Content recommendation method, device, computer equipment and computer-readable storage medium

Abstract

The embodiments of the present application disclose a content recommendation method, apparatus, computer device, and computer-readable storage medium. The method comprises obtaining a data change message for recommended content in a content recommendation list, wherein the content recommendation list includes at least one recommended content sorted according to a preset rule; parsing the data change message to obtain the content type of the recommended content in the content recommendation list; obtaining a status list corresponding to the content recommendation list when the content type is a specified content type; performing corresponding adjustments to the recommended status of the recommended content in the content recommendation list to obtain an adjusted status list; and performing content recommendation processing on the content recommendation list based on the adjusted status list. This solution determines the content type of the recommended content based on the message generated by target content identification and adjusts the status list accordingly. This solution can de-recommend recommended content in the content recommendation list that belongs to the specified content type, thereby improving the effectiveness of content recommendation.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a content recommendation method, apparatus, computer device, and computer-readable storage medium. Background Art

[0002] When recommending content, the recommended content is typically ranked based on factors such as exposure and clickthrough rates, and a certain number of recommended content is determined and displayed. However, in order to get their own recommended content displayed, some users add illegal content to their content, inducing users to click on it and increasing its clickthrough rate, thereby ensuring that it is selected as recommended content and displayed. However, this type of recommended content is often low-quality, resulting in a large amount of low-quality content in the recommended content, and poor content recommendation results. Summary of the Invention

[0003] The embodiments of the present application provide a content recommendation method, apparatus, computer device, and computer-readable storage medium, which can improve the effect of content recommendation.

[0004] An embodiment of the present application provides a content recommendation method, including:

[0005] Obtaining a data change message for recommended content in a content recommendation list, wherein the content recommendation list includes at least one recommended content sorted according to a preset rule, and the data change message includes a message generated by performing target content identification on the recommended content;

[0006] Parsing the data change message to obtain the content type of the recommended content in the content recommendation list;

[0007] When the content type is a specified content type, obtaining a status list corresponding to the content recommendation list, the status list including the recommendation status of the recommended content in the content recommendation list;

[0008] Performing corresponding adjustment processing on the recommendation status of the recommended content in the content recommendation list to obtain an adjusted status list;

[0009] Perform content recommendation processing on the content recommendation list according to the adjusted status list.

[0010] Accordingly, an embodiment of the present application further provides a content recommendation device, including:

[0011] a message acquiring unit, configured to acquire a data change message for a recommended content in a content recommendation list, wherein the content recommendation list includes at least one recommended content sorted according to a preset rule, and the data change message includes a message generated by performing target content identification on the recommended content;

[0012] A parsing unit, configured to parse the data change message to obtain the content type of the recommended content in the content recommendation list;

[0013] a list acquisition unit, configured to acquire a status list corresponding to the content recommendation list when the content type is a specified content type, the status list including the recommendation status of the recommended content in the content recommendation list;

[0014] An adjusting unit, configured to adjust the recommendation status of the recommended content in the content recommendation list accordingly to obtain an adjusted status list;

[0015] A recommendation unit is configured to perform content recommendation processing on the content recommendation list according to the adjusted status list.

[0016] Correspondingly, an embodiment of the present application also provides a computer device, including a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute any content recommendation method provided in the embodiment of the present application.

[0017] Accordingly, an embodiment of the present application further provides a computer-readable storage medium, which is used to store a computer program, and the computer program is loaded by a processor to execute any content recommendation method provided in the embodiment of the present application.

[0018] The embodiment of the present application obtains a data change message for recommended content in a content recommendation list, wherein the content recommendation list includes at least one recommended content sorted according to a preset rule, and the data change message includes a message generated by target content identification for the recommended content; the data change message is parsed to obtain the content type of the recommended content in the content recommendation list; when the content type is a specified content type, a status list corresponding to the content recommendation list is obtained, the status list includes the recommended status of the recommended content in the content recommendation list; the recommended status of the recommended content in the content recommendation list is adjusted accordingly to obtain an adjusted status list; and content recommendation processing is performed on the content recommendation list based on the adjusted status list. This solution determines the content type of the recommended content based on the message generated by target content identification for the recommended content in the content recommendation list, and adjusts the status list accordingly. It can thus improve the effect of content recommendation by not recommending the recommended content in the content recommendation list that belongs to the specified content type. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 This is a scene diagram of the content recommendation method provided in the embodiment of the present application;

[0021] Figure 2 This is a flowchart of the content recommendation method provided by an embodiment of the present application;

[0022] Figure 3 is another flow chart of the content recommendation method provided by an embodiment of the present application;

[0023] Figure 4 This is another flow chart of the content recommendation method provided by the embodiment of the present application;

[0024] Figure 5 Schematic diagram of a content recommendation device provided in an embodiment of the present application;

[0025] Figure 6 It is a structural diagram of the computer structure provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0027] The present invention provides a content recommendation method, apparatus, computer device, and computer-readable storage medium. The content recommendation apparatus can be integrated into a computer device, which can be a server or a terminal.

[0028] The terminal may include a mobile phone, a wearable smart device, a tablet computer, a laptop computer, a personal computer (PC), and a vehicle-mounted computer.

[0029] Among them, the server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, as well as big data and artificial intelligence platforms.

[0030] For example, Figure 1 As shown, a computer device obtains a data change message for recommended content in a content recommendation list and a status list corresponding to the content recommendation list, the status list including the recommended status of the recommended content in the content recommendation list; parses the data change message to obtain the content type of the recommended content in the content recommendation list; when the content type is a specified content type, adjusts the recommendation status of the recommended content in the content recommendation list accordingly to obtain an adjusted status list; and performs content recommendation processing on the content recommendation list based on the adjusted status list. This solution determines the content type of the recommended content based on a message generated by target content identification of the recommended content in the content recommendation list and adjusts the status list accordingly. This solution can de-recommend recommended content in the content recommendation list that belongs to the specified content type, thereby improving the effectiveness of content recommendation.

[0031] The content recommendation list may be a list obtained by sorting recommended content samples in a database according to preset rules. For example, the content recommendation list may be a list containing a certain number of recommended content samples, sorted according to data such as the number of clicks, exposure, or conversion rate of each recommended content sample in the database within a preset period or history. The content recommendation list may include at least one recommended content. The recommended content may be content that appears in the content recommendation list after being sorted according to preset rules. The recommended content may be articles, videos, pictures, or other works.

[0032] The status list may be a list corresponding to the content recommendation list, and the status list may include the recommendation status of the recommended content in the content recommendation list, and the recommendation status of the recommended content determines whether to recommend the recommended content.

[0033] It should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments.

[0034] This embodiment will be described from the perspective of a content recommendation device. The content recommendation device may be integrated into a computer device, which may be a server or a terminal.

[0035] The present application provides a content recommendation method, such as Figure 2As shown, the specific process of the content recommendation method can be as follows:

[0036] 101. Obtain a data change message for recommended content in a content recommendation list, where the content recommendation list includes at least one recommended content sorted according to a preset rule, and the data change message includes a message generated by performing target content identification on the recommended content.

[0037] The data change message may include a message generated according to a result of target content identification performed on the recommended content in the content list.

[0038] For example, a content review service could identify the target content in a content recommendation list, determine the recommended content type, and generate a corresponding data change message based on the recommended content type, storing the data change message in a database or on a blockchain. A data processing service could retrieve the data change message from a database or blockchain, or the content review service could send the data change message to the data processing service over the network. The data processing service would then adjust the status list based on the data change message.

[0039] Among them, the content review service and data processing service can be implemented through corresponding code programs. The code programs of the content review service and data processing service can be stored on the same server or on different servers, which is not limited here.

[0040] In one embodiment, target content may be identified for recommended content samples in a database or blockchain, and a data change message may be generated. That is, before the step of "receiving a data change message for recommended content in a content recommendation list," the content recommendation method may further specifically include:

[0041] Obtaining a recommended content sample from a database, where the recommended content sample includes recommended content in a content recommendation list;

[0042] Performing target content recognition on the recommended content sample to obtain a content type of the recommended content sample, where the recommended content sample includes the recommended content in the content recommendation list;

[0043] A data change message is generated based on the content type of the recommended content sample.

[0044] The database may store recommended content samples posted by users, the recommended content samples may include recommended content in a content recommendation list, and the recommended content may be content determined by sorting the recommended content samples according to preset rules.

[0045] Among them, the content type may include a specified content type, such as an abnormal content type, which indicates that the recommended content contains abnormal content, for example, it contains misleading text content to induce users to click on the recommended content, or contains illegal controls, such as buttons that imitate system functions, buttons that turn off functions, and fake function buttons.

[0046] For example, it may be to perform text content recognition on the recommended content sample and determine the content type of the recommended content sample based on the text content contained in the recommended content sample; or it may be to perform image processing on the recommended content to identify whether the recommended content contains abnormal content and then determine the content type of the recommended content; or it may be to determine the content type of the recommended content in response to the user's operation, for example, to determine the content type of the recommended content specified by the user as the specified content type.

[0047] If the content type of the recommended content sample is an abnormal content type, the content type of the recommended content sample in the database is updated to the abnormal content type, or its corresponding review status is set to review failure.

[0048] The binary file binlog records user statements for database updates, such as statements and timestamps for changing database tables and data table contents. Therefore, the binlog generates data update messages based on content type updates or review status updates of recommended content samples.

[0049] In one embodiment, decoupling can be achieved through a message queue. For an existing content recommendation list, there is no need to modify the original code program. Instead, the recommendation status of the recommended content can be determined by simply obtaining a data update message from the message queue. That is, after the step of "generating a data change message based on the content type of the recommended content sample," the following steps may also be included:

[0050] Send data change messages to the message queue;

[0051] Get data change messages for recommended content in the content recommendation list, including:

[0052] Get data change messages for recommended content in the content recommendation list from the message queue.

[0053] Among them, the message queue can be Pulsar, Kafka, etc.

[0054] For example, the data change message can be pushed to the message queue, and then the data change message can be obtained from the message queue. This can decouple the process of identifying the target content of the recommended content samples and the process of adjusting the status list. At the same time, asynchronous processing is realized to avoid problems such as machine crashes caused by computing power mismatch.

[0055] Typically, a content recommendation list includes recommended content and related information about the recommended content, such as content ID, author ID, and ranking ID. A status list corresponding to the content recommendation list can be established based on the related information about the recommended content. Recommended content in the content recommendation list can be recommended based on the status list. That is, before the step of "obtaining data change information for recommended content in the content recommendation list," the content recommendation method can further include:

[0056] Get the recommended status and associated content identifier of the recommended content in the content recommendation list;

[0057] A status list is established according to the associated content identifier and the recommendation status of the recommended content, and the status list is associated with the content recommendation list through the associated content identifier.

[0058] The recommendation status may be a status determined based on the result of target content identification of the recommended content. For example, if the target content identification is performed on the recommended content and it is determined to be a specified content type, the recommendation status is a recommendable status, otherwise it is a non-recommended status.

[0059] For example, the recommendation status of the recommended content in the content recommendation list may be obtained, and the content ID of the recommended content may be obtained as the associated content identifier of the recommended content, and the associated content identifier may be used as a query field to establish a status list.

[0060] Optionally, when a content recommendation list is generated, the recommendation status of the recommended content can be pre-set to recommended. After the target content is identified for the recommended content, the recommendation status is adjusted according to the identification result, or when the content recommendation list is generated, the historical target content identification result is obtained to determine the recommendation status of the recommended content.

[0061] By generating a status list corresponding to a content recommendation list through associated content identifiers of recommended content, it is unnecessary to modify the existing content recommendation list, thereby improving the convenience of integrating the content recommendation method into the existing content recommendation list.

[0062] In one embodiment, the step of “performing target content identification on the recommended content sample to obtain the content type of the recommended content sample” may specifically include:

[0063] Perform text information recognition on the recommended content sample to obtain the text information contained in the recommended content;

[0064] Determine the content type of the recommended content sample based on the text information.

[0065] For example, if the recommended content sample is presented in the form of an article, text information recognition can be performed on the recommended content sample based on preset text information, and text information matching the preset text information can be searched in the recommended content sample to obtain the text information contained in the recommended content. If the text information is empty, the content type of the recommended content is determined to be normal content type. If the text information is not empty, the content type of the recommended content is determined to be abnormal content type. The preset text information can be pre-set illegal text information.

[0066] If the recommended content sample is presented in the form of a video, you can obtain the keyframe image or cover image of the video, perform optical character recognition (OCR) on the keyframe image or cover image, identify the text information contained in the keyframe image or cover image, and compare the text information with the preset text information. If there is a match, the content type of the recommended content sample is determined to be an abnormal content type. If there is no match, the content type of the recommended content sample is determined to be a normal content type. If the recommended content sample is presented in the form of an image, you can directly perform OCR recognition on the recommended content to identify the text information contained in the recommended content.

[0067] If the recommended content sample is presented in the form of a video or image, in addition to text information recognition, image recognition can also be performed to identify whether the recommended content contains illegal controls. That is, in one embodiment, the step of "performing target content recognition on the recommended content sample to obtain the content type of the recommended content sample" can specifically include:

[0068] Get reference content corresponding to the preset content type;

[0069] Perform content matching on the reference content and the recommended content sample to obtain a matching result between the reference content and the recommended content sample;

[0070] The content type of the recommended content sample is determined based on the matching results.

[0071] The reference content may include multiple illegal contents, which are used to match with the recommended content to determine the content type of the recommended content. For example, it may be a picture containing illegal controls.

[0072] For example, it can be to obtain reference content corresponding to the abnormal content type, that is, to obtain a picture containing illegal controls, use the picture as a template, and perform template matching with the recommended content in picture form, or perform template matching with the key frame picture and cover picture of the recommended content in video form. If matching content is found in the recommended content, the matching result is a match, and the content type of the recommended content is an abnormal content type. If no matching content is found in the recommended content, the matching result is a mismatch, and the content type of the recommended content is a normal content type.

[0073] In one embodiment, the content type of the recommended content may be predicted using a neural network model. Specifically, the step of "performing target content identification on the recommended content sample to obtain the content type of the recommended content sample" may include:

[0074] Extracting content features of the recommended content samples in the content recommendation list to obtain content feature information of the recommended content samples;

[0075] Predict the content type of the recommended content sample based on content feature information.

[0076] For example, specifically, convolution processing may be performed on the recommended content sample or the key frame image (such as the cover image) of the recommended content sample to obtain convolution feature information of the key frame image, and content feature extraction may be performed on the convolution feature information to obtain content feature information of the recommended content sample; based on the content feature information, it may be predicted whether the recommended content contains illegal controls; if so, the content type of the recommended content sample is the specified content type; if not, the content type of the recommended content sample is the normal content type.

[0077] Optionally, the above process can be implemented through an abnormal content recognition model. The initial abnormal content recognition model is trained with image samples containing illegal controls to obtain an abnormal content recognition model. The content feature information of the recommended content sample is extracted through the abnormal content recognition model, and the content type of the recommended content sample is predicted based on the content feature information.

[0078] 102. Parse and process the data change message to obtain the content type of the recommended content in the content recommendation list.

[0079] For example, the data change message may be parsed to determine the recommended content indicated by the data change message, and the content type of the recommended content may be determined to be a specified content type.

[0080] Since the data change message records the change record of the content type or review status of the recommended content sample, if the recommended content sample is recommended content in the content recommendation list, the content type of the recommended content can be determined based on the data change record. That is, in one embodiment, the step of "parsing the data change message to obtain the content type of the recommended content in the content recommendation list" may specifically include:

[0081] Parse the data structure of the data change message to obtain the content identifier corresponding to the data change message;

[0082] If the content identifier matches the recommended content in the content recommendation list, the content type of the recommended content is determined to be a specified content type.

[0083] For example, binlog can generate data change records based on content type updates or review status updates of recommended content samples, and generate data update messages based on protocol data exchange format tools, such as Protobuf or JSON. The data update messages can include the data in the data change records in a preset format.

[0084] By parsing the data structure of the data update message, the content identifier contained in the data change message can be extracted. Based on the content identifier, it can be determined whether the recommended content sample with content type update or review status update is the recommended content in the content recommendation list, and the content type of the recommended content can be determined. For example, if the content identifier matches the recommended content in the content recommendation list, it can be determined whether it is a content type update or a review status update.

[0085] Optionally, in order to increase the speed of determining the content type of the recommended content, the data of the recommendation status or review status corresponding to the recommended content sample can be stored in a preset data table, and the location identifier can be used to determine whether the data change message is a data change message corresponding to the recommended content sample. If so, the data structure of the data update message is parsed, that is, the step of "parsing the data structure of the data change message to obtain the content identifier corresponding to the data change message" can specifically include:

[0086] Get the location identifier in the data change message;

[0087] If the location identifier is the target location identifier, the data structure of the data change message is parsed to obtain the content identifier included in the data change message.

[0088] The location identifier may be the identifier of the data table where the data change occurs and the data change message is generated. The target location identifier may be the identifier of the data table where the recommended content sample is located.

[0089] For example, specifically, the location identifier in the data change message can be obtained. If the location identifier is consistent with the target location identifier, the data change message is determined to be the data change message corresponding to the recommended content sample, and the data structure of the data update message is parsed to determine whether the recommended content sample is the recommended content in the content recommendation list.

[0090] Optionally, identifying abnormal content in recommended content samples and determining the content type of recommended content in the content recommendation list can be implemented through two systems. The review system is responsible for identifying abnormal content in recommended content samples, and the data processing system is responsible for determining the content type of recommended content. The review system and the data processing system are responsible for transmitting data change messages through message middleware.

[0091] For example, the audit system can identify abnormal content in recommended content samples in the database and determine the content type of the recommended content sample. If the content type of the recommended content sample is abnormal, the corresponding audit status is set to "failed." Binlog generates data change messages based on updates to the content type or audit status of the recommended content sample and sends them to a messaging middleware such as Kafka or Pulsar.

[0092] The data processing system obtains the data update message from the message middleware, and determines whether it is a data change message corresponding to the modification of the review status of the recommended content sample based on the location identifier contained in the data update message. If so, the data structure of the data update message is parsed to determine whether it is an event of modifying the review status to fail. If so, the content identifier is used to determine whether it is a recommended content in the content recommendation list.

[0093] In one embodiment, an administrator may manage the content recommendation list on a management platform corresponding to the content recommendation method. For example, the administrator may input an object identifier of the recommended content through the management platform, obtain the object identifier input by the administrator, and determine the content type of the corresponding recommended content as an abnormal content type. The content recommendation method may further include:

[0094] Get the object ID of the recommended content in the content recommendation list;

[0095] Determine the corresponding recommended content in the content recommendation list according to the object identifier;

[0096] The content type of the recommended content is determined to be an abnormal content type.

[0097] Among them, the object identifier can be an identifier related to the recommended content, and the recommended content can be located according to the object identifier. For example, it can be the above-mentioned content identifier, or the account ID of the author who published the recommended content. According to the account ID, all recommended content published by the user can be queried.

[0098] For example, the object identifier of the recommended content input by the user can be obtained. For example, if an administrator finds that a recommended content of an abnormal content type exists in the content recommendation list, the administrator can input the object identifier of the recommended content of the abnormal content type through the management platform corresponding to the content recommendation method. Based on the object identifier, the corresponding recommended content in the content recommendation list is determined, and the content type of the recommended content is determined to be an abnormal content type.

[0099] 103. When the content type is a specified content type, obtain a status list corresponding to the content recommendation list, where the status list includes recommendation statuses of the recommended contents in the content recommendation list.

[0100] For example, specifically, when the content type is a specified content type, a status list corresponding to the content recommendation list is obtained.

[0101] In order to strengthen the management of recommended content in the content recommendation list, target content identification can be performed on the recommended content in the content recommendation list. The specific process can refer to the process of target content identification for recommended content samples, which will not be described in detail here. Target content identification can also be performed on newly recommended content in the content recommendation list compared to the historical recommendation list to improve the recognition efficiency of recommended content in the content recommendation list. That is, before the step of "when the content type is a specified content type, obtaining a status list corresponding to the content recommendation list", the following can also be included:

[0102] Get a list of historical content recommendations;

[0103] Determine new recommended content based on the historical content recommendation list and the content recommendation list;

[0104] Target content recognition is performed on the newly added content to obtain the content type of the newly added recommended content.

[0105] For example, the content recommendation list may be updated once a day, and the historical content recommendation list of the previous day may be obtained to determine the newly added recommended content in the content recommendation list, and target content recognition may be performed on the newly added recommended content in the content recommendation list to obtain the content type of the newly added recommended content.

[0106] 104. Perform corresponding adjustment processing on the recommendation status of the recommended content in the content recommendation list to obtain an adjusted status list.

[0107] For example, specifically, when the recommendation status in the status list is recommended, when the content type is a specified content type, the recommendation status of the recommended content is changed to not recommended, and an adjusted status list is obtained based on the changed recommendation status.

[0108] In one embodiment, the object identifier may be a common identifier for recommended content in content recommendation lists in different scenarios. For example, each recommended content in the content recommendation list generally corresponds to a unique identifier, namely, a content identifier, and identifiers such as the account ID of the author who published the recommended content. Therefore, corresponding status lists may be set for different types of object identifiers so that the recommendation status of the corresponding recommended content can be quickly modified according to the object identifier. That is, the step of "adjusting the recommendation status of the recommended content in the status list to correspond to the specified content type to obtain an adjusted status list" may specifically include:

[0109] Determine the corresponding status list according to the type of object identification;

[0110] For the corresponding status list, abnormal adjustment processing is performed on the recommended status of the recommended content to obtain an adjusted status list.

[0111] The object identifier may include at least two types of object identifiers, for example, a content identifier and an author's account ID, or other identifiers may be set according to the characteristics of the scene, which is not limited here.

[0112] Among them, the status list can be a status list set according to different types of object identifiers, for example, a status list with content identifier as the association key with the content recommendation list, and a status list with the account ID of the author who published the recommended content as the association key with the content recommendation list.

[0113] For example, the corresponding status list may be determined based on the type of object identifier, and the recommended status of the recommended content in the corresponding status list may be abnormally adjusted. For example, if the object identifier is a content identifier, the corresponding status list may be determined to be a status list with the content identifier as the associated key, and based on the object identifier, the recommended status of the corresponding recommended content may be abnormally adjusted in the status list with the content identifier as the associated key. If the object identifier is a user ID, the corresponding status list may be determined to be a status list with the user ID as the associated key, and based on the object identifier, the recommended status of the recommended content published by the user ID may be abnormally adjusted in the status list with the content identifier as the associated key. This allows abnormal adjustment of the recommended status of multiple recommended contents at once.

[0114] It can be understood that the content recommendation list is based on the ranking order number as the sorting primary key. Since the content recommendation list usually includes the content identifier corresponding to the recommended content and the author's account ID, the content identifier and account ID can be used as the association key with the content recommendation list to set two corresponding status lists. The content identifier and account ID are used as common fields of the content recommendation list. Therefore, the status list can be more easily connected to the content recommendation list in different scenarios to recommend the recommended content in the content recommendation list and improve the content recommendation effect.

[0115] Optionally, when abnormal adjustment is performed on the recommended status corresponding to the recommended content in the status list, the time of the abnormal adjustment, the reason for the abnormal adjustment (for example, containing misleading text or containing illegal controls, etc.), and the abnormal adjustment operation method (manual or system) can be recorded.

[0116] 105. Perform content recommendation processing on the content recommendation list according to the adjusted status list.

[0117] For example, it can be determined whether to recommend the recommended content in the content recommendation list based on the recommendation status of each recommended content in the adjusted status list. If the recommendation status is a recommendable status, the recommended content is recommended. If the recommendation status is a non-recommended status, the recommended content is not recommended, and the recommended content is not included in the content recommendation list viewed by the user.

[0118] Optionally, the content recommendation list may be adjusted according to the status list, and then recommendations may be made based on the adjusted content recommendation list, i.e., the step of "processing content recommendations on the content recommendation list according to the adjusted status list" may specifically include:

[0119] Get the recommended content corresponding to the recommended status in the status list;

[0120] Sort the recommended content corresponding to the recommendable status according to preset rules to obtain an adjusted content recommendation list;

[0121] Content recommendation is performed based on the adjusted content recommendation list.

[0122] For example, the content recommendation list may be specifically sorted for the recommended content in the recommended state to obtain an adjusted content recommendation list, that is, the adjusted content recommendation list does not include the recommended content in the not recommended state, and recommendation processing is performed based on the adjusted content recommendation list.

[0123] As can be seen from the above, the embodiment of the present application obtains a data change message for the recommended content in the content recommendation list, the content recommendation list includes at least one recommended content sorted according to a preset rule, and the data change message includes a message generated by target content identification for the recommended content; the data change message is parsed to obtain the content type of the recommended content in the content recommendation list; when the content type is a specified content type, a status list corresponding to the content recommendation list is obtained, the status list includes the recommended status of the recommended content in the content recommendation list; the recommended status of the recommended content in the content recommendation list is adjusted accordingly to obtain an adjusted status list; and content recommendation processing is performed on the content recommendation list based on the adjusted status list. This scheme determines the content type of the recommended content based on the message generated by target content identification for the recommended content in the content recommendation list, and adjusts the status list accordingly. It can improve the effect of content recommendation by not recommending the recommended content in the content recommendation list that belongs to the specified content type.

[0124] Based on the above embodiments, further detailed description will be given below with examples.

[0125] This embodiment will describe the recommended content from the perspective of a content recommendation system, where the recommended content is presented in the form of pictures or videos. The content recommendation system may specifically include an audit system and a data processing system.

[0126] The present application provides a content recommendation method, such as Figure 3 As shown, the specific process of the content recommendation method can be as follows:

[0127] 201. The review system obtains a sample of recommended content from the database.

[0128] For example, the review system may obtain recommended content samples from a database, and optionally, may also obtain recommended content samples from a blockchain.

[0129] 202. The review system identifies abnormal content in the recommended content sample and changes the corresponding review status to failed review based on the recognition result.

[0130] For example, specifically for recommended content presented in the form of a video, the review system can obtain the key frame image or cover image of the video, perform optical character recognition (OCR) on the key frame image or cover image, identify the text information contained in the key frame image or cover image, and compare the text information with the preset text information. If there is a match, the content type of the recommended content is determined to be an abnormal content type, and the corresponding review status is changed to review failure; if there is no match, the recommended content sample is identified through the abnormal content recognition model to identify whether the recommended content sample contains illegal controls. If it contains illegal controls, the content type of the recommended content is determined to be an abnormal content type, and the corresponding review status is changed to review failure. If it does not contain illegal controls, the content type of the recommended content is determined to be a normal content type, and the review status is not modified.

[0131] For identifying abnormal content in recommended content presented in the form of images, you can refer to the above method of identifying abnormal content in key frame images or cover images, which will not be repeated here.

[0132] Optionally, the recommended content samples may be screened to select the recommended content in the content recommendation list from the recommended content samples, and only the recommended content in the content recommendation list may be identified as abnormal.

[0133] The key frame image may be an image used as a video cover, or a video frame at the end of a video.

[0134] 203. The binlog in the audit system generates a data update message based on the audit status of the recommended content sample, and sends the data update message to the message middleware.

[0135] For example, the binlog in the audit system may generate data update messages based on the audit status of the recommended content samples, and push the data update messages to the message middleware, such as Kafka or Pulsar.

[0136] 204. The data processing system obtains the data update message from the message middleware.

[0137] For example, the data processing system may subscribe to data update messages from the message middleware, or receive data update messages pushed by the message middleware.

[0138] The review system and the data processing system can be decoupled through message middleware. There is no need for communication between the two systems. The review system generates binlog data change messages, and the data processing system detects binlog data change messages to complete the real-time update of the recommendation status of the recommended content.

[0139] 205. The data processing system determines whether the location identifier included in the data update message is the target location identifier. If so, execute step 206. Otherwise, the content recommendation method flow continues.

[0140] For example, the data corresponding to the content recommendation sample in the database (including the review status) can be saved in a specified data table (the corresponding identifier is the target location identifier), and the data update message can include the identifier of the data table where the data update occurs (i.e., the location identifier). If the location identifier is the target location identifier, the data processing system determines that the data update message is a data change message corresponding to the recommended content sample and executes step 206. Otherwise, the data change message is irrelevant to the recommended content sample, that is, it is irrelevant to the recommended content in the recommendation list, and the process exits.

[0141] 206. The data processing system parses the data structure of the data update message to determine whether it is an audit status failure event. If so, execute step 207. If not, end the content recommendation method process.

[0142] For example, the data processing system may parse the data structure of the data update message, obtain the processing event corresponding to the data update message, and determine whether the data update message corresponds to a review status failure event. If so, the data processing system executes step 207; if not, the content recommendation method process ends.

[0143] 207. The data processing system parses the data structure of the data update message to obtain a content identifier of the recommended content sample.

[0144] For example, the data processing system may parse the data structure of the data update message and extract the content identifier of the recommended content sample from the data update message.

[0145] 208. The data processing system determines whether the recommended content sample with updated review status is recommended content in the content recommendation list based on the content identifier. If so, execute step 209. If not, end the content recommendation method process.

[0146] For example, the data processing system may specifically determine whether the recommended content sample with updated review status is recommended content in the content recommendation list based on the content identifier. If so, the data processing system executes step 209; otherwise, the content recommendation method process ends.

[0147] 209. The data processing system modifies the recommendation status of the corresponding recommended content in the status list to a non-recommended status according to the content identifier, so that the client updates the displayed recommendation list based on the modified recommendation status.

[0148] For example, the data processing system may query the recommendation status of the corresponding recommended content from the status list according to the content identifier, modify the recommended status to the non-recommended status, and obtain an adjusted status list. The client displays the corresponding recommended content based on the adjusted status list and the content recommendation list, and can realize real-time offline processing of the recommended content belonging to the abnormal content type displayed on the client.

[0149] Optionally, the recommendation status is modified and the time of the abnormal adjustment, the reason for the abnormal adjustment (such as the inclusion of misleading text or illegal controls, etc.), and the abnormal adjustment operation method (system), etc. are recorded.

[0150] In addition to determining the recommendation status of the recommended content in the content recommendation list based on the data change message, it is also possible to simultaneously perform target content identification on the recommended content in the content recommendation list to determine the recommendation status of the recommended content. The specific process can be referred to the description of steps 301-304 of the following embodiment and will not be repeated here.

[0151] As can be seen from the above, the audit system of the embodiment of the present application obtains a recommended content sample from the database; the audit system identifies abnormal content on the recommended content sample, and changes the corresponding audit status to audit failure according to the identification result. Binlog generates a data update message according to the audit status update of the recommended content sample, and sends the data update message to the message middleware. The data processing system obtains the data update message from the message middleware, and determines whether it is a data change message corresponding to the recommended content sample according to the location identifier contained in the data update message. If so, the data processing system parses the data structure of the data update message to determine whether it is an audit failure event. If so, the data structure of the data update message is parsed to obtain the content identifier of the recommended content sample, and determines whether the recommended content sample with the updated audit status is a recommended content in the content recommendation list according to the content identifier. If so, the recommended status of the corresponding recommended content in the status list is modified to a non-recommended status according to the content identifier, so that the client updates the displayed recommendation list based on the modified recommendation status. This solution identifies abnormal content on the recommended content in the content recommendation list and determines the recommended status of the recommended content according to the identification result. It can not recommend the recommended content in the content recommendation list that belongs to the abnormal content type, thereby improving the effect of content recommendation.

[0152] Based on the above embodiments, further detailed description will be given below with examples.

[0153] In this embodiment, from the perspective of a content recommendation device, the recommended content is content presented in the form of pictures or videos, and the content recommendation list is described as a content ranking list. The content recommendation device can be specifically integrated into a computer device, which can be a server or other device.

[0154] The present application provides a content recommendation method, such as Figure 4 As shown, the specific process of the content recommendation method can be as follows:

[0155] 301. The server determines recommended content to be identified based on the content ranking list and the historical content ranking list.

[0156] The historical ranking list may be a content ranking list of a previous ranking period of the content ranking list. For example, if the ranking period is one day and the content ranking list is today's ranking list, then the historical ranking list may be yesterday's content ranking list.

[0157] For example, the server may obtain a content ranking list and a historical content ranking list, compare the content ranking list with the historical ranking list, and determine the recommended content in the content ranking list that is different from the historical ranking list as the recommended content to be identified.

[0158] 302. The server performs text information recognition and illegal control recognition on the recommended content to be recognized, and determines the content type of the recommended content to be recognized based on the recognition results.

[0159] The specific implementation process of text information identification and illegal control identification for the recommended content to be identified can be referred to the description in the previous embodiment and will not be repeated here.

[0160] 303. If the content type of the recommended content to be identified is an abnormal content type, the server modifies the recommendation status corresponding to the content to be identified in the status list to a non-recommended status, thereby obtaining an adjusted status list.

[0161] For example, if the content type of the recommended content to be identified is an abnormal content type, the server modifies the recommendation status corresponding to the recommended content to be identified in the status list to a non-recommended status, and obtains an adjusted status list.

[0162] 304. The server performs recommendation processing on the content ranking list based on the adjusted status list.

[0163] For example, if the recommended status of a recommended content in the adjusted status list is "recommended," the server will recommend it. If the recommended status of a recommended content in the adjusted status list is "not recommended," the server will not recommend it. When users view content ranking data on the client, they will not see recommended content of unusual content types, only recommended content of normal content types. This reduces the appearance of illegal recommended content on the rankings and improves the effectiveness of content recommendations.

[0164] It is understandable that in addition to identifying abnormal content for the recommended content to be identified, abnormal content identification can also be performed on other recommended content on the content ranking list to avoid other recommended content on the content ranking list being recommended content of abnormal content types, thereby enhancing the reliability of the content ranking list.

[0165] It is understandable that the server can also obtain data update messages from the message middleware, determine the recommendation status of the recommended content in the content recommendation list based on the data update messages, and determine the recommendation status of the recommended content in the content recommendation list based on the object identifier input by the user.

[0166] As can be seen from the above, the server of the embodiment of the present application determines the recommended content to be identified based on the content ranking list and the historical content ranking list, performs text information identification and illegal control identification on the recommended content to be identified, and determines the content type of the recommended content to be identified based on the identification results. If the content type of the recommended content to be identified is an abnormal content type, the server modifies the corresponding recommendation status of the content to be identified in the status list to a non-recommended status, obtains an adjusted status list, and performs recommendation processing on the content ranking list based on the adjusted status list. This solution identifies abnormal content in the recommended content in the content recommendation list and determines the recommended status of the recommended content based on the identification results. It can not recommend the recommended content in the content recommendation list that belongs to the abnormal content type, thereby improving the effect of content recommendation.

[0167] In order to facilitate better implementation of the content recommendation method provided in the embodiment of the present application, a content recommendation device is also provided in one embodiment. The meanings of the terms are the same as those in the above-mentioned content recommendation method, and the specific implementation details can be referred to the description in the method embodiment.

[0168] The content recommendation device can be integrated into a computer device, such as Figure 5 As shown, the content recommendation device may include: a message acquisition unit 401, a parsing unit 402, a list acquisition unit 403, an adjustment unit 404 and a recommendation unit 405, specifically as follows:

[0169] (1) Message acquisition unit 401: used to acquire a data change message for a recommended content in a content recommendation list, wherein the content recommendation list includes at least one recommended content sorted according to a preset rule, and the data change message includes a message generated by identifying a target content for the recommended content.

[0170] Optionally, the content recommendation device may further include a sample acquisition subunit, a first identification subunit, a message generation subunit, and a first type determination subunit. Specifically:

[0171] Sample acquisition unit: used to obtain recommended content samples from the database, where the recommended content samples include recommended content in the content recommendation list;

[0172] Identification unit: used to identify target content of the recommended content sample and obtain the content type of the recommended content sample, where the recommended content sample includes the recommended content in the content recommendation list;

[0173] Message generation unit: used to generate data change messages based on the content type of the recommended content sample.

[0174] Optionally, the content recommendation device may further include a sending unit, specifically:

[0175] Sending unit: used to send data change messages to the message queue;

[0176] The message acquisition unit 401 may also be configured to: acquire data change messages for recommended content in the content recommendation list from the message queue.

[0177] Optionally, the content recommendation device may further include a data acquisition unit and a creation unit, specifically:

[0178] A data acquisition unit, configured to acquire a recommendation status and an associated content identifier of a recommended content in a content recommendation list;

[0179] The establishing unit is used to establish a status list according to the associated content identifier and the recommendation status of the recommended content, and the status list is associated with the content recommendation list through the associated content identifier.

[0180] Optionally, the identification unit may include a second identification subunit and a second type determination subunit, specifically:

[0181] The second identification subunit is used to identify text information of the recommended content sample and obtain the text information contained in the recommended content sample;

[0182] The second type determination subunit is configured to determine the content type of the recommended content sample according to the text information.

[0183] Optionally, the identification unit may include a second acquisition subunit, a matching subunit, and a third type determination subunit, specifically:

[0184] The second acquisition subunit is used to acquire reference content corresponding to a preset content type;

[0185] Matching subunit: used to match the reference content with the recommended content sample to obtain the matching result between the reference content and the recommended content;

[0186] The third type determination subunit is configured to determine the content type of the recommended content sample according to the matching result.

[0187] Optionally, the identification unit may include an extraction subunit and a fourth type determination subunit, specifically:

[0188] Extraction subunit: used to extract content features of recommended content samples to obtain content feature information of recommended content;

[0189] The fourth type determination subunit is configured to predict the content type of the recommended content sample based on the content feature information.

[0190] (2) The parsing unit 402 parses the data change message to obtain the content type of the recommended content in the content recommendation list.

[0191] Optionally, the parsing unit 402 may include a structure parsing subunit and a determination subunit, specifically:

[0192] Structure parsing subunit: used to parse the data structure of the data change message and obtain the content identifier corresponding to the data change message;

[0193] Determining subunit: for determining that the content type of the recommended content is a specified content type if the content identifier matches the recommended content in the content recommendation list.

[0194] Optionally, the structure analysis subunit may include an acquisition module and an analysis module, specifically:

[0195] Acquisition module: used to obtain the location identifier carried in the data change message;

[0196] Parsing module: used for parsing the data structure of the data change message if the location identifier is the target location identifier, and obtaining the content identifier contained in the data change message.

[0197] The content recommendation device may further include an identification acquisition unit, a content determination unit, and a type determination unit. Specifically:

[0198] Identification acquisition unit: used to obtain the object identification of the recommended content in the content recommendation list;

[0199] Content determination unit: used to determine the corresponding recommended content in the content recommendation list according to the object identifier;

[0200] Type determination unit: used to determine the content type of the recommended content as an abnormal content type.

[0201] (3) List acquisition unit 403: used to acquire a status list corresponding to the content recommendation list when the content type is a specified content type, where the status list includes the recommendation status of the recommended content in the content recommendation list.

[0202] Optionally, the content recommendation device may further include a history list acquisition unit, a new content determination unit, and a new content identification unit. Specifically:

[0203] History list acquisition unit: used to obtain a history content recommendation list;

[0204] New content determination unit: used to determine new recommended content based on the historical content recommendation list and the content recommendation list;

[0205] New content identification unit: used to identify the target content of the new content and obtain the content type of the new recommended content.

[0206] (4) Adjustment unit 404: used to adjust the recommended status of the recommended content in the status list accordingly to obtain an adjusted status list. Optionally, the adjustment unit 404 may include a list determination subunit and a first adjustment subunit, specifically:

[0207] List determination subunit: used to determine the corresponding state list according to the object identifier;

[0208] Adjustment subunit: used to perform abnormal adjustment processing on the recommendation status of the recommended content in the corresponding status list to obtain an adjusted status list.

[0209] (5) Recommendation unit 405: configured to perform content recommendation processing on the content recommendation list according to the adjusted status list.

[0210] Optionally, the recommendation unit 405 may include a third acquisition subunit, a ranking subunit, and a recommendation subunit, specifically:

[0211] The third acquisition subunit is used to acquire the recommended content corresponding to the recommendable status in the status list;

[0212] Sorting subunit: used to sort the recommended content corresponding to the recommendable status according to preset rules to obtain an adjusted content recommendation list;

[0213] Recommendation subunit: used to make content recommendations based on the adjusted content recommendation list.

[0214] In the content recommendation device of the present embodiment, a message acquisition unit 401 acquires a data change message for recommended content in a content recommendation list. The content recommendation list includes at least one recommended content sorted according to a preset rule. The data change message includes a message generated by performing target content identification on the recommended content. A parsing unit 402 parses the data change message to obtain the content type of the recommended content in the content recommendation list. When the content type is a specified content type, a list acquisition unit 403 acquires a status list corresponding to the content recommendation list. The status list includes the recommended status of the recommended content in the content recommendation list. An adjustment unit 404 adjusts the recommended status of the recommended content in the content recommendation list accordingly to obtain an adjusted status list. A recommendation unit 405 performs content recommendation on the content recommendation list based on the adjusted status list. This solution determines the content type of the recommended content based on the message generated by performing target content identification on the recommended content in the content recommendation list and adjusts the status list accordingly. This solution can de-recommend recommended content in the content recommendation list that belongs to the specified content type, thereby improving the effectiveness of content recommendation.

[0215] The embodiment of the present application also provides a computer device, which can be a terminal or a server. Figure 6 , which shows a schematic diagram of the structure of the computer device involved in the embodiment of the present application, specifically:

[0216] The computer device may include one or more processing core processors 1001, one or more computer readable storage media memories 1002, a power supply 1003, an input unit 1004 and other components. Those skilled in the art will understand that Figure 6 The computer device structure shown in the figure does not constitute a limitation on the computer device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange components differently.

[0217] Processor 1001 is the control center of the computer device. It connects the various components of the entire computer device using various interfaces and lines. By running or executing software programs and / or modules stored in memory 1002 and accessing data stored in memory 1002, it performs various functions of the computer device and processes data, thereby performing overall testing of the computer device. Optionally, processor 1001 may include one or more processing cores; preferably, processor 1001 may integrate an application processor and a modem processor, wherein the application processor primarily processes the operating system, user interface, and computer programs, while the modem processor primarily handles wireless communications. It is understood that the modem processor may not be integrated into processor 1001.

[0218] The memory 1002 can be used to store software programs and modules. The processor 1001 executes various functional applications and data processing by running the software programs and modules stored in the memory 1002. The memory 1002 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, a computer program required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 1002 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage device. Accordingly, the memory 1002 may also include a memory controller to provide the processor 1001 with access to the memory 1002.

[0219] The computer device also includes a power supply 1003 for supplying power to various components. Preferably, the power supply 1003 can be logically connected to the processor 1001 via a power management system, thereby enabling the power management system to manage charging, discharging, and power consumption. The power supply 1003 can also include one or more DC or AC power supplies, a recharging system, a power failure detection circuit, a power converter or inverter, a power status indicator, and other arbitrary components.

[0220] The computer device may further include an input unit 1004, which may be configured to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0221] Although not shown, the computer device may further include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 1001 in the computer device will load the executable files corresponding to one or more computer program processes into the memory 1002 according to the following instructions, and the processor 1001 will run the computer programs stored in the memory 1002 to implement various functions as follows:

[0222] Obtaining a data change message for a recommended content in a content recommendation list, where the content recommendation list includes at least one recommended content sorted according to a preset rule, and the data change message includes a message generated by identifying a target content for the recommended content;

[0223] Parse the data change message to obtain the content type of the recommended content in the content recommendation list;

[0224] When the content type is a specified content type, obtaining a status list corresponding to the content recommendation list, the status list including the recommendation status of the recommended content in the content recommendation list;

[0225] Adjust the recommendation status of the recommended content in the content recommendation list accordingly to obtain an adjusted status list;

[0226] Perform content recommendation processing on the content recommendation list according to the adjusted status list.

[0227] The specific implementation of the above operations can be found in the previous embodiments and will not be described in detail here.

[0228] As can be seen from the above, the computer device of the embodiment of the present application can obtain data change messages for recommended content in a content recommendation list, the content recommendation list includes at least one recommended content sorted according to preset rules, and the data change message includes a message generated by target content identification for the recommended content; the data change message is parsed and processed to obtain the content type of the recommended content in the content recommendation list; when the content type is a specified content type, a status list corresponding to the content recommendation list is obtained, the status list includes the recommended status of the recommended content in the content recommendation list; the recommended status of the recommended content in the content recommendation list is adjusted accordingly to obtain an adjusted status list; content recommendation processing is performed on the content recommendation list based on the adjusted status list. This scheme determines the content type of the recommended content based on the message generated by target content identification for the recommended content in the content recommendation list, and adjusts the status list accordingly. It can improve the effect of content recommendation by not recommending the recommended content in the content recommendation list that belongs to the specified content type.

[0229] According to one aspect of the present application, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in various optional implementations of the above embodiments.

[0230] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be accomplished by a computer program, or by controlling related hardware through a computer program. The computer program may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0231] To this end, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. The computer program can be loaded by a processor to execute any content recommendation method provided in the embodiment of the present application.

[0232] The specific implementation of the above operations can be found in the previous embodiments and will not be repeated here.

[0233] The computer-readable storage medium may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0234] Since the computer program stored in the computer-readable storage medium can execute any of the content recommendation methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any of the content recommendation methods provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0235] The above is a detailed introduction to a content recommendation method, device, computer equipment and computer-readable storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, based on the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A content recommendation method, characterized in that: include: Obtaining a recommended content sample from a database, performing target content recognition on the recommended content sample, and obtaining a content type of the recommended content sample; a data change message generated based on the content type of the recommended content sample; Obtaining a data change message for a recommended content in a content recommendation list, wherein the content recommendation list includes at least one recommended content sorted according to a preset rule; Parsing the data change message to obtain the content type of the recommended content in the content recommendation list; When the content type is a specified content type, obtaining a status list corresponding to the content recommendation list, the status list including the recommendation status of the recommended content in the content recommendation list; Performing corresponding adjustment processing on the recommendation status of the recommended content in the content recommendation list to obtain an adjusted status list; Perform content recommendation processing on the content recommendation list according to the adjusted status list.

2. The method according to claim 1, characterized in that The parsing of the data change message to obtain the content type of the recommended content in the content recommendation list includes: Parsing the data structure of the data change message to obtain a content identifier included in the data change message, including: obtaining a location identifier in the data change message, the location identifier including a location identifier of a storage location where the data change occurs in a database; if the location identifier is a target location identifier, parsing the data structure of the data change message to obtain a content identifier included in the data change message; If the content identifier matches the recommended content in the content recommendation list, it is determined that the content type of the recommended content is a specified content type.

3. The method according to claim 1, characterized in that The method further comprises: Sending the data change message to the message queue; The acquiring of the data change message for the recommended content in the content recommendation list includes: A data change message for the recommended content in the content recommendation list is obtained from the message queue.

4. The method according to claim 1, wherein The performing target content identification on the recommended content sample to obtain the content type of the recommended content sample includes: Performing text information recognition on the recommended content sample to obtain text information contained in the recommended content; The content type of the recommended content sample is determined according to the text information.

5. The method according to claim 1, wherein The performing target content identification on the recommended content sample to obtain the content type of the recommended content sample includes: Get reference content corresponding to the preset content type; Performing content matching on the reference content and the recommended content sample to obtain a matching result between the reference content and the recommended content sample; The content type of the recommended content sample is determined according to the matching result.

6. The method according to claim 1, characterized in that The step of identifying abnormal content on the recommended content in the content recommendation list to obtain the content type of the recommended content sample includes: Extracting content features of the recommended content in the content recommendation list to obtain content feature information of the recommended content; The content type of the recommended content is predicted according to the content feature information.

7. The method according to claim 1, characterized in that When the content type is a specified content type, before obtaining the status list corresponding to the content recommendation list, the method further includes: Get a list of historical content recommendations; Determining new recommended content based on the historical content recommendation list and the content recommendation list; Target content identification is performed on the newly added recommended content to obtain the content type of the newly added recommended content.

8. The method according to claim 1, characterized in that Before obtaining the data change message for the recommended content in the content recommendation list, the method further includes: Obtaining the recommendation status and associated content identifier of the recommended content in the content recommendation list; A status list is established according to the associated content identifier and the recommendation status of the recommended content, and the status list is associated with the content recommendation list through the associated content identifier.

9. The method according to any one of claims 1 to 8, characterized in that The recommendation status includes a recommendable status and a non-recommended status, and performing content recommendation processing on the content recommendation list according to the adjusted status list includes: Obtain the recommended content corresponding to the recommended status in the status list; Sorting the recommended content corresponding to the recommendable status according to preset rules to obtain an adjusted content recommendation list; Content recommendation is performed based on the adjusted content recommendation list.

10. A content recommendation device, characterized in that: include: A sample acquisition unit, used to acquire recommended content samples from a database; an identification unit, configured to identify target content of the recommended content sample and obtain a content type of the recommended content sample; a message generating unit, configured to generate a data change message based on the content type of the recommended content sample; A message acquisition unit, configured to acquire a data change message for a recommended content in a content recommendation list, wherein the content recommendation list includes at least one recommended content sorted according to a preset rule; A parsing unit, configured to parse the data change message to obtain the content type of the recommended content in the content recommendation list; a list acquisition unit, configured to acquire a status list corresponding to the content recommendation list when the content type is a specified content type, the status list including the recommendation status of the recommended content in the content recommendation list; An adjusting unit, configured to adjust the recommendation status of the recommended content in the content recommendation list accordingly to obtain an adjusted status list; A recommendation unit is configured to perform content recommendation processing on the content recommendation list according to the adjusted status list.

11. A computer device, characterized in that: The invention comprises a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the content recommendation method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that The computer-readable storage medium is used to store a computer program, and the computer program is loaded by a processor to execute the content recommendation method according to any one of claims 1 to 9.

13. A computer program product, characterized in that The computer program product includes computer instructions, which are stored in a computer-readable storage medium; a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the content recommendation method described in any one of claims 1 to 9.

Images