Illegal gang account identification method and device, computer equipment and readable storage medium

Abstract

The invention relates to the technical field of illegal account identification, and provides an illegal gang account identification method and device, computer equipment and a readable storage medium. The method comprises the steps of determining a target account set of a current time window according to text content published in the current time window of a non-first time window; obtaining a target account set screened by the current time window according to at least one of account consistency and text content similarity; according to the similarity of the text content, clustering the text content published by the screened target account set in the current time window to obtain a suspected violation gang library of the current time window; and determining the illegal gang account according to at least one of the times that the same account is classified into the suspected illegal gang libraries corresponding to the plurality of time windows and the times that the same text content is classified into the suspected illegal gang libraries corresponding to the plurality of time windows. By adopting the method, the illegal gang account identification accuracy is improved.

Description

Technical Field

[0001] The present application relates to the technical field of illegal account identification, and in particular to an illegal group account identification method, device, computer equipment, computer-readable storage medium and computer program product. Background Art

[0002] In order to maintain a healthy community environment and encourage the creation of authentic and original content, content creation platforms will identify various group violations, such as group violations of sending a large number of similar texts in a short period of time to attract traffic.

[0003] Currently, we can identify a small number of illegal group accounts based on policy configurations such as multiple accounts on the same device, multiple devices on the same account, and clustering devices corresponding to accounts with similar illegal behaviors during the same period. However, the current illegal group account identification methods have problems with lag and easy omissions in identifying short-term and high-frequency group illegal behaviors, resulting in a low accuracy rate in illegal group account identification. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for identifying accounts of illegal gangs in response to the above technical problems.

[0005] In the first aspect, the present application provides a method for identifying illegal gang accounts, including:

[0006] If the current time window is not the first time window, determine the target account set corresponding to the current time window according to the text content published in the current time window; wherein each text content is published by a corresponding account;

[0007] Determine, based on at least one of account consistency and text content similarity, whether to classify at least some of the accounts in the target account set and the text content published by the some of the accounts into a suspected illegal group library corresponding to a previous time window, so as to obtain a filtered target account set corresponding to the current time window;

[0008] Clustering the text content posted by the screened target account set within the current time window according to the similarity of the text content, to obtain a suspected illegal group library corresponding to the current time window;

[0009] Determine whether the corresponding account belongs to the illegal group account based on at least one of the number of times the same account is included in the suspected illegal group library corresponding to multiple time windows and the number of times the same text content is included in the suspected illegal group library corresponding to multiple time windows.

[0010] In one embodiment, determining the target account set corresponding to the current time window according to the text content published in the current time window includes:

[0011] According to the text content published in the current time window, determine the account that published the text content in the current time window;

[0012] Determining an initial account set corresponding to the current time window according to the accounts that publish text content within the current time window;

[0013] Get illegal accounts and whitelist accounts;

[0014] The illegal accounts and the whitelist accounts are removed from the initial account set to obtain a target account set corresponding to the current time window.

[0015] In one embodiment, after obtaining the suspected illegal group database corresponding to the current time window, the method further includes:

[0016] Get the text content posted by the violating account in the current time window;

[0017] Based on the similarity between the text content posted by the violating account in the current time window and the suspected violating text content posted by the suspected violating group account in the current time window, determine whether to include the violating account and the content posted by the violating account into the suspected violating group library corresponding to the current time window.

[0018] In one embodiment, determining whether to classify at least some of the accounts in the target account set and the text contents published by the some of the accounts into a suspected illegal group library corresponding to a previous time window based on at least one of account consistency and text content similarity to obtain a filtered target account set corresponding to the current time window includes:

[0019] Comparing the consistency of each account in the target account set with each suspected illegal group account in the suspected illegal group database corresponding to the previous time window in turn;

[0020] When the accounts in the target account set are consistent with the suspected illegal gang accounts in the suspected illegal gang library corresponding to the previous time window, the accounts in the target account set and the text content published by the accounts are classified into the suspected illegal gang library corresponding to the previous time window, and the accounts in the target account set are removed from the target account set to obtain the filtered target account set corresponding to the current time window;

[0021] When the accounts in the target account set are inconsistent with the suspected illegal group accounts in the suspected illegal group library corresponding to the previous time window, determine whether to include the accounts in the target account set and the text contents posted by some of the accounts into the suspected illegal group library corresponding to the previous time window based on the similarity of the text contents, and remove the accounts in the target account set from the target account set to obtain the filtered target account set corresponding to the current time window.

[0022] In one embodiment, determining whether to classify the accounts in the target account set and the text contents posted by the partial accounts into the suspected illegal group library corresponding to the previous time window according to the text content similarity includes:

[0023] Using the account number in the target account set as the account number to be queried;

[0024] From each suspected illegal group in the suspected illegal group database corresponding to the previous time window, select at least one suspected illegal text content posted by the account of the suspected illegal group in the previous time window as the representative text content of each suspected illegal group;

[0025] Determine similar representative text content corresponding to the account to be queried according to the similarity between the text content published by the account to be queried in the current time window and each representative text content;

[0026] The account to be queried and the text content published by the account to be queried are classified into the suspected illegal group corresponding to the similar representative text content.

[0027] In one embodiment, determining similar representative text content corresponding to the account to be queried according to the similarity between the text content published by the account to be queried in the current time window and each representative text content includes:

[0028] Obtaining a hash value of the account to be queried in the current window according to the text content published by the account to be queried in the current time window;

[0029] Get the hash value of each representative text content;

[0030] According to the closeness between the hash values, among the accounts to which the representative text contents belong, the accounts that are classified into the same set as the account to be queried are determined to obtain a target account subset;

[0031] Based on the similarity between the text content published by the account to be queried in the current time window and the representative text content of other accounts in the target account subset, similar representative text content corresponding to the account to be queried is determined.

[0032] In one embodiment, clustering the text content posted by the screened target account set within the current time window according to the similarity of the text content to obtain a suspected illegal group library corresponding to the current time window includes:

[0033] Obtaining a hash value of each text content published by the filtered target account set within the current time window;

[0034] According to the closeness between the hash values, the text content corresponding to the filtered target account set in the current time window is divided to obtain a plurality of target text content subsets;

[0035] For any target text content subset, clustering the text contents in the target text content subset according to the similarity of the text contents in the target text content subset to obtain multiple clusters;

[0036] When the number of text contents included in the cluster is greater than or equal to the number of text contents included in the minimum cluster, the text contents included in the cluster are used as suspected illegal text contents to obtain a suspected illegal text content set corresponding to the current time window;

[0037] According to the suspected illegal text content set corresponding to the current time window, obtain the suspected illegal group account set corresponding to the current time window;

[0038] According to the suspected illegal text content set and the suspected illegal group account set corresponding to the current time window, a suspected illegal group library corresponding to the current time window is obtained.

[0039] In the second aspect, the present application also provides a device for identifying illegal gang accounts, including:

[0040] A target account set acquisition module, used to determine the target account set corresponding to the current time window according to the text content published in the current time window if the current time window is not the first time window; wherein each text content is published by a corresponding account;

[0041] A target account set screening module, used to determine whether to classify at least some of the accounts in the target account set and the text contents published by the accounts into a suspected illegal gang library corresponding to a previous time window, based on at least one of account consistency and text content similarity, so as to obtain a screened target account set corresponding to the current time window;

[0042] A suspected illegal group library acquisition module is used to cluster the text content posted by the screened target account set within the current time window according to the similarity of the text content, and obtain the suspected illegal group library corresponding to the current time window;

[0043] The illegal group account determination module is used to determine whether the corresponding account belongs to the illegal group based on at least one of the number of times the same account is included in the suspected illegal group library corresponding to multiple time windows and the number of times the same text content is included in the suspected illegal group library corresponding to multiple time windows.

[0044] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the above method.

[0045] In a fourth aspect, the present application further provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and the computer program is executed by a processor to execute the above method.

[0046] In a fifth aspect, the present application further provides a computer program product, wherein the computer program product comprises a computer program, and the computer program is executed by a processor to execute the above method.

[0047] The above-mentioned method, device, computer equipment, computer-readable storage medium and computer program product for identifying accounts of illegal groups, if the current time window is not the first time window, determine the target account set corresponding to the current time window based on the text content published in the current time window; wherein each text content is published by a corresponding account; determine whether to include at least part of the accounts in the target account set and the text content published by some of the accounts into the suspected illegal group library corresponding to the previous time window based on at least one of account consistency and text content similarity, so as to obtain a filtered target account set corresponding to the current time window; cluster the text content published by the filtered target account set in the current time window based on the similarity of the text content, so as to obtain a suspected illegal group library corresponding to the current time window; determine whether the corresponding account belongs to an illegal group account based on at least one of the number of times the same account is included in the suspected illegal group library corresponding to multiple time windows and the number of times the same text content is included in the suspected illegal group library corresponding to multiple time windows. In the solution provided by the present application, the length of the time window can be set to be shorter. In this case, the filtered target account set corresponding to the current time window can be clustered according to the similarity of the text content published in the time window with a shorter time length, so as to obtain a library of suspected illegal gangs corresponding to the current time window. This can identify short-term and high-frequency group violations, alleviate the lag in identifying group violations, and improve the accuracy of identifying illegal gang accounts. In addition, the present application determines whether to include at least part of the accounts in the target account set into the library of suspected illegal gangs corresponding to the previous time window based on at least one of the account consistency and the text content similarity. This can avoid the omission of some accounts that actually belong to illegal gangs due to the division of time windows, and further improve the accuracy of identifying illegal gang accounts. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 A diagram of an application environment of a method for identifying illegal group accounts in an embodiment;

[0050] Figure 2 A flowchart of a method for identifying illegal group accounts in one embodiment;

[0051] Figure 3 A schematic diagram of a process of obtaining a target account set corresponding to a current time window in an embodiment;

[0052] Figure 4 It is a structural block diagram of a device for identifying illegal group accounts in one embodiment;

[0053] Figure 5 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0055] The present application embodiment provides a method for identifying illegal gang accounts, which can be executed by a computer device, such as Figure 1 As shown, the computer device can obtain the text content published in the current time window, and then determine whether the corresponding account belongs to the account of the illegal group. It can be understood that the computer device can be implemented through a server, or through a terminal, or through an interactive system between a terminal and a server. In this embodiment, the method includes Figure 2 The steps shown are:

[0056] Step S201, if the current time window is not the first time window, determine the target account set corresponding to the current time window according to the account of the text content published in the current time window; wherein each text content is published by a corresponding account.

[0057] Accounts that publish text content may include accounts that modify their account nicknames and accounts that create content; for accounts that modify their account nicknames, the modified account nickname is text content, and for accounts that create content, the created content is text content.

[0058] The length of the time window can be set to be shorter, for example, the length of the time window can be set to 1 minute or 5 minutes.

[0059] The text content published on designated channels such as the platform's public screen, private chat, square, room name, room content, etc. can be integrated by channel to determine the account that published the text content within a certain time period. The account that published the text content within the current time window can be determined based on the current time window.

[0060] If the current time window is not the first time window, the account that publishes the text content in the current time window is determined based on the text content published in the current time window, and the account that publishes the text content in the current time window is determined as the target account corresponding to the current time window; based on the target account corresponding to the current time window, the target account set corresponding to the current time window is obtained.

[0061] Step S202, based on at least one of account consistency and text content similarity, determine whether to include at least some of the accounts in the target account set and the text content published by some of the accounts into the suspected illegal group library corresponding to the previous time window to obtain a filtered target account set corresponding to the current time window.

[0062] The suspected illegal gang database includes a suspected illegal text content set and a suspected illegal gang account set, wherein the suspected illegal text content corresponds to the suspected illegal gang account one by one.

[0063] When the target account set corresponding to the current time window is screened, it may be screened based on at least one of account consistency and text content similarity.

[0064] If the target account set corresponding to the current time window is filtered based on account consistency, the following steps can be performed specifically:

[0065] For the time window before the current time window, a suspected illegal gang database corresponding to the previous time window can be obtained; based on the suspected illegal gang database corresponding to the previous time window, the suspected illegal gang account corresponding to the previous time window is obtained.

[0066] Taking account A in the target account set corresponding to the current time window as an example, the following steps can be performed: compare the consistency of account A with each suspected illegal group account of the suspected illegal group corresponding to the previous time window; if account A is consistent with a suspected illegal group account corresponding to the previous time window, then account A and the text content published by account A are classified into the suspected illegal group to which the suspected illegal group account belongs in the previous time window, and account A is removed from the target account set corresponding to the current time window. In this way, each account in the target account set corresponding to the current time window is processed to obtain the filtered target account set corresponding to the current time window.

[0067] If the target account set corresponding to the current time window is filtered based on the text content similarity, the following steps can be performed specifically:

[0068] For the time window before the current time window, a suspected illegal group library corresponding to the previous time window can be obtained; based on the suspected illegal group library corresponding to the previous time window, the suspected illegal text content corresponding to the previous time window is obtained.

[0069] Taking account B in the target account set corresponding to the current time window as an example, the following steps can be performed: compare the similarity between the text content of account B and each suspected illegal text content of the suspected illegal group corresponding to the previous time window; if the similarity between the text content of account B and a suspected illegal text content corresponding to the previous time window meets the similarity threshold, account B and the text content published by account B are classified into the suspected illegal group to which the suspected illegal text content belongs in the previous time window, and account B is removed from the target account set corresponding to the current time window. In this way, each account in the target account set corresponding to the current time window is processed to obtain the filtered target account set corresponding to the current time window.

[0070] If the target account set corresponding to the current time window is filtered based on account consistency and text content similarity, the following steps can be performed specifically:

[0071] For the time window before the current time window, the suspected illegal gang database corresponding to the previous time window can be obtained; according to the suspected illegal gang database corresponding to the previous time window, the suspected illegal gang account corresponding to the previous time window and the suspected illegal text content corresponding to the above suspected illegal gang are obtained.

[0072] Taking account C in the target account set corresponding to the current time window as an example, the following steps can be performed: compare the consistency of account C with each suspected illegal group account of the suspected illegal group corresponding to the previous time window; if account C is inconsistent with all suspected illegal group accounts of the suspected illegal group corresponding to the previous time window, compare the similarity of the text content of account C with each suspected illegal text content corresponding to the previous time window; if the similarity between the text content of account C and a suspected illegal text content corresponding to the previous time window meets the similarity threshold, account C and the text content published by account C are classified into the suspected illegal group to which the suspected illegal text content belongs in the previous time window, and account C is removed from the target account set corresponding to the current time window. In this way, each account in the target account set corresponding to the current time window is processed to obtain the filtered target account set corresponding to the current time window.

[0073] Step S203: cluster the text contents posted by the screened target account set within the current time window according to the similarity of the text contents, and obtain a library of suspected illegal groups corresponding to the current time window.

[0074] The account number plus the timestamp can be used as the unique identifier (ID) of the text content. On the one hand, it meets the requirement that the text content has uniqueness during clustering, and on the other hand, it is convenient to trace back the text content. Each text content has a corresponding account number.

[0075] The initial text content published in the current time window can be obtained, and the initial text content can be standardized to obtain the text content published in the current time window. The specific standardization process is as follows: identify and remove spaces and invisible characters; convert all letters to lowercase; identify traditional Chinese characters and convert them to simplified Chinese; convert non-half-width characters in the text content to half-width; truncate the text content, for example, truncate the text content to 30 characters.

[0076] The similarity of any two text contents published in the current time window can be calculated. If the similarity of two text contents is higher than the similarity threshold, the accounts that published the two text contents can be classified into the same cluster, thus forming multiple clusters. Each cluster includes several text contents. The text contents included in the same cluster can be determined as a suspected illegal group in the current time window, thereby forming a suspected illegal text content set corresponding to the current time window. Based on the suspected illegal text content set corresponding to the current time window, the suspected illegal group account set corresponding to the current time window can be obtained; based on the suspected illegal text content set and the suspected illegal group account set corresponding to the current time window, the suspected illegal group library corresponding to the current time window can be obtained.

[0077] In this embodiment, if the current time window is the first time window, the target account set corresponding to the current time window can be determined based on the text content published in the current time window; at this time, there is no need to screen the target account set. At this time, the text content published by the target account set in the current time window can be clustered according to the similarity of the text content to obtain a library of suspected illegal gangs corresponding to the current time window.

[0078] According to the above method, a database of suspected illegal groups corresponding to multiple time windows within a set period can be obtained, and the process proceeds to step S204.

[0079] Step S204, determining whether the corresponding account belongs to an illegal group account based on at least one of the number of times the same account is included in the suspected illegal group library corresponding to multiple time windows and the number of times the same text content is included in the suspected illegal group library corresponding to multiple time windows.

[0080] The account number threshold and text number threshold can be determined based on actual conditions.

[0081] If the number of times the same account is included in the suspected illegal gang database corresponding to multiple time windows is used to determine whether the corresponding account belongs to an illegal gang, the following steps can be performed specifically: If the set period includes 10 time windows, the suspected illegal gang database corresponding to these 10 time windows can be obtained in the above manner. For each account in the suspected illegal gang database corresponding to these 10 time windows, the following processing is performed, taking account D as an example: Determine the number of times account D appears in the suspected illegal gang database corresponding to these 10 time windows. If the number is greater than or equal to the account number threshold, it is determined that account D belongs to an illegal gang account. If the number is less than the account number threshold, it is determined that account D does not belong to an illegal gang account.

[0082] Among them, if in the suspected illegal gang database corresponding to 10 time windows, the suspected illegal gang database corresponding to 5 time windows includes account B, it can be determined that the number of times account B appears in the suspected illegal gang database corresponding to the 10 time windows is 5.

[0083] If the number of times the same text content is included in the suspected illegal group library corresponding to multiple time windows is used to determine whether the corresponding account belongs to the illegal group, the following steps can be performed specifically: If the set period includes 10 time windows, the suspected illegal group library corresponding to these 10 time windows can be obtained in the above manner. For each suspected illegal text content in the suspected illegal group library corresponding to these 10 time windows, the following processing is performed, taking the suspected illegal text content M as an example: determine the number of times the suspected illegal text content M appears in the suspected illegal group library corresponding to these 10 time windows. If the number is greater than or equal to the text number threshold, it is determined that the suspected illegal group account that published the suspected illegal text content M belongs to the illegal group account. If the number is less than the text number threshold, it is determined that the suspected illegal group account that published the suspected illegal text content M does not belong to the illegal group account.

[0084] Among them, it is also possible to determine whether the corresponding account belongs to the account of the illegal group according to the number of times the same phrase is included in the suspected illegal group library corresponding to multiple time windows. Specifically, the following steps can be performed: If the set period includes 10 time windows, the suspected illegal group library corresponding to these 10 time windows can be obtained in the above manner. Determine the phrases whose number of appearances in the suspected illegal group library corresponding to these 10 time windows is greater than or equal to the text number threshold, taking phrase N as an example: Determine the suspected illegal text content corresponding to phrase N in these 10 time windows, and determine the suspected illegal group account that publishes the above suspected illegal text content as the account of the illegal group.

[0085] If whether the corresponding account belongs to the account of the illegal gang is determined based on the number of times the same account is included in the suspected illegal gang library corresponding to multiple time windows and the number of times the same text content is included in the suspected illegal gang library corresponding to multiple time windows, then it is possible to first determine whether the corresponding account belongs to the account of the illegal gang based on the number of times the same account is included in the suspected illegal gang library corresponding to multiple time windows, and then determine whether the corresponding account belongs to the account of the illegal gang based on the number of times the same text content is included in the suspected illegal gang library corresponding to multiple time windows; it is also possible to first determine whether the corresponding account belongs to the account of the illegal gang based on the number of times the same text content is included in the suspected illegal gang library corresponding to multiple time windows, and then determine whether the corresponding account belongs to the account of the illegal gang based on the number of times the same account is included in the suspected illegal gang library corresponding to multiple time windows.

[0086] In the above-mentioned method for identifying accounts of illegal gangs, the length of the time window can be set to be shorter. At this time, the filtered target account set corresponding to the current time window can be clustered according to the similarity of the text content published in the time window with a shorter time length, so as to obtain a library of suspected illegal gangs corresponding to the current time window. This can reduce the magnitude of clustered data and improve calculation efficiency. It can also identify short-term and high-frequency group violations, alleviate the lag in identifying group violations, and improve the accuracy of identifying accounts of illegal gangs. In addition, the present application determines whether to include at least part of the accounts of the target account set into the library of suspected illegal gangs corresponding to the previous time window according to at least one of the account consistency and the text content similarity, so as to avoid the omission of some accounts that actually belong to illegal gangs due to the division of time windows, and further improve the accuracy of identifying accounts of illegal gangs.

[0087] In one embodiment, according to the account that publishes the text content in the current time window, the target account set corresponding to the current time window is determined. The specific steps are as follows: Figure 3 As shown: Step S301, determine the account that publishes the text content in the current time window based on the text content published in the current time window; Step S302, determine the initial account set corresponding to the current time window based on the account that publishes the text content in the current time window; Step S303, obtain the illegal accounts and whitelist accounts; Step S304, remove the illegal accounts and whitelist accounts from the initial account set to obtain the target account set corresponding to the current time window.

[0088] Based on the text content published in the current time window, determine the account that published the text content in the current time window, determine the initial account corresponding to the current time window based on the account that published the text content in the current time window, and obtain the initial account set corresponding to the current time window based on the initial account corresponding to the current time window.

[0089] Violating accounts refer to accounts that have violated regulations and been active recently; whitelist accounts refer to accounts with a lower probability of engaging in illegal group behavior, such as higher-level accounts.

[0090] Violating accounts and whitelist accounts can be obtained, and the violating accounts and whitelist accounts can be removed from the initial account set to obtain the target account set corresponding to the current time window. The above method of removing violating accounts and whitelist accounts can be called a policy funnel filtering method.

[0091] In this embodiment, illegal accounts and whitelist accounts are removed from the initial account set to obtain a target account set corresponding to the current time window, which can reduce invalid calculations in the process of identifying illegal group accounts and reduce the computational burden.

[0092] In one of the embodiments, after obtaining the suspected illegal group library corresponding to the current time window, the method provided by the present application also includes: obtaining the text content posted by the illegal account in the current time window; determining whether to include the illegal account and the content posted by the illegal account into the suspected illegal group library corresponding to the current time window based on the similarity between the text content posted by the illegal account in the current time window and the suspected illegal text content posted by the suspected illegal group account in the current time window.

[0093] You can get the text content posted by the violating account in the current time window. The violating account refers to the account that has violated the rules recently and has been active, but this type of account does not necessarily belong to the violating gang account and needs further identification.

[0094] In each suspected illegal group in the suspected illegal group library corresponding to the previous time window, at least one text content published by the suspected illegal group account in the previous time window can be selected as the representative text content of each suspected illegal group. For example, in the N suspected illegal groups in the suspected illegal group library corresponding to the previous time window, a suspected illegal text content published by the suspected illegal group account in the previous time window is randomly selected as the representative text content of the suspected illegal group, and N representative text contents are obtained.

[0095] The similar representative text content corresponding to the violating account can be determined based on the similarity between the text content posted by the violating account in the current time window and each representative text content; the violating account can be classified into the suspected violating group corresponding to the similar representative text content.

[0096] The specific steps for determining similar representative text content corresponding to the violating account are as follows: the hash value of the violating account in the current window can be obtained based on the text content published by the violating account in the current time window; the hash value of each representative text content can be obtained; the hash value can be the minimum hash value (MinHash). According to the closeness between the two hash values, the accounts that belong to the representative text content can be determined to be in the same set as the violating account, and the third subset of the target account can be obtained; according to the similarity between the text content published by the violating account in the current time window and the representative text content of other accounts in the third subset of the target account, the similar representative text content corresponding to the violating account can be determined.

[0097] In this embodiment, whether to include the violating account into the suspected violating group library corresponding to the current time window is determined based on the similarity between the text content posted by the violating account in the current time window and the suspected violating text content posted by the suspected violating group account in the current time window. This can avoid the omission of violating accounts that actually belong to the violating group and further improve the accuracy of identifying the violating group accounts.

[0098] In one of the embodiments, it is determined whether to include at least some of the accounts and the text content published by some of the accounts in the target account set into the suspected illegal gang library corresponding to the previous time window according to at least one of the account consistency and the text content similarity, so as to obtain the filtered target account set corresponding to the current time window, and the specific steps are as follows: the consistency of each account in the target account set is sequentially compared with each suspected illegal gang account in the suspected illegal gang library corresponding to the previous time window; when the account in the target account set is consistent with the suspected illegal gang account in the suspected illegal gang library corresponding to the previous time window, the account in the target account set and the text content published by the account are included into the suspected illegal gang library corresponding to the previous time window, and the account in the target account set is removed from the target account set to obtain the filtered target account set corresponding to the current time window; when the account in the target account set is inconsistent with the suspected illegal gang account in the suspected illegal gang library corresponding to the previous time window, it is determined whether to include the account in the target account set and the text content published by the account into the suspected illegal gang library corresponding to the previous time window according to the text content similarity, and the account in the target account set is removed from the target account set to obtain the filtered target account set corresponding to the current time window.

[0099] For the time window before the current time window, the suspected illegal gang database corresponding to the previous time window can be obtained; based on the suspected illegal gang database corresponding to the previous time window, the suspected illegal gang account corresponding to the previous time window and the suspected illegal text content of the above suspected illegal gang account in the previous time window are obtained.

[0100] The consistency of each account in the target account set with each suspected illegal gang account in the suspected illegal gang library corresponding to the previous time window can be compared in turn; when the account in the target account set is consistent with the suspected illegal gang account in the suspected illegal gang account library corresponding to the previous time window, it means that the account in the target account set meets the entry criteria. At this time, the account in the target account set and the text content published by the account can be classified into the suspected illegal gang library corresponding to the previous time window, and the account in the target account set can be removed from the target account set to obtain the filtered target account set corresponding to the current time window.

[0101] When the accounts in the target account set are inconsistent with the suspected illegal gang accounts in the suspected illegal gang database corresponding to the previous time window, it means that the accounts in the target account set do not meet the entry criteria. At this time, based on the text content similarity, it is further determined whether the accounts in the target account set meet the entry criteria.

[0102] When it is determined that the accounts in the target account set meet the entry criteria, the accounts in the target account set and the text content published by the accounts are added to the suspected illegal gang library corresponding to the previous time window, and the accounts in the target account set are removed from the target account set to obtain the filtered target account set corresponding to the current time window.

[0103] Taking account C in the target account set corresponding to the current time window as an example, account C can be compared for consistency with each suspected illegal group account of the suspected illegal group corresponding to the previous time window; if account C is inconsistent with all suspected illegal group accounts of the suspected illegal group corresponding to the previous time window, the text content of account C is compared for similarity with each suspected illegal text content of the suspected illegal group corresponding to the previous time window; if the similarity between the text content of account C and a suspected illegal text content corresponding to the previous time window meets the similarity threshold, account C and the text content published by account C are classified into the suspected illegal group to which the account belongs in the previous time window, and account C is removed from the target account set corresponding to the current time window. In this way, each account in the target account set corresponding to the current time window is processed to obtain the filtered target account set corresponding to the current time window.

[0104] In this embodiment, it is first determined whether to include at least part of the accounts in the target account set into the suspected illegal gang library corresponding to the previous time window based on the account consistency; when the accounts in the target account set are inconsistent with the accounts in the suspected illegal gang library corresponding to the previous time window, it is determined whether to include at least part of the accounts in the target account set into the suspected illegal gang library corresponding to the previous time window based on the text content similarity, thereby obtaining the filtered target account set corresponding to the current time window, which can reduce invalid calculations in the process of identifying illegal gang accounts and reduce the computational burden; in addition, it can avoid the omission of some accounts that actually belong to illegal gangs due to the division of time windows, thereby further improving the accuracy of identifying illegal gang accounts.

[0105] In one of the embodiments, based on the similarity of text content, it is determined whether to include the accounts in the target account set and the text content posted by some accounts into the suspected illegal gang library corresponding to the previous time window. The specific steps are as follows: the accounts in the target account set are used as the accounts to be queried; in each suspected illegal gang in the suspected illegal gang library corresponding to the previous time window, at least one suspected illegal text content posted by the account of the suspected illegal gang in the previous time window is selected as the representative text content of each suspected illegal gang; based on the similarity between the text content posted by the account to be queried in the current time window and each representative text content, similar representative text content corresponding to the account to be queried is determined; and the account to be queried and the text content posted by the account to be queried are included into the suspected illegal gang corresponding to the similar representative text content.

[0106] When the account number of the target account set is inconsistent with the suspected illegal group account number in the suspected illegal group database corresponding to the previous time window, the account number of the target account set is used as the account to be queried.

[0107] In each suspected illegal group in the suspected illegal group database corresponding to the previous time window, at least one suspected illegal group account published in the previous time window is randomly selected as the representative text content of each suspected illegal group. For example, in the N suspected illegal groups in the suspected illegal group database corresponding to the previous time window, a suspected illegal text content published in the previous time window by a suspected illegal account is randomly selected as the representative text content of the suspected illegal group, and N representative text contents are obtained.

[0108] Based on the similarity between the text content published by the account to be queried in the current time window and each representative text content, the similar representative text content corresponding to the account to be queried is determined; the account to be queried and the text content published by the account to be queried are classified into the suspected violation group corresponding to the similar representative text content.

[0109] In this embodiment, similar representative text content corresponding to the account to be queried is determined based on the similarity between the text content published by the account to be queried in the current time window and each representative text content; the account to be queried and the text content published by the account to be queried are classified into the suspected illegal group corresponding to the similar representative text content, which can avoid the omission of some accounts that actually belong to the illegal group due to the division of the time window, and further improve the accuracy of identifying the illegal group account.

[0110] In one of the embodiments, similar representative text content corresponding to the account to be queried is determined based on the similarity between the text content published by the account to be queried in the current time window and each representative text content. The specific steps are as follows: based on the text content published by the account to be queried in the current time window, a hash value of the account to be queried in the current window is obtained; the hash value of each representative text content is obtained; based on the proximity between each pair of hash values, among the accounts to which each representative text content belongs, accounts that are classified into the same set as the account to be queried are determined to obtain a target account subset; based on the similarity between the text content published by the account to be queried in the current time window and the representative text content of other accounts in the target account subset, similar representative text content corresponding to the account to be queried is determined.

[0111] The hash value of the account to be queried in the current window can be obtained according to the text content published by the account to be queried in the current time window, wherein the hash value can be a minimum hash value (MinHash).

[0112] A hash value of each representative text content may be obtained, wherein the hash value may be a minimum hash value (MinHash).

[0113] According to the closeness between the hash values, the accounts that are classified into the same set as the account to be queried can be determined among the accounts to which the text content belongs, and the target account subset can be obtained. A locality sensitive hashing (LSH) index can be set for each account in the target account subset.

[0114] The similar representative text content corresponding to the account to be queried can be determined based on the similarity between the text content published by the account to be queried in the current time window and the representative text content of other accounts in the target account subset. Specifically, when there is a representative text content whose similarity with the text content published by the account to be queried in the current time window is higher than the similarity threshold, the representative text content is determined as the similar representative text content corresponding to the account to be queried; when there are multiple representative text contents whose similarity with the text content published by the account to be queried in the current time window is higher than the similarity threshold, the magnitude of the suspected illegal groups corresponding to the above multiple representative text contents is determined, and the representative text content corresponding to the suspected illegal group with the largest magnitude is determined as the similar representative text content corresponding to the account to be queried.

[0115] In this embodiment, based on the degree of proximity between the hash value of the account to be queried in the current window and the hash value of each representative text content, among the accounts to which each representative text content belongs, the accounts that are classified into the same set as the account to be queried are determined to obtain a target account subset; based on the similarity between the text content published by the account to be queried in the current time window and the representative text content of other accounts in the target account subset, similar representative text content corresponding to the account to be queried is determined, thereby effectively improving the calculation efficiency without reducing the calculation accuracy.

[0116] In one of the embodiments, the text contents posted by the screened target account set in the current time window are clustered according to the similarity of the text contents to obtain a suspected illegal group library corresponding to the current time window, and the specific steps are as follows: obtain the hash value of each text content posted by the screened target account set in the current time window; divide the text contents corresponding to the screened target account set in the current time window according to the proximity between the hash values ​​to obtain several target text content subsets; for any target text content subset, cluster the text contents in the target text content subset according to the similarity of the text contents in the target text content subset to obtain multiple clusters; when the number of text contents included in the cluster is greater than or equal to the number of text contents included in the minimum cluster, take the text contents included in the cluster as suspected illegal text contents to obtain a suspected illegal text content set corresponding to the current time window; obtain a suspected illegal group account set corresponding to the current time window according to the suspected illegal text content set corresponding to the current time window; obtain a suspected illegal group library corresponding to the current time window according to the suspected illegal text content set and the suspected illegal group account set corresponding to the current time window.

[0117] Obtain the hash value of each text content published by the filtered target account set within the current time window, where the hash value may be a minimum hash value (MinHash).

[0118] According to the closeness between the hash values ​​of each pair, the text content corresponding to the filtered target account set in the current time window can be divided to obtain several target text content subsets, among which a local sensitive hash index can be set for each text content in the target text content subset.

[0119] For any target text content subset, the text content in the target text content subset is clustered according to the similarity of the text content in the target text content subset to obtain multiple clusters; when the number of text contents included in the cluster is greater than or equal to the number of text contents included in the minimum cluster, the text content included in the cluster is used as the suspected illegal text content to obtain the suspected illegal text content set corresponding to the current time window; according to the suspected illegal text content set corresponding to the current time window, the suspected illegal group account set corresponding to the current time window is obtained; the suspected illegal text content set and the suspected illegal group account set corresponding to the current time window are used as the suspected illegal group library corresponding to the current time window. Among them, the number of text contents included in the minimum cluster can be set according to the actual situation, for example, the number of text contents included in the minimum cluster can be set to 5.

[0120] After obtaining the suspected illegal group database corresponding to multiple time windows, some suspected illegal group accounts can be marked according to the policy configuration plan, and different degrees of punishment can be imposed later. The specific policy configuration plan is as follows: the clustering process in which the number of text contents included in the smallest cluster is greater than or equal to 5 is determined as the target clustering process, and the cluster with a score greater than or equal to the score threshold in the target clustering process is used as the target cluster. The suspected illegal group accounts corresponding to the suspected illegal texts in the target cluster that are located before the set ranking are marked, and different degrees of punishment can be imposed later. Among them, the score threshold can be set according to the actual situation, for example, the score threshold can be set to 0.8; the set ranking can be set according to the actual situation, for example, the ranking can be set to 15%.

[0121] In this embodiment, the text content corresponding to the filtered target account set in the current time window is divided according to the degree of proximity between the hash values ​​of the text content published by the filtered target account set in the current time window to obtain several target text content subsets; for any target text content subset, the text content in the target text content subset is clustered according to the similarity of the text content in the target text content subset to obtain the suspected illegal text content set corresponding to the current time window, and then obtain the suspected illegal gang library corresponding to the current time window, which can effectively improve the clustering efficiency without reducing the clustering accuracy.

[0122] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0123] Based on the same inventive concept, the embodiment of the present application also provides a device for identifying illegal group accounts for implementing the illegal group account identification method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the device for identifying illegal group accounts provided below can refer to the limitations of the illegal group account identification method above, and will not be repeated here.

[0124] In an exemplary embodiment, Figure 4As shown, a device for identifying illegal gang accounts is provided, wherein:

[0125] The target account set acquisition module 401 is used to determine the target account set corresponding to the current time window according to the text content published in the current time window if the current time window is not the first time window; each text content is published by a corresponding account;

[0126] The target account set screening module 402 is used to determine whether to classify at least some of the accounts in the target account set and the text contents published by the accounts into the suspected illegal gang library corresponding to the previous time window, so as to obtain the screened target account set corresponding to the current time window, based on at least one of the account consistency and the text content similarity;

[0127] The suspected illegal group library acquisition module 403 is used to cluster the text content posted by the screened target account set within the current time window according to the similarity of the text content, and obtain the suspected illegal group library corresponding to the current time window;

[0128] The illegal group account determination module 404 is used to determine whether the corresponding account belongs to an illegal group account based on at least one of the number of times the same account is included in the suspected illegal group library corresponding to multiple time windows and the number of times the same text content is included in the suspected illegal group library corresponding to multiple time windows.

[0129] In one of the embodiments, the target account set acquisition module 401 is also used to: determine the account that publishes text content in the current time window based on the text content published in the current time window; determine the initial account set corresponding to the current time window based on the account that publishes text content in the current time window; obtain the violating account and the whitelist account; remove the violating account and the whitelist account from the initial account set to obtain the target account set corresponding to the current time window.

[0130] In one of the embodiments, the device also includes a violation account storage module, which is used to: obtain the text content posted by the violation account in the current time window; and determine whether to include the violation account and the content posted by the violation account into the suspected violation group library corresponding to the current time window based on the similarity between the text content posted by the violation account in the current time window and the suspected violation text content posted by the suspected violation group account in the current time window.

[0131] In one embodiment, the target account set screening module 402 is further used to: sequentially compare the consistency of each account in the target account set with each suspected illegal group account in the suspected illegal group library corresponding to the previous time window; when the account in the target account set is consistent with the suspected illegal group account in the suspected illegal group library corresponding to the previous time window, the account in the target account set and the text content published by the account are classified into the suspected illegal group library corresponding to the previous time window, and the account in the target account set is removed from the target account set to obtain the filtered target account set corresponding to the current time window; when the account in the target account set is inconsistent with the suspected illegal group account in the suspected illegal group library corresponding to the previous time window, determine whether to classify the account in the target account set and the text content published by the part of the accounts into the suspected illegal group library corresponding to the previous time window according to the text content similarity, and remove the account in the target account set from the target account set to obtain the filtered target account set corresponding to the current time window.

[0132] In one of the embodiments, the target account set screening module 402 is further used to: use the accounts in the target account set as the accounts to be queried; from each suspected violating group in the suspected violating group library corresponding to the previous time window, select at least one suspected violating text content published by the account of the suspected violating group in the previous time window as the representative text content of each suspected violating group; determine the similar representative text content corresponding to the account to be queried based on the similarity between the text content published by the account to be queried in the current time window and each representative text content; and classify the account to be queried and the text content published by the account to be queried into the suspected violating group corresponding to the similar representative text content.

[0133] In one embodiment, the target account set screening module 402 is further used to: obtain a hash value of the account to be queried in the current window based on the text content published by the account to be queried in the current time window; obtain the hash value of each representative text content; determine the accounts that are classified into the same set as the account to be queried among the accounts to which the representative text content belongs based on the proximity between each other's hash values, and obtain a target account subset; determine similar representative text content corresponding to the account to be queried based on the similarity between the text content published by the account to be queried in the current time window and the representative text content of other accounts in the target account subset.

[0134] In one of the embodiments, the suspected illegal group library acquisition module 403 is also used to: obtain the hash value of each text content published by the filtered target account set within the current time window; divide the text content corresponding to the filtered target account set in the current time window according to the proximity between the hash values, and obtain a number of target text content subsets; for any target text content subset, cluster the text content in the target text content subset according to the similarity of the text content in the target text content subset to obtain multiple clusters; when the number of text contents included in the cluster is greater than or equal to the number of text contents included in the smallest cluster, use the text content included in the cluster as suspected illegal text content to obtain the suspected illegal text content set corresponding to the current time window; obtain the suspected illegal group account set corresponding to the current time window according to the suspected illegal text content set corresponding to the current time window; obtain the suspected illegal group library corresponding to the current time window according to the suspected illegal text content set and the suspected illegal group account set corresponding to the current time window.

[0135] Each module in the above-mentioned illegal gang account identification device can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of the above modules.

[0136] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data of an embodiment of the method for identifying an illegal gang account. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for identifying an illegal gang account is implemented.

[0137] Those skilled in the art will understand that Figure 5The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0138] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0139] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0140] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0141] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0142] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0143] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0144] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for identifying illegal group accounts, characterized in that: The method comprises: If the current time window is not the first time window, determine the target account set corresponding to the current time window according to the text content published in the current time window; wherein each text content is published by a corresponding account; Determine, based on at least one of account consistency and text content similarity, whether to classify at least some of the accounts in the target account set and the text content published by the some of the accounts into a suspected illegal group library corresponding to a previous time window, so as to obtain a filtered target account set corresponding to the current time window; Clustering the text content posted by the screened target account set within the current time window according to the similarity of the text content, to obtain a suspected illegal group library corresponding to the current time window; Determine whether the corresponding account belongs to the illegal group account based on at least one of the number of times the same account is included in the suspected illegal group library corresponding to multiple time windows and the number of times the same text content is included in the suspected illegal group library corresponding to multiple time windows.

2. The method according to claim 1, characterized in that The step of determining a target account set corresponding to the current time window according to the text content published in the current time window includes: According to the text content published in the current time window, determine the account that published the text content in the current time window; Determining an initial account set corresponding to the current time window according to the accounts that publish text content within the current time window; Get illegal accounts and whitelist accounts; The illegal accounts and the whitelist accounts are removed from the initial account set to obtain a target account set corresponding to the current time window.

3. The method according to claim 2, characterized in that After obtaining the suspected illegal group database corresponding to the current time window, the method further includes: Get the text content posted by the violating account in the current time window; Based on the similarity between the text content posted by the violating account in the current time window and the suspected violating text content posted by the suspected violating group account in the current time window, determine whether to include the violating account and the content posted by the violating account into the suspected violating group library corresponding to the current time window.

4. The method according to claim 1, characterized in that: The determining, based on at least one of account consistency and text content similarity, whether to classify at least some of the accounts in the target account set and the text content published by the some of the accounts into a suspected illegal group library corresponding to a previous time window, so as to obtain a filtered target account set corresponding to the current time window, includes: Comparing the consistency of each account in the target account set with each suspected illegal group account in the suspected illegal group database corresponding to the previous time window in turn; When the accounts in the target account set are consistent with the suspected illegal gang accounts in the suspected illegal gang library corresponding to the previous time window, the accounts in the target account set and the text content published by the accounts are classified into the suspected illegal gang library corresponding to the previous time window, and the accounts in the target account set are removed from the target account set to obtain the filtered target account set corresponding to the current time window; When the accounts in the target account set are inconsistent with the suspected illegal group accounts in the suspected illegal group library corresponding to the previous time window, determine whether to include the accounts in the target account set and the text contents posted by some of the accounts into the suspected illegal group library corresponding to the previous time window based on the similarity of the text contents, and remove the accounts in the target account set from the target account set to obtain the filtered target account set corresponding to the current time window.

5. The method according to claim 4, characterized in that The determining, based on the text content similarity, whether to classify the text content posted by the accounts in the target account set and the partial accounts into the suspected illegal group library corresponding to the previous time window includes: Using the account number in the target account set as the account number to be queried; From each suspected illegal group in the suspected illegal group database corresponding to the previous time window, select at least one suspected illegal text content posted by the account of the suspected illegal group in the previous time window as the representative text content of each suspected illegal group; Determine similar representative text content corresponding to the account to be queried according to the similarity between the text content published by the account to be queried in the current time window and each representative text content; The account to be queried and the text content published by the account to be queried are classified into the suspected illegal group corresponding to the similar representative text content.

6. The method according to claim 5, characterized in that The step of determining similar representative text content corresponding to the account to be queried according to the similarity between the text content published by the account to be queried in the current time window and each representative text content includes: Obtaining a hash value of the account to be queried in the current window according to the text content published by the account to be queried in the current time window; Get the hash value of each representative text content; According to the closeness between the hash values, among the accounts to which the representative text contents belong, the accounts that are classified into the same set as the account to be queried are determined to obtain a target account subset; Based on the similarity between the text content published by the account to be queried in the current time window and the representative text content of other accounts in the target account subset, similar representative text content corresponding to the account to be queried is determined.

7. The method according to claim 1, characterized in that According to the similarity of the text content, the text content published by the screened target account set within the current time window is clustered to obtain a suspected illegal group library corresponding to the current time window, including: Obtaining a hash value of each text content published by the filtered target account set within the current time window; According to the closeness between the hash values, the text content corresponding to the filtered target account set in the current time window is divided to obtain a plurality of target text content subsets; For any target text content subset, clustering the text contents in the target text content subset according to the similarity of the text contents in the target text content subset to obtain multiple clusters; When the number of text contents included in the cluster is greater than or equal to the number of text contents included in the minimum cluster, the text contents included in the cluster are used as suspected illegal text contents to obtain a suspected illegal text content set corresponding to the current time window; According to the suspected illegal text content set corresponding to the current time window, obtain the suspected illegal group account set corresponding to the current time window; According to the suspected illegal text content set and the suspected illegal group account set corresponding to the current time window, a suspected illegal group library corresponding to the current time window is obtained.

8. A device for identifying illegal group accounts, characterized in that: The device comprises: A target account set acquisition module, used to determine the target account set corresponding to the current time window according to the text content published in the current time window if the current time window is not the first time window; wherein each text content is published by a corresponding account; A target account set screening module, used to determine whether to classify at least some of the accounts in the target account set and the text contents published by the accounts into a suspected illegal gang library corresponding to a previous time window, based on at least one of account consistency and text content similarity, so as to obtain a screened target account set corresponding to the current time window; A suspected illegal group library acquisition module is used to cluster the text content posted by the screened target account set within the current time window according to the similarity of the text content, and obtain the suspected illegal group library corresponding to the current time window; The illegal group account determination module is used to determine whether the corresponding account belongs to the illegal group based on at least one of the number of times the same account is included in the suspected illegal group library corresponding to multiple time windows and the number of times the same text content is included in the suspected illegal group library corresponding to multiple time windows.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

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

Images