Comment information identification method and device, electronic equipment and storage medium

Abstract

The application discloses a comment information identification method and device and electronic equipment, and belongs to the technical field of electronics. The method comprises the following steps: obtaining target information of M comment information, the target information comprising at least one of the following: first information and second information, the first information being relationship chain information used for representing the relationship and cheating condition of N comment users, the second information being score information used for representing the credit of the N comment users, the N comment users being users publishing the M comment information, and M and N being positive integers; and based on the target information, performing identification processing on the M comment information, and determining false comment information from the M comment information.

Description

Technical Field

[0001] This application belongs to the field of electronic technology, specifically relating to a method, apparatus, electronic device, and storage medium for identifying comment information. Background Technology

[0002] Currently, users can get a preliminary understanding of an application by reading its reviews before downloading it, thus deciding whether to download it. However, if there is false information in the application's reviews, it can mislead users' choices.

[0003] In related technologies, electronic devices can identify and filter out false information in comment sections by analyzing the wording of comments. However, since electronic devices only identify false information through the wording of comments, if the comments posted by cheating users are very similar to those of normal users in terms of wording and frequency, the electronic devices may not be able to accurately identify the false information in the comment section. Thus, the accuracy of electronic devices in identifying false information in comment sections is relatively poor. Summary of the Invention

[0004] The purpose of this application is to provide a method, apparatus, electronic device, and storage medium for identifying comment information, which can solve the problem of poor accuracy in identifying false information in the comment section by electronic devices.

[0005] In a first aspect, embodiments of this application provide a method for identifying comment information. This method includes: acquiring target information for M comment messages, the target information including at least one of the following: first information and second information, the first information being relationship chain information characterizing the relationships and cheating situations of N comment users, and the second information being rating information characterizing the creditworthiness of the N comment users, where the N comment users are users who posted the M comment messages, and M and N are both positive integers; and based on the target information, processing the M comment messages to identify and determine false comment information from the M comment messages.

[0006] Secondly, embodiments of this application provide a comment information identification device, which includes an acquisition module and a processing module. The acquisition module is used to acquire target information from M comment messages. The target information includes at least one of the following: first information and second information. The first information is relationship chain information characterizing the relationships and cheating situations of N comment users, and the second information is rating information characterizing the creditworthiness of the N comment users. The N comment users are users who posted the M comment messages, and M and N are both positive integers. The processing module is used to identify and process the M comment messages based on the target information, and determine false comment information from the M comment messages.

[0007] Thirdly, embodiments of this application provide an electronic device including a processor and a memory, wherein the memory stores programs or instructions executable on the processor, and the programs or instructions, when executed by the processor, implement the steps of the method described in the first aspect.

[0008] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0009] Fifthly, embodiments of this application provide a chip, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.

[0010] In a sixth aspect, embodiments of this application provide a computer program product stored in a storage medium, which is executed by at least one processor to implement the method described in the first aspect.

[0011] In this embodiment, the electronic device can acquire at least one of a first piece of information and a second piece of information from M comments. The first piece of information is the relationship chain information of N commenting users corresponding to the M comments, and the second piece of information is the credit rating information of the N commenting users. Based on the first and second pieces of information, the electronic device identifies the M comments to determine fraudulent comments. In this solution, because the electronic device can identify the M comments using the relationship chain information and credit rating information of the users corresponding to the comments, it avoids the situation where the electronic device only recognizes the textual description of the comments and fails to identify more subtle fraudulent comments. Thus, the electronic device can better identify subtle fraudulent comments through multiple dimensions of information. Moreover, identifying comments through multiple dimensions improves the accuracy of the electronic device in identifying fraudulent information compared to the prior art which only recognizes the textual description of comments. Attached Figure Description

[0012] Figure 1 This is a flowchart of a comment information recognition method provided in an embodiment of this application;

[0013] Figure 2 This is a schematic diagram of an example of a target relationship diagram provided in an embodiment of this application;

[0014] Figure 3 This is one of the schematic diagrams illustrating an example of a comment information recognition method provided in this application embodiment;

[0015] Figure 4 This is a second schematic diagram illustrating an example of a comment information recognition method provided in this application embodiment;

[0016] Figure 5 This is a third example of a comment information recognition method provided in this application embodiment;

[0017] Figure 6 This is a schematic diagram of the structure of a comment information recognition device provided in an embodiment of this application;

[0018] Figure 7 This is one of the hardware structure diagrams of an electronic device provided in the embodiments of this application;

[0019] Figure 8 This is a second schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0022] The comment information recognition method provided in this application will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.

[0023] Currently, with the development of electronic devices, the number of applications on these devices is increasing. Before downloading an application, users can determine whether to download it based on reviews in the application's review section. However, driven by profit, some individuals or organizations create fake reviews to mislead users. This cheating not only makes it difficult for users to choose the applications they need but also constitutes unfair competition for other applications. In existing technologies, electronic devices analyze fake reviews only within the context of the review, leading to a biased assessment and an inability to identify hidden fake reviews. This is because some individuals or groups have evolved their cheating methods in the long-term struggle against platform anti-cheating measures, creating reviews that are similar to those of normal users in terms of wording and frequency. Therefore, the accuracy of electronic devices in identifying fake review information is poor.

[0024] In this embodiment, the electronic device can acquire at least one of the first and second information from M comment pieces. The first information is the relationship chain information of N commenting users corresponding to the M comment pieces, and the second information is the credit rating information of the N commenting users. Based on the first and second information, the electronic device identifies the M comment pieces to determine fraudulent comment information. In this solution, because the electronic device can identify the M comment pieces through the relationship chain information and credit rating information of the users corresponding to the comment pieces, it avoids the situation where the electronic device only recognizes the textual description of the comment information and fails to identify more subtle fraudulent comment information. Thus, the electronic device can better identify subtle fraudulent comment information through multiple dimensions of information. Moreover, identifying comment information through multiple dimensions improves the accuracy of the electronic device in identifying fraudulent information compared to the prior art which only recognizes the textual description of the comment.

[0025] The entity performing the comment information recognition provided in this application embodiment can be a comment information recognition device, which can be an electronic device or a functional module within an electronic device. The following description uses an electronic device as an example to illustrate the technical solution provided in this application embodiment.

[0026] This application provides a method for identifying comment information. Figure 1 A flowchart illustrating a comment information recognition method provided in an embodiment of this application is shown. Figure 1 As shown, the comment information recognition method provided in this application embodiment may include the following steps 201 and 202.

[0027] Step 201: The electronic device acquires the target information of M comments.

[0028] In this embodiment of the application, the target information includes at least one of the following: first information and second information. The first information is relationship chain information used to characterize the relationship and cheating situation of N commenting users. The second information is rating information used to characterize the creditworthiness of N commenting users. The N commenting users are users who publish M comment information, and M and N are both positive integers.

[0029] In this embodiment of the application, the electronic device can identify and process M comment information through first information and second information to determine false comment information from the M comment information.

[0030] Optionally, in this embodiment of the application, the above-mentioned M comment information can correspond one-to-one with N comment users; or, each of the N comment users can correspond to multiple comment information in the M comment information.

[0031] In this embodiment of the application, each of the above M comment information corresponds to the relationship chain information and credit rating information of the commenting user who posted the comment information.

[0032] It can be understood that the above M comments were sent by N users in the application's review area.

[0033] It should be noted that the aforementioned false review information refers to messages sent by cheating users that do not match the application's description.

[0034] For example, suppose the application mentioned above is a game application, but the comment information sent by the cheating user is information from an instant messaging application. In this case, the electronic device can determine that the information is false comment information; or, if the application functions performed by the application do not include the application functions in the comment information sent by the cheating user, the electronic device can also determine that the comment information sent by the cheating user is false comment information.

[0035] In this embodiment of the application, the relationship chain information of one of the N commenting users can be used to indicate other user information that is connected to a commenting user.

[0036] In this embodiment of the application, each credit score in the credit score information of the above N commenting users can be used to indicate whether each commenting user is a cheating user.

[0037] Optionally, in this embodiment of the application, the above-mentioned label information includes the first information; the above-mentioned step 201 can be specifically implemented by the following steps 201a to 201c.

[0038] Step 201a: The electronic device obtains the third and fourth information corresponding to N commenting users.

[0039] In this embodiment of the application, the third information includes at least one of the device identification information and network protocol IP address information used by N comment users when they post M comment messages; the fourth information is used to characterize users among the N comment users who have cheated in the past.

[0040] In this embodiment of the application, when each of the N commenting users sends a comment, the electronic device can record the third and fourth information of each user when sending the comment, thereby obtaining the N third and fourth information corresponding to the N commenting users.

[0041] For example, the fourth information mentioned above may include a target identifier, so that the electronic device can determine, based on the target identifier, a user among N commenting users who has cheated in the past.

[0042] For example, after a cheating user is identified in the past, the electronic device can mark the cheating user so that the electronic device can quickly identify the historical cheating user when identifying cheating users in the future.

[0043] Optionally, in the embodiments of this application, the target identifier can be any of the following: a number identifier, a special symbol identifier, a letter identifier, or a color identifier.

[0044] Optionally, in this embodiment of the application, the electronic device can obtain the third and fourth information corresponding to the comment information of N commenting users across the entire platform, and then construct a relationship diagram based on the third and fourth information of the N commenting users.

[0045] It should be noted that the above-mentioned platforms refer to different applications used by users to register for commenting.

[0046] In this embodiment, the electronic device can integrate the third and fourth information generated by the commenting user across the entire platform to construct a relationship graph. This makes the electronic device more comprehensive in determining whether a commenting user is a cheating user and the relationship network of cheating users.

[0047] Step 201b: The electronic device constructs a target relationship graph based on N commenting users and third-party information.

[0048] In this embodiment of the application, the target relationship graph is used to characterize the relationship between K comment users among N comment users whose third information satisfies the preset conditions, where K is a positive integer less than or equal to N.

[0049] Optionally, in this embodiment of the application, the above-mentioned preset conditions include at least one of the following: the device used by the commenting user when posting the comment information is the same device, and the IP address of the device used by the commenting user when posting the comment information is the same IP address.

[0050] For example, such as Figure 2 As shown, assume there are 4 commenting users, namely user 1 ( Figure 2 (represented by A in the text), User 2 ( Figure 2 (represented by B in Chinese), User 3 ( Figure 2 (represented by C) and User 4 ( Figure 2 (Represented by D in the diagram). User 1 and User 2 use the same device to comment on the application. The electronic device can construct an edge between User 1 and User 2. User 2 and User 3 use the same IP address when posting comments. The electronic device can construct an edge between User 2 and User 3. User 3 and User 4 use the same device to comment on the application. The electronic device can construct an edge between User 3 and User 4. User 4 and User 1 use the same IP address when posting comments. The electronic device can construct an edge between User 4 and User 1. Thus, we obtain the relationship graph 10, which represents the relationship between each commenting user and other users.

[0051] Optionally, in this embodiment of the application, after obtaining the target relationship graph, the electronic device can perform vectorization processing on the target relationship graph to obtain a vector-represented target relationship graph, so that the electronic device can generate first information based on the vectorized target relationship graph.

[0052] Step 201c: The electronic device generates the first information based on the target relationship diagram and the fourth information.

[0053] It is understandable that since the target relationship diagram above is determined based on K users who meet the preset conditions, the target relationship diagram can intuitively reflect other users who are related to each commenting user.

[0054] In this embodiment of the application, the electronic device can intuitively determine other users who are related to each commenting user through N relationship diagrams, and thus perform qualitative processing of commenting users based on the N relationship diagrams.

[0055] Optionally, in the embodiments of this application, step 201c can be implemented by steps 301 and 302 as described below.

[0056] Step 301: The electronic device determines the largest connected graph from the target relation graph.

[0057] In this embodiment of the application, the maximum connectivity graph is used to indicate the one-degree connection relationship between the nodes where the K commenting users are located.

[0058] In this embodiment of the application, the electronic device can determine the maximum connected graph from the target relation graph through a graph learning algorithm.

[0059] Step 302: The electronic device performs statistical processing on the user nodes connected to the nodes of the L cheating users to obtain the first relationship chain information, and performs statistical processing on the other user nodes in the maximum connected graph to obtain the second relationship chain information.

[0060] In this embodiment of the application, the above-mentioned L cheating users are users who have cheated in the past among the N commenting users, and the other users are users other than the L cheating users among the N commenting users, where L is a positive integer less than or equal to N.

[0061] In this embodiment of the application, the first information includes first relationship chain information and second relationship chain information; the first relationship chain information includes at least one of the following: the number of user nodes connected to the nodes where the L cheating users are located, and the proportion of the L cheating users in the largest connected graph; the second relationship chain information includes the number of user nodes connected to the nodes where each of the other users is located.

[0062] For example, an electronic device can determine the proportion of the L cheating users in the maximum connected graph based on the number of user nodes connected to the nodes where the L cheating users are located, and thus characterize the cheating user group based on the proportion.

[0063] It should be noted that, for a single user, the aforementioned connection refers to a direct connection with that single user. For example, combining... Figure 2 User 2 is a first-degree connection to User 1.

[0064] In this embodiment of the application, the electronic device can count the proportion of cheating users in the maximum connected graph and other users connected to cheating users, thereby identifying cheating groups and increasing the accuracy of the electronic device in identifying false comment information.

[0065] Optionally, in this embodiment of the application, the target information includes the second information; step 201 can be specifically implemented through steps 501 to 504 below.

[0066] Step 501: The electronic device acquires the behavioral information and fourth information of N commenting users.

[0067] In this embodiment of the application, the aforementioned behavioral information includes at least one of the following: registration information for each of the N commenting users, login information for each commenting user, and comment information for each commenting user.

[0068] In this embodiment of the application, the electronic device can obtain the behavior information of N commenting users across the entire platform, thereby determining the full-link behavior of the N commenting users and constructing relevant features corresponding to the behavior information.

[0069] For example, taking a commenting user as an example, the behavioral information obtained by the electronic device includes: the time information of the user's posting of the comment information on the entire platform, the IP address information of the user's posting of the comment information on the entire platform, the device information of the user when posting the comment information on the entire platform (such as device model and device name), and the frequency of the user's posting of the comment information on the entire platform.

[0070] In this embodiment, the electronic device can integrate behavioral data generated by commenting users across all platforms, thereby fully uncovering the characteristics of fraudulent comment-making groups. This allows the electronic device to more comprehensively determine whether a commenting user is a cheater and the network of relationships among cheaters.

[0071] Step 502: Based on the fourth information, the electronic device determines L cheating users among the N commenting users, where L is a positive integer less than or equal to N.

[0072] In this embodiment of the application, the aforementioned L cheating users are users who have cheated in the past among the N commenting users.

[0073] In this embodiment of the application, the fourth information mentioned above may include a target identifier, so that the electronic device can determine L cheating users from N commenting users based on the target identifier.

[0074] Step 503: The electronic device determines the first negative training sample based on the behavior information of the L cheating users and the L cheating users, and determines the first positive training sample based on the behavior information of other users and the other users.

[0075] In this embodiment of the application, other users are users other than L cheating users among N commenting users, where L is a positive integer less than or equal to N.

[0076] Optionally, in this embodiment of the application, after obtaining the fourth information, the electronic device can filter the cheating users indicated by the fourth information to obtain L cheating users.

[0077] Specifically, the electronic device can select L cheating users from the cheating users indicated by the fourth information based on a preset threshold.

[0078] For example, an electronic device can determine L cheating users as those whose cheating counts among the cheating users indicated by the fourth information exceed a preset threshold.

[0079] Step 504: The electronic device uses a tree model to identify and process the first negative training sample and the first positive training sample to obtain the second information.

[0080] For example, an electronic device can use the eXtreme Gradient Boosting (XGBoost) algorithm to identify the first negative training sample and the first positive training sample, thereby obtaining the second information.

[0081] In this embodiment of the application, the electronic device can determine whether each commenting user is a cheating user by scoring the creditworthiness of each commenting user.

[0082] For example, suppose user 1 has a credit score of 50, user 2 has a credit score of 60, and user 3 has a signal score of 80. If the preset threshold is 50, then user 1 is a cheating user.

[0083] In this embodiment, the electronic device can determine whether a user is cheating by analyzing the user's behavior throughout the entire process, thereby improving the accuracy of the electronic device in identifying cheating users.

[0084] Step 202: Based on the target information, the electronic device identifies and processes the M comment messages and determines the false comment messages from the M comment messages.

[0085] In this embodiment of the application, the electronic device can identify and process M comment information through information from multiple dimensions in order to determine false comment information from the M comment information.

[0086] Optionally, in the embodiments of this application, step 202 above can be specifically implemented by steps 202a to 202d below.

[0087] Step 202a: The electronic device acquires the feature information of M comments.

[0088] Optionally, in this embodiment of the application, the aforementioned feature information may include at least one of the following: semantic feature information of M comment information, text feature information of M comment information, and user-to-historical comment information corresponding to the M comment information.

[0089] Step 202b: The electronic device concatenates the relationship chain information, rating information, and feature information of the commenting user corresponding to each comment information to obtain M fifth pieces of information.

[0090] In this embodiment of the application, the electronic device can use M pieces of fifth information as training samples, and thus obtain the credit score of M comment information based on the M pieces of fifth information.

[0091] Step 202c: The electronic device determines the second negative training sample based on the L cheating users and the fifth information corresponding to the L cheating users, and determines the second positive training sample based on the fifth information corresponding to other users.

[0092] In this embodiment of the application, the above-mentioned L cheating users are users who have cheated in the past among the N commenting users, and the other users are users other than the L cheating users among the N commenting users, where L is a positive integer less than or equal to N.

[0093] Step 202d: The electronic device uses a tree model to identify and process the second negative training sample and the second positive training sample to obtain a credit score for each of the M comment information. The credit score is used to characterize whether the comment information is fake.

[0094] In this embodiment of the application, the electronic device can use xgboost to identify and process the second negative training sample and the second positive training sample, thereby obtaining a credit score for each of the M comment information.

[0095] This application provides a method for identifying comment information. An electronic device can acquire at least one of a first piece of information and a second piece of information from M comment pieces. The first piece of information is the relationship chain information of N commenting users corresponding to the M comment pieces, and the second piece of information is the credit rating information of the N commenting users. Based on the first and second pieces of information, the electronic device identifies the M comment pieces to determine fraudulent comment information. In this solution, because the electronic device can identify the M comment pieces through the relationship chain information and credit rating information of the users corresponding to the comment pieces, it avoids the situation where the electronic device only recognizes the textual description of the comment information and fails to identify more subtle fraudulent comment information. Thus, the electronic device can better identify subtle fraudulent comment information through multiple dimensions of information. Moreover, identifying comment information through multiple dimensions improves the accuracy of the electronic device in identifying fraudulent information compared to the prior art that only recognizes the textual description of the comment.

[0096] For example, such as Figure 3 As shown below, the comment information recognition method provided in this application embodiment will be further explained through specific examples. The comment information recognition method provided in this application embodiment can be implemented through the following steps 21 to 24.

[0097] Step 21: Electronic devices acquire behavioral data of commenting users across the entire platform.

[0098] In this embodiment of the application, the electronic device can construct a relationship graph by acquiring the behavior data of commenting users across the entire platform.

[0099] Step 22: Electronic devices identify users who comment as cheating groups.

[0100] In this embodiment of the application, the electronic device can use graph algorithms to mine groups in the relationship graph, thereby enabling quantitative analysis of the group characteristics of commenting users.

[0101] Step 23: The electronic device calculates the comprehensive behavioral credit score of the reviewing user.

[0102] In this embodiment of the application, the electronic device can collect historical malicious behavior samples as labels, combine them with behavioral features, and use a tree model to model the creditworthiness of the commenting user to obtain the comprehensive behavioral credit score of the commenting user.

[0103] For example, such as Figure 4 As shown, after acquiring the behavioral data of commenting users across the entire platform, the electronic device can construct a relationship graph 11 and behavioral feature information 12 based on the behavioral data. This behavioral feature information 12 includes time distribution features, IP address clustering features, device usage features, login count statistics, and active scenario count statistics. The electronic device can perform graph computation on the relationship graph 11 and obtain historical cheating user information from the black seed database through a preset interface, thereby obtaining cheating group information 13. This cheating group information 13 includes the size of the community, the proportion of black seeds in the community, the number of nodes first-order connected to the black seed node, and the proportion of other black seeds first-order connected to the black seed node. Simultaneously, the electronic device can perform tagging and credit modeling on the behavioral feature information 12 and the black seed feature information to obtain the comprehensive behavioral credit score of the commenting user.

[0104] It should be noted that the aforementioned black seed characteristic information was obtained by the electronic device through a black seed database.

[0105] Step 24: The electronic device identifies fake comments posted by the commenting user.

[0106] In this embodiment, the electronic device can collect historical fake comments as tags, and combine them with comment content features, environmental features, commenter group features, and commenter behavior credit scores to use logistic regression to model the authenticity of comments and identify fake comments.

[0107] For example, such as Figure 5 As shown, the electronic device can concatenate the comprehensive behavioral credit score of the commenting user corresponding to each comment information, the current comment characteristics of the comment information, and the cheating gang information 13 of the commenting user corresponding to each comment information to obtain the first information corresponding to each comment information. Then, the historical fake comments are tagged and processed, and a comment credibility model 14 is established based on the tagged historical fake comments and the first information. Then, fake comments are identified based on the comment credibility model 14.

[0108] It should be noted that the comment information recognition method provided in this application can be executed by a comment information recognition device, an electronic device, or a functional module or entity within an electronic device. This application uses a comment information recognition device executing the comment information recognition method as an example to illustrate the comment information recognition device provided in this application.

[0109] Figure 6 A schematic diagram of a possible structure of the comment information recognition device involved in an embodiment of this application is shown. For example... Figure 6 As shown, the comment information recognition device 70 may include an acquisition module 71 and a processing module 72.

[0110] The acquisition module 71 is used to acquire target information from M comment messages. The target information includes at least one of the following: first information and second information. The first information is a relationship chain information representing the relationship and cheating situation of N comment users. The second information is a rating information representing the creditworthiness of the N comment users. The N comment users are users who posted the M comment messages, and M and N are both positive integers. The processing module 72 is used to identify and process the M comment messages based on the target information, and determine the fraudulent comment messages from the M comment messages.

[0111] In one possible implementation, the target information includes first information; the acquisition module 71 is specifically used to acquire third and fourth information corresponding to N comment users; wherein, the third information includes at least one of the device identification information and network protocol IP address information used by the N comment users when they publish M comment messages; the fourth information is used to characterize users among the N comment users who have cheated in the past; and based on the N comment users and the third information, a target relationship graph is constructed, which is used to characterize the relationship between K comment users among the N comment users whose third information satisfies the preset conditions, where K is a positive integer less than or equal to N; and the first information is generated based on the target relationship graph and the fourth information.

[0112] In one possible implementation, the aforementioned preset conditions include at least one of the following: the device used by the commenting user when posting the comment information is the same device, and the IP address of the device used by the commenting user when posting the comment information is the same IP address.

[0113] In one possible implementation, the acquisition module 71 is specifically used to determine the maximum connected graph from the target relationship graph, the maximum connected graph indicating the one-degree connection relationship between the nodes where the K commenting users are located; and to perform statistical processing on the user nodes connected to the nodes where the L cheating users are located in the maximum connected graph to obtain first relationship chain information, and to perform statistical processing on the other user nodes in the maximum connected graph to obtain second relationship chain information, where the L cheating users are users who have cheated in the past among the N commenting users, and the other users are users other than the L cheating users among the N commenting users, and L is a positive integer less than or equal to N; wherein, the first information includes the first relationship chain information and the second relationship chain information; the first relationship chain information includes at least one of the following: the number of user nodes connected to the nodes where the L cheating users are located, and the proportion of the L cheating users in the maximum connected graph; the second relationship chain information includes the number of user nodes connected to the nodes where each of the other users is located.

[0114] In one possible implementation, the target information includes second information; the acquisition module 71 is specifically used to acquire behavioral information and fourth information of N commenting users; wherein, the behavioral information includes at least one of the following: registration information, login information, and comment information of each commenting user among the N commenting users; and based on the fourth information, to determine L cheating users among the N commenting users, where L are users who have cheated in the past, and L is a positive integer less than or equal to N; and based on the L cheating users and their corresponding behavioral information, to determine a first negative training sample, and based on the behavioral information of other users and their corresponding behavioral information, to determine a first positive training sample, where other users are users other than the L cheating users among the N commenting users; and through a tree model, to identify and process the first negative training sample and the first positive training sample to obtain the second information.

[0115] In one possible implementation, the processing module 72 is specifically used to obtain feature information of M comment information; and to concatenate the relationship chain information, rating information and feature information of the commenting user corresponding to each comment information to obtain M pieces of fifth information; and to determine the second negative training sample based on L cheating users and the fifth information corresponding to L cheating users, and to determine the second positive training sample based on the fifth information corresponding to other users and other users, where L cheating users are users who have cheated in the past among N commenting users, and other users are users other than L cheating users among N commenting users, and L is a positive integer less than or equal to N; and to perform identification processing on the second negative training sample and the second positive training sample through a tree model to obtain the credit score of each comment information in the M comment information, and the credit score is used to characterize whether the comment information is fake comment information.

[0116] This application provides a comment information recognition device. Because the device can identify M comments using the user's relationship chain information and credit rating information corresponding to the comment information, it avoids the problem of only recognizing the textual description of the comments and failing to identify more subtle false comments. Thus, the device can better identify subtle false comments through multiple dimensions of information. Furthermore, by recognizing comment information from multiple dimensions, compared to existing technologies that only recognize the textual description of comments, the accuracy of electronic devices in identifying false information is improved.

[0117] The comment information recognition device in this application embodiment can be a device, or a component, integrated circuit, or chip in an electronic device. The device can be a mobile electronic device or a non-mobile electronic device. For example, a mobile electronic device can be a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM, or self-service machine, etc. This application embodiment does not specifically limit the device.

[0118] The comment information recognition device in this application embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this application embodiment does not specifically limit it.

[0119] The comment information recognition device provided in this application embodiment can achieve... Figures 1 to 5 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0120] Optionally, such as Figure 7As shown, this application embodiment also provides an electronic device 90, including a processor 91 and a memory 92. The memory 92 stores a program or instructions that can run on the processor 91. When the program or instructions are executed by the processor 91, they implement the various steps of the above-described comment information recognition device method embodiment and can achieve the same technical effect. To avoid repetition, they will not be described again here.

[0121] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0122] Figure 8 A schematic diagram of the hardware structure of an electronic device to implement an embodiment of this application.

[0123] The electronic device 100 includes, but is not limited to, components such as: radio frequency unit 101, network module 102, audio output unit 103, input unit 104, sensor 105, display unit 106, user input unit 107, interface unit 108, memory 109, and processor 110.

[0124] Those skilled in the art will understand that the electronic device 100 may also include a power supply (such as a battery) for supplying power to various components. The power supply may be logically connected to the processor 110 through a power management system, thereby enabling functions such as managing charging, discharging, and power consumption through the power management system. Figure 8 The electronic device structure shown does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements, which will not be elaborated here.

[0125] The processor 110 is used to acquire target information from M comments, the target information including at least one of the following: first information and second information, the first information being relationship chain information used to characterize the relationship and cheating situation of N comment users, and the second information being rating information used to characterize the creditworthiness of N comment users, where N comment users are users who posted M comments, and M and N are both positive integers; and based on the target information, to identify and process the M comments to determine fake comments from the M comments.

[0126] This application provides an electronic device that can identify M comments by using the user's relationship chain information and credit rating information corresponding to the comment information. This avoids the problem of the electronic device only recognizing the textual description of the comment information and failing to identify more subtle false comment information. In this way, the electronic device can better identify subtle false comment information through multiple dimensions of information. Moreover, by recognizing the comment information through multiple dimensions, the accuracy of the electronic device in identifying false information is improved compared to the prior art which only recognizes the textual description of the comment.

[0127] Optionally, in this embodiment of the application, the processor 110 is specifically used to obtain third information and fourth information corresponding to N commenting users; wherein, the third information includes at least one of the device identification information and network protocol IP address information used by the N commenting users when they publish M commenting information; the fourth information is used to characterize users among the N commenting users who have cheated in the past; and constructs a target relationship graph based on the N commenting users and the third information, the target relationship graph being used to characterize the relationship between K commenting users among the N commenting users whose third information satisfies a preset condition, where K is a positive integer less than or equal to N; and generates first information based on the target relationship graph and the fourth information.

[0128] Optionally, in this embodiment of the application, the processor 110 is specifically used to determine the maximum connected graph from the target relationship graph, the maximum connected graph being used to indicate the one-degree connection relationship between the nodes where the K commenting users are located; and to perform statistical processing on the user nodes connected to the nodes where the L cheating users are located in the maximum connected graph to obtain first relationship chain information, and to perform statistical processing on the other user nodes in the maximum connected graph to obtain second relationship chain information, wherein the L cheating users are users who have cheated in the past among the N commenting users, and the other users are users other than the L cheating users among the N commenting users, and L is a positive integer less than or equal to N; wherein, the first information includes the first relationship chain information and the second relationship chain information; the first relationship chain information includes at least one of the following: the number of user nodes connected to the nodes where the L cheating users are located, and the proportion of the L cheating users in the maximum connected graph; the second relationship chain information includes the number of user nodes connected to the nodes where each of the other users is located.

[0129] Optionally, in this embodiment of the application, the processor 110 is specifically used to obtain behavioral information and fourth information of N commenting users; wherein, the behavioral information includes at least one of the following: registration information, login information, and comment information of each commenting user among the N commenting users; and based on the fourth information, to determine L cheating users among the N commenting users, where L cheating users are users who have cheated in the past among the N commenting users, and L is a positive integer less than or equal to N; and based on the L cheating users and the behavioral information corresponding to the L cheating users, to determine a first negative training sample, and based on the behavioral information corresponding to other users and other users, to determine a first positive training sample, where L cheating users are users who have cheated in the past among the N commenting users, and other users are users other than the L cheating users among the N commenting users, and L is a positive integer less than or equal to N; and to perform identification processing on the first negative training sample and the first positive training sample through a tree model to obtain second information.

[0130] Optionally, in this embodiment, the processor 110 is specifically used to obtain feature information of M comment information; concatenate the relationship chain information, rating information, and feature information of the comment user corresponding to each comment information to obtain M pieces of fifth information; determine the second negative training sample based on L cheating users and the fifth information corresponding to L cheating users, and determine the second positive training sample based on the fifth information corresponding to other users and other users, where L cheating users are users who have cheated in the past among N comment users, and other users are users other than L cheating users among N comment users, and L is a positive integer less than or equal to N; and perform identification processing on the second negative training sample and the second positive training sample through a tree model to obtain the credit score of each comment information in the M comment information, where the credit score is used to characterize whether the comment information is fake comment information.

[0131] The electronic device provided in this application embodiment can implement the various processes implemented in the above method embodiments and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0132] The beneficial effects of the various implementation methods in this embodiment can be found in the beneficial effects of the corresponding implementation methods in the above method embodiments. To avoid repetition, they will not be repeated here.

[0133] It should be understood that, in this embodiment, the input unit 104 may include a graphics processing unit (GPU) 1041 and a microphone 1042. The GPU 1041 processes image data of still images or videos obtained by an image capture device (such as a camera) in video capture mode or image capture mode. The display unit 106 may include a display panel 1061, which may be configured in the form of a liquid crystal display, an organic light-emitting diode, or the like. The user input unit 107 includes at least one of a touch panel 1071 and other input devices 1072. The touch panel 1071 is also called a touch screen. The touch panel 1071 may include a touch detection device and a touch controller. Other input devices 1072 may include, but are not limited to, a physical keyboard, function keys (such as volume control buttons, power buttons, etc.), a trackball, a mouse, and a joystick, which will not be described in detail here.

[0134] The memory 109 can be used to store software programs and various data. The memory 109 may primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, the memory 109 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DRRAM). The memory 109 in the embodiments of this application includes, but is not limited to, these and any other suitable types of memory.

[0135] Processor 110 may include one or more processing units; optionally, processor 110 integrates an application processor and a modem processor, wherein the application processor mainly handles operations involving the operating system, user interface, and applications, and the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into processor 110.

[0136] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0137] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0138] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above method embodiments and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0139] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0140] This application provides a computer program product stored in a storage medium. The program product is executed by at least one processor to implement the various processes of the comment information recognition method embodiment described above, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0141] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0142] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0143] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A comment information identifying method characterized by comprising: The method comprises: obtaining target information of M pieces of comment information, the target information comprising: first information and second information, the first information being relationship chain information for representing relationships and cheating conditions of N comment users, the second information being score information for representing credit degrees of the N comment users, the N comment users being users publishing the M pieces of comment information, M and N being positive integers; based on the target information, performing identification processing on the M pieces of comment information to determine false comment information from the M pieces of comment information; wherein the step of obtaining the first information comprises: obtaining third information and fourth information corresponding to the N comment users; wherein the third information comprises at least one of identification information of a device used by the N comment users to publish the M pieces of comment information and network protocol IP address information, and the fourth information is used to represent users that have historically cheated among the N comment users; constructing a target relationship graph according to the N comment users and the third information, the target relationship graph being used to represent relationships between K comment users that satisfy a preset condition in the third information among the N comment users, K being a positive integer less than or equal to N; generating the first information according to the target relationship graph and the fourth information.

2. The method of claim 1, wherein, The preset condition comprises at least one of the following: the device used by the comment users to publish the comment information is the same device, and the IP address of the device used by the comment users to publish the comment information is the same IP address.

3. The method according to claim 1 or 2, characterized in that, The generating of the first information according to the target relationship graph and the fourth information comprises: determining a maximum connected graph from the target relationship graph, the maximum connected graph being used to indicate a one-degree connection relationship between nodes where the K comment users are located; statistically processing user nodes connected to nodes where L cheating users are located in the maximum connected graph to obtain first relationship chain information, and statistically processing other user nodes in the maximum connected graph to obtain second relationship chain information, the L cheating users being users that have historically cheated among the N comment users, the other users being users other than the L cheating users among the N comment users, L being a positive integer less than or equal to N; wherein the first information comprises the first relationship chain information and the second relationship chain information; the first relationship chain information comprises at least one of the following: a number of user nodes connected to nodes where the L cheating users are located, and a proportion of the L cheating users in the maximum connected graph; and the second relationship chain information comprises a number of user nodes connected to nodes where each user among the other users is located.

4. The method of claim 1, wherein, The target information comprises the second information; and the obtaining of the target information of the M pieces of comment information comprises: obtaining behavior information and fourth information of the N comment users; wherein the behavior information comprises at least one of the following: registration information corresponding to each comment user among the N comment users, login information corresponding to the each comment user, and comment information corresponding to the each comment user. According to the fourth information, L cheating users in the N comment users are determined, the L cheating users are users that have cheated historically in the N comment users, and L is a positive integer less than or equal to N; According to the L cheating users and behavior information corresponding to the L cheating users, a first negative training sample is determined, and according to other users and behavior information corresponding to the other users, a first positive training sample is determined, the other users being users other than the L cheating users in the N comment users; The first negative training sample and the first positive training sample are identified by a tree model to obtain the second information.

5. The method of claim 1, wherein, The identification processing of the M comment information based on the target information includes: Obtaining feature information of the M comment information; Respectively, the relationship chain information, the score information and the feature information of each comment information corresponding to the comment user are spliced to obtain M fifth information; According to the L cheating users and the fifth information corresponding to the L cheating users, a second negative training sample is determined, and according to other users and the fifth information corresponding to the other users, a second positive training sample is determined, the L cheating users being users that have cheated historically in the N comment users, the other users being users other than the L cheating users in the N comment users, and L being a positive integer less than or equal to N; The second negative training sample and the second positive training sample are identified by a tree model to obtain a credit score of each comment information in the M comment information, and the credit score is used to represent whether the comment information is a false comment information.

6. A comment information identifying apparatus characterized by comprising: The comment information identification device includes an acquisition module and a processing module. The acquisition module is used to acquire target information of M comment information, the target information including first information and second information, the first information being relationship chain information used to represent relationships and cheating conditions of N comment users, the second information being score information used to represent credit degrees of the N comment users, the N comment users being users publishing the M comment information, and M and N being positive integers; The processing module is used to identify the M comment information based on the target information to determine false comment information from the M comment information. The acquisition module is specifically used to acquire third information and fourth information corresponding to the N comment users, the third information including at least one of identification information and network protocol IP address information of devices used by the N comment users to publish the M comment information, and the fourth information being used to represent users that have cheated historically in the N comment users; a target relationship graph is constructed according to the N comment users and the third information, the target relationship graph being used to represent relationships between K comment users in the N comment users whose third information satisfies a preset condition, K being a positive integer less than or equal to N; and the first information is generated according to the target relationship graph and the fourth information.

7. An electronic device, comprising: An article of manufacture includes a processor, a memory, and a program or instructions stored on the memory and executable on the processor, the program or instructions, when executed by the processor, implement the steps of the comment information recognition method of any one of claims 1 to 5.

8. A readable storage medium, characterized by, A program or instructions are stored on the readable storage medium, the program or instructions, when executed by the processor, implement the steps of the comment information recognition method of any one of claims 1 to 5.

Images