chat message, the processed message and the recommendation-related parameters.

[0072] In step a11, it is determined whether to recommend the new chat message based on the recommendation parameters. If recommended, step a12 is executed; if not, the phrase in the new chat message is stored in the database as a phrase in a message that the target user does not care about.

[0073] In step a12, the new chat message is recommended to the target user, and the phrases of the new chat message are stored in the database as phrases of messages that the target user is interested in.

[0074] The specific contents of the above steps a1 to a12 can be found in the relevant descriptions in the above embodiments and will not be repeated here.

[0075] This application also provides a device for recommending group chat messages. Figure 7 This is a structural diagram of a group chat message recommendation device provided in an embodiment of the present application, such as Figure 7 As shown, the group chat message recommendation device 400 may include a pre-processing module 401 , a recommendation parameter determination module 402 , a recommendation module 403 and a display module 404 .

[0076] The pre-processing module 401 can be used to convert chat messages in the chat group into phrase data, where the phrase data includes phrases. A phrase includes keywords of a corresponding chat message, and the chat messages include chat messages of the target user and chat messages of group friends.

[0077] The recommended parameter determination module 402 can be used to obtain recommended parameters for the chat message to be determined based on the phrases of the chat message to be determined, the phrases of the target user's chat message, and the phrases of the group friend user's chat messages that form a reply pair with the target user's chat message.

[0078] The recommendation module 403 may be configured to determine that the chat message to be determined is a recommended message when the recommendation parameter is greater than a preset recommendation trigger threshold.

[0079] The display module 404 may be configured to display a prompt icon on the user interface, where the prompt icon is configured to prompt the target user to view the recommended message.

[0080] In some embodiments, the recommendation parameter determination module 402 can be specifically used to: obtain the word repetition rate between the phrases of the chat message to be determined and the phrases of the chat messages of each target user, and the word repetition rate between the phrases of the chat message to be determined and the phrases of the chat messages of the group friend users in the reply pair; calculate the recommended parameters of the chat message to be determined based on the word repetition rate, the weight coefficient corresponding to the phrases of the chat message to be determined and the weight coefficient corresponding to the phrases of the chat messages of the group friend users in the reply pair.

[0081] In some embodiments, the recommendation parameter determination module 402 can be specifically used to: use a pre-trained word vector model to convert the phrases of the target user's chat message and the phrases of the group user's chat message in the reply pair into a first word vector, and convert the phrases of the chat message to be determined into a second word vector. The word vector model is trained based on the phrases and vocabulary obtained by converting the chat messages in the chat group, and the vocabulary is constructed based on the phrases obtained by converting the chat messages in the chat group; based on the first word vector and the weight coefficient corresponding to the first word vector, the user interest vector of the target user is obtained; based on the similarity between the second word vector and the user interest vector, the recommended parameters of the chat message to be determined are obtained.

[0082] In some embodiments, the group chat message recommendation device 400 may further include a parameter adjustment module. The parameter adjustment module may be configured to: lower the recommendation trigger threshold if the number of times a group friend user's chat message in a reply pair is not determined as a recommended message is greater than or equal to a first preset number threshold; and raise the recommendation trigger threshold if the number of times a recommended message is not determined as a group friend user's chat message in a reply pair is greater than or equal to a second preset number threshold.

[0083] In some embodiments, the parameter adjustment module may be used to adjust the recommendation triggering threshold according to the target user's feedback operation on the recommendation message.

[0084] In some embodiments, pre-processing module 401 may be configured to: input a chat message into a large language model and convert the chat message into phrase data using the large language model; or segment the chat message using a word segmentation algorithm and obtain phrase data based on the segmented words. If the chat message includes a file, the phrase may also include file keywords obtained based on the file information.

[0085] In some examples, the phrase data also includes a phrase reply match score. The phrase reply match score is based on the reply method between the chat message of the group user and the chat message of the target user. The reply methods include explicit reply method, implicit reply method, and no reply method. The phrase reply match score corresponding to the explicit reply method is higher than the phrase reply match score corresponding to the implicit reply method, and the phrase reply match score corresponding to the implicit reply method is higher than the phrase reply match score corresponding to the no reply method. The phrase reply match score is positively correlated with the weight coefficient.

[0086] In some embodiments, if the number of chat groups is more than two and it is determined to use the independent group chat recommendation function, the recommended messages in one chat group are determined through the chat messages in one chat group; if the number of chat groups is more than two and it is determined not to use the independent group chat recommendation function, the recommended messages in each chat group are determined comprehensively through the chat messages in all chat groups.

[0087] In some embodiments, the parameter adjustment module can be used to: obtain the configuration parameters of the user terminal of the target user; when the performance of the user terminal represented by the configuration parameters is lower than the preset performance standard, use the first determination method as the method for determining the recommended parameters, and the first determination method includes a method for obtaining the recommended parameters based on the word repetition rate; when the performance of the user terminal represented by the configuration parameters is higher than or equal to the preset performance standard, use the second determination method as the method for determining the recommended parameters, and the second determination method includes a method for obtaining the recommended parameters based on the word vector model.

[0088] In some embodiments, the parameter adjustment module may be configured to: execute processing logic corresponding to the recommendation-related parameters after the setting input by the user, based on the setting input of the recommendation-related parameters.

[0089] In the case where the recommendation-related parameters include recommendation sensitivity, the processing logic corresponding to the recommendation sensitivity includes: adjusting the recommendation trigger threshold according to the recommendation sensitivity parameter.

[0090] In the case where the recommendation-related parameters include the recommended message sending time, the processing logic corresponding to the recommended message sending time includes: obtaining the chat message to be determined from the chat messages of group friends that meet the recommended message sending time.

[0091] In the case where the recommendation-related parameters include the recommended message sender, the processing logic corresponding to the recommended message sender includes: obtaining the chat message to be determined from the chat messages of the group friend users whose sender is the recommended message sender.

[0092] The recommendation-related parameters include independent group chat recommendation parameters, and the processing logic corresponding to the independent group chat recommendation parameters includes: if the independent group chat recommendation parameters indicate the use of the independent group chat recommendation function, the recommended messages in a chat group are determined through the chat messages in a chat group; if the independent group chat recommendation parameters indicate that the independent group chat recommendation function is not used, the recommended messages in each chat group are comprehensively determined through the chat messages in all chat groups.

[0093] The recommendation-related parameters include a recommendation parameter determination method, and the processing logic corresponding to the recommendation parameter determination method includes: if the recommendation parameter determination method represents a first determination method, then a method of obtaining the recommendation parameter based on the word repetition rate is used; if the recommendation parameter determination method represents a second determination method, then a method of obtaining the recommendation parameter based on the word vector model is used.

[0094] It should be noted that the group chat message recommendation device 400 is a device corresponding to the above-mentioned group chat message recommendation method. All implementation methods in the above-mentioned method embodiments are applicable to the embodiments of the device and can achieve the same technical effects, which will not be repeated here.

[0095] The present application also provides a terminal device. Figure 8 A schematic diagram of the structure of a terminal device provided in an embodiment of the present application is shown in FIG. Figure 8 As shown, the terminal device 500 includes a memory 501 , a processor 502 , and a computer program stored in the memory 501 and executable on the processor 502 .

[0096] In some examples, the processor 502 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0097] The memory 501 may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Therefore, generally, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method for recommending group chat messages according to the embodiments of the present application.

[0098] The processor 502 runs a computer program corresponding to the executable program code by reading the executable program code stored in the memory 501, so as to implement the method for recommending group chat messages in the above embodiment.

[0099] In some examples, the terminal device 500 may further include a communication interface 503 and a bus 504. Figure 8 As shown, the memory 501 , the processor 502 , and the communication interface 503 are connected via a bus 504 and communicate with each other.

[0100] The communication interface 503 is mainly used to implement communication between the modules, devices, units and / or equipment in the embodiment of the present application. Input devices and / or output devices can also be connected through the communication interface 503.

[0101] The bus 504 includes hardware, software, or both, and couples the components of the terminal device 500 to each other. By way of example, and not limitation, the bus 504 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-E) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, the bus 504 may include one or more buses. Although embodiments herein describe and illustrate a particular bus, this application contemplates any suitable bus or interconnect.

[0102] The present application also provides a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the method for recommending group chat messages in the above-mentioned embodiment can be implemented, and the same technical effects can be achieved. To avoid repetition, the above-mentioned computer-readable storage medium may include a non-transitory computer-readable storage medium, such as a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc., and is not limited here.

[0103] The present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the method for recommending group chat messages in the above embodiment and can achieve the same technical effect. To avoid repetition, it will not be described here.

[0104] It should be understood that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. For device embodiments, equipment embodiments, computer-readable storage medium embodiments, and computer program product embodiments, the relevant parts can be referred to the description section of the method embodiment. This application is not limited to the specific steps and structures described above and shown in the figures. Those skilled in the art can make various changes, modifications and additions, or change the order of the steps after understanding the spirit of this application. In addition, for the sake of brevity, a detailed description of known method technologies is omitted here.

[0105] Aspects of the present application have been described above with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine so that these instructions executed via the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. This processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor or a field programmable logic circuit. It is also understood that each box in the block diagram and / or the flowchart and the combination of the boxes in the block diagram and / or the flowchart can also be implemented by the dedicated hardware that performs the specified function or action, or can be implemented by the combination of dedicated hardware and computer instructions.

[0106] Those skilled in the art should understand that the above embodiments are illustrative rather than restrictive. Different technical features appearing in different embodiments can be combined to achieve beneficial effects. Based on a study of the drawings, the specification and the claims, those skilled in the art should be able to understand and implement other variations of the disclosed embodiments. In the claims, the term "comprising" does not exclude other devices or steps; the quantifier "one" does not exclude a plurality; the terms "first" and "second" are used to identify names rather than to indicate any specific order. Any figure marks in the claims should not be understood as limiting the scope of protection. The functions of multiple parts appearing in the claims can be implemented by a separate hardware or software module. The fact that certain technical features appear in different dependent claims does not mean that these technical features cannot be combined to achieve beneficial effects.

Claims

1. A method for recommending group chat messages, characterized in that: include: Convert chat messages in the chat group into phrase data, where the phrase data includes phrases, and a phrase includes keywords of a corresponding chat message, and the chat messages include chat messages of the target user and chat messages of group members; Obtaining recommended parameters for the chat message to be determined based on the phrase of the chat message to be determined, the phrase of the target user's chat message, and the phrase of the chat message of the group friend user that forms a reply pair with the chat message of the target user; In a case where the recommendation parameter is greater than a preset recommendation trigger threshold, determining the chat message to be determined as a recommended message; A prompt mark is displayed on the user interface, where the prompt mark is used to prompt the target user to view the recommended message.

2. The method according to claim 1, characterized in that The step of obtaining the recommended parameters of the chat message to be determined based on the phrase of the chat message to be determined, the phrase of the target user's chat message, and the phrase of the chat message of the group friend user that forms a reply pair with the chat message of the target user includes: Obtaining a word repetition rate between the phrase in the chat message to be determined and the phrase in the chat message of each target user, and a word repetition rate between the phrase in the chat message to be determined and the phrase in the chat message of the group friend user in the reply pair; The recommendation parameters of the chat message to be determined are calculated based on the word repetition rate, the weight coefficient corresponding to the phrase of the chat message to be determined, and the weight coefficient corresponding to the phrase of the chat message of the group friend user in the reply pair.

3. The method according to claim 1, characterized in that The step of obtaining the recommended parameters of the chat message to be determined based on the phrase of the chat message to be determined, the phrase of the target user's chat message, and the phrase of the chat message of the group friend user that forms a reply pair with the chat message of the target user includes: Using a pre-trained word vector model, the phrases in the target user's chat message and the phrases in the group member user's chat message in the reply pair are converted into a first word vector, and the phrases in the chat message to be determined are converted into a second word vector, wherein the word vector model is trained based on the phrases converted from the chat messages in the chat group and a vocabulary, and the vocabulary is constructed based on the phrases converted from the chat messages in the chat group; Obtaining a user interest vector of the target user according to the first word vector and a weight coefficient corresponding to the first word vector; The recommendation parameters of the chat message to be determined are obtained according to the similarity between the second word vector and the user interest vector.

4. The method according to claim 1, wherein Also includes: If the number of times that the chat message of the group friend user in the formed reply pair is not determined as the recommended message is greater than or equal to a first preset number threshold, lowering the recommendation trigger threshold; In the case that the number of times that the recommended message is not determined as a reply to the chat message of the group friend user is greater than or equal to the second preset number threshold, the recommendation trigger threshold is increased.

5. The method according to claim 1, wherein Also includes: The recommendation trigger threshold is adjusted according to the target user's feedback operation on the recommendation message.

6. The method according to claim 1, characterized in that The converting chat messages in the chat group into phrase data includes: Inputting the chat message into the large language model, and converting the chat message into phrase data through the large language model; or, Use the word segmentation algorithm to segment the chat message, and obtain phrase data based on the words obtained by word segmentation; In the case where the chat message includes a file, the phrase also includes file keywords obtained based on the file information of the file.

7. The method according to claim 2 or 3, characterized in that The phrase data also includes a phrase reply matching score, which is obtained based on the reply method between the chat message of the group user and the chat message of the target user. The reply method includes explicit reply method, implicit reply method and no reply method. The phrase reply matching score corresponding to the explicit reply method is higher than the phrase reply matching score corresponding to the implicit reply method, and the phrase reply matching score corresponding to the implicit reply method is higher than the phrase reply matching score corresponding to the no reply method. The phrase response matching score is positively correlated with the weight coefficient.

8. The method according to claim 1, characterized in that If the number of chat groups is more than two and it is determined to use the independent group chat recommendation function, the recommended message in the chat group is determined based on the chat messages in the chat group; If the number of chat groups is more than two and it is determined that the independent group chat recommendation function is not used, the recommended messages in each chat group are determined comprehensively through the chat messages in all chat groups.

9. The method according to claim 1, characterized in that Before obtaining the recommendation parameters of the chat message to be determined based on the phrase of the chat message to be determined, the phrase of the target user's chat message, and the phrase of the chat message of the group friend user that forms a reply pair with the chat message of the target user, the method further includes: Obtaining configuration parameters of a user terminal of the target user; When the user terminal performance represented by the configuration parameters is lower than a preset performance standard, a first determination method is used as the method for determining the recommended parameters, wherein the first determination method includes obtaining the recommended parameters based on a word repetition rate; When the user terminal performance represented by the configuration parameters is higher than or equal to the preset performance standard, the second determination method is used as the determination method of the recommended parameters, and the second determination method includes a method of obtaining the recommended parameters based on a word vector model.

10. The method according to claim 1, characterized in that Also includes: According to the setting input of the recommendation-related parameters by the user, executing the processing logic corresponding to the recommendation-related parameters after the setting input; In the case where the recommendation-related parameters include recommendation sensitivity, the processing logic corresponding to the recommendation sensitivity includes: adjusting the recommendation trigger threshold according to the recommendation sensitivity parameter; In the case where the recommendation-related parameters include a recommended message sending time, the processing logic corresponding to the recommended message sending time includes: obtaining the chat message to be determined from the chat messages of the group friend users that meet the recommended message sending time; In the case where the recommendation-related parameters include a sender of the recommendation message, the processing logic corresponding to the sender of the recommendation message includes: obtaining the chat message to be determined from the chat messages of the group friend user whose sender is the sender of the recommendation message; The recommendation-related parameters include an independent group chat recommendation parameter, and the processing logic corresponding to the independent group chat recommendation parameter includes: if the independent group chat recommendation parameter indicates that the independent group chat recommendation function is used, determining the recommended message in the chat group based on chat messages in the chat group; if the independent group chat recommendation parameter indicates that the independent group chat recommendation function is not used, comprehensively determining the recommended message in each chat group based on chat messages in all chat groups; The recommendation-related parameters include a recommendation parameter determination method, and the processing logic corresponding to the recommendation parameter determination method includes: if the recommendation parameter determination method represents a first determination method, then a method based on word repetition rate is used to obtain the recommendation parameter; if the recommendation parameter determination method represents a second determination method, then a method based on a word vector model is used to obtain the recommendation parameter.

11. A method for recommending group chat messages, characterized in that: include: A pre-processing module, configured to convert chat messages within a chat group into phrase data, wherein the phrase data includes phrases, each phrase includes keywords of a corresponding chat message, and the chat messages include chat messages of a target user and chat messages of group members; a recommendation parameter determination module, configured to obtain recommended parameters for the chat message to be determined based on a phrase in the chat message to be determined, a phrase in the chat message of the target user, and a phrase in the chat message of the group friend user that forms a reply pair with the chat message of the target user; a recommendation module, configured to determine that the chat message to be determined is a recommended message if the recommendation parameter is greater than a preset recommendation trigger threshold; The display module is used to display a prompt mark on the user interface, where the prompt mark is used to prompt the target user to view the recommended message.

12. A terminal device, characterized in that: include: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the method for recommending group chat messages according to any one of claims 1 to 10 is implemented.

13. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer program instructions, which, when executed by a processor, implement the method for recommending group chat messages according to any one of claims 1 to 10.

14. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the method for recommending group chat messages according to any one of claims 1 to 10.

Images