Resource recommendation method and device, storage medium and computer equipment

Abstract

According to the resource recommendation method and device, the storage medium and the computer equipment provided by the invention, when the resources are recommended to the target user, the user tag of the target user and the multiple target resources to be recommended can be determined firstly, and then the picture features corresponding to the pictures uploaded by the target user are obtained; in this way, the total recommendation value of each target resource can be calculated based on the user tag, the picture feature and the file feature, and finally the final resource recommended to the target user is determined according to the total recommendation value of each target resource. In the process, the recommended resources can be directly determined according to the picture features, the text features and the like uploaded by the user without manually tagging the user or tagging the target resources needing to be pushed, so that intelligent recommendation is performed on the user according to the preference of the user, the requirement of the user for obtaining personalized resources is met, the user rights and interests are guaranteed, and the user experience is improved. And the push opening rate and the resource subscription rate are improved.

Description

Technical Field

[0001] The present application relates to the field of emerging information technology, and in particular to a resource recommendation method, device, storage medium and computer equipment. Background Art

[0002] With the rapid development of the Internet era, information is growing exponentially. In the case of great pressure on mobile phone storage, cloud storage service has emerged as an important solution. At present, with the increase in the number of users and user activity, the file storage volume is getting larger and larger. It is imperative to provide users with personalized services, enhance differentiated competitive advantages, and recommend resources based on user preferences.

[0003] In the prior art, when recommending resources to users in resource plazas, most of them are based on simple collaborative filtering or content recommendation algorithms. These algorithms often ignore the individual differences between different users, and thus cannot effectively and accurately recommend resources that meet their personalized needs to users. Summary of the invention

[0004] The purpose of this application is to solve at least one of the above-mentioned technical defects, especially the technical defect that the prior art often ignores the individual differences between different users when recommending resources to users, and thus cannot effectively and accurately recommend resources that meet their personalized needs to users.

[0005] The present application provides a resource recommendation method, the method comprising:

[0006] When recommending resources to a target user, determining a user tag of the target user and a plurality of target resources to be recommended;

[0007] Acquire image features corresponding to the image uploaded by the target user, and file features corresponding to the file uploaded by the target user;

[0008] Calculate the total recommendation value of each target resource based on the user tag, the image feature and the file feature;

[0009] The final resource recommended to the target user is determined according to the total recommendation value of each target resource.

[0010] Optionally, the calculating the total recommendation value of each target resource based on the user tag, the image feature, and the file feature includes:

[0011] Determining a first recommendation value of the user tag relative to each target resource according to the similarity between the user tag and each target resource;

[0012] Determining, according to the similarity between the image feature and each target resource, a second recommendation value of the image feature relative to each target resource;

[0013] Determining a third recommendation value of the file feature relative to each target resource according to the similarity between the file feature and each target resource;

[0014] Based on the first recommendation value, the second recommendation value, and the third recommendation value corresponding to each target resource, the total recommendation value of each target resource is calculated.

[0015] Optionally, when there are multiple user tags, determining the first recommendation value of the user tag relative to each target resource according to the similarity between the user tag and each target resource includes:

[0016] For each target resource, calculate the similarity between each user tag and the target resource, and select at least one user tag whose similarity is greater than a first similarity threshold from each user tag as an initial user tag;

[0017] Sorting the initial user tags according to the similarity, and selecting a first preset number of initial user tags from the initial user tags as target user tags according to the first sorting result;

[0018] According to the similarity between each target user tag and the target resource, a first recommendation value of the user tag relative to the target resource is calculated.

[0019] Optionally, when there are multiple picture features, determining the second recommendation value of the picture feature relative to each target resource according to the similarity between the picture feature and each target resource includes:

[0020] For each target resource, calculate the similarity between each image feature and the target resource, and select at least one image feature whose similarity is greater than a second similarity threshold from each image feature as an initial image feature;

[0021] Sorting the initial picture features according to their similarity, and selecting a second preset number of initial picture features from the initial picture features as target picture features according to the second sorting result;

[0022] According to the similarity between each target image feature and the target resource, a second recommendation value of the image feature relative to the target resource is calculated.

[0023] Optionally, when there are multiple file features, determining the third recommendation value of the file feature relative to each target resource according to the similarity between the file feature and each target resource includes:

[0024] For each target resource, calculate the similarity between each file feature and the target resource, and select at least one file feature whose similarity is greater than a third similarity threshold from each file feature as an initial file feature;

[0025] Sorting the initial file features according to the similarity, and selecting a third preset number of initial file features from the initial file features as target file features according to the third sorting result;

[0026] According to the similarity between each target file feature and the target resource, a third recommendation value of the file feature relative to the target resource is calculated.

[0027] Optionally, the calculating the total recommendation value of each target resource based on the first recommendation value, the second recommendation value, and the third recommendation value corresponding to each target resource includes:

[0028] Determine a first weight corresponding to the user tag, a second weight corresponding to the image feature, and a third weight corresponding to the text feature;

[0029] For each target resource, the total recommendation value of the target resource is calculated based on the first recommendation value and first weight of the user tag relative to the target resource, the second recommendation value and second weight of the image feature relative to the target resource, and the third recommendation value and third weight of the text feature relative to the target resource.

[0030] Optionally, determining the final resource recommended to the target user according to the total recommendation value of each target resource includes:

[0031] Selecting at least one target resource whose total recommendation value is greater than a preset recommendation threshold from each target resource as a final resource recommended to the target user;

[0032] Alternatively, at least one target resource whose total recommendation value is greater than a preset recommendation threshold is selected from each target resource as a recommendable resource;

[0033] The recommendable resources are sorted according to the total recommendation value, and a fourth preset number of recommendable resources are selected from the recommendable resources according to the fourth sorting result as final resources recommended to the target user.

[0034] The present application also provides a resource recommendation device, comprising:

[0035] A data acquisition module, used to determine the user tag of the target user and a plurality of target resources to be recommended when recommending resources to the target user;

[0036] A feature acquisition module, used to acquire image features corresponding to the image uploaded by the target user, and file features corresponding to the file uploaded by the target user;

[0037] A recommendation value calculation module, used to calculate the total recommendation value of each target resource based on the user tag, the image feature and the file feature;

[0038] The resource recommendation module is used to determine the final resource recommended to the target user according to the total recommendation value of each target resource.

[0039] The present application also provides a computer-readable storage medium, in which computer-readable instructions are stored. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the resource recommendation method described in any one of the above embodiments.

[0040] The present application also provides a computer device, comprising: one or more processors, and a memory;

[0041] The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the steps of the resource recommendation method described in any one of the above embodiments are performed.

[0042] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0043] The resource recommendation method, device, storage medium and computer equipment provided by the present application can first determine the user tag of the target user and multiple target resources to be recommended when recommending resources to the target user, and then obtain the image features corresponding to the images uploaded by the target user, and the file features corresponding to the files uploaded by the target user, so that the total recommendation value of each target resource can be calculated based on the user tag, image features and file features, and finally the final resource recommended to the target user can be determined based on the total recommendation value of each target resource. This process does not require manual labeling of users, nor does it require labeling of target resources to be pushed. The recommended resources can be determined directly based on the image features and text features uploaded by the user, thereby realizing intelligent recommendations for users based on their preferences, meeting the needs of users to obtain personalized resources, protecting user rights, and increasing push open rates and resource subscription rates, thereby increasing the possibility of conversion and subscription, promoting marketing, and increasing revenue. At the same time, it can promote user activity and achieve the purpose of ensuring user stickiness. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art 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 drawings can be obtained based on these drawings without paying creative labor.

[0045] Figure 1 A schematic diagram of a resource recommendation method according to an embodiment of the present invention;

[0046] Figure 2 A schematic diagram of a process for calculating the total recommended value of each target resource provided in an embodiment of the present application;

[0047] Figure 3 A schematic diagram of the structure of a resource recommendation device provided in an embodiment of the present application;

[0048] Figure 4 A schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0050] In one embodiment, Figure 1 As shown, Figure 1 A schematic diagram of a resource recommendation method provided in an embodiment of the present application; the present application provides a resource recommendation method, which may include:

[0051] S110: When recommending resources to a target user, determine a user tag of the target user and a plurality of target resources to be recommended.

[0052] In this step, users can use the client of the business system for cloud storage services. The business system has developed a resource recommendation function. When users use the resource plaza function on the client discovery page, the resource recommendation function can be triggered for calculation. If the system has set a list of target resources to be recommended, the recommended values ​​of the user and the target resources will be calculated at this time. Those that meet the threshold will be recommended to the user and displayed in the target resource list.

[0053] Specifically, when recommending resources to a target user, the resource recommendation function of the present application can first determine the user tag of the target user and multiple target resources to be recommended. In this way, the target resources can be screened according to the degree of match between the user tag of the target user and each target resource, thereby providing the user with target resources that meet their needs.

[0054] Among them, the user tag of this application refers to the feature tag maintained by the system for the user, such as texts such as travel and food. The user tag can be one, multiple, or empty; the target resources of this application refer to a series of resources to be recommended to users, such as movies, songs, food, e-books, learning tutoring tools, etc., including information such as resource name and resource format type. The specific settings can be made according to the actual situation and are not restricted here.

[0055] The target users of this application may be users who subscribe to the resource recommendation function, or they may be new users, old users, or other types of users of this system. The specific settings can be made according to the actual situation and are not restricted here.

[0056] S120: Acquire image features corresponding to the image uploaded by the target user, and file features corresponding to the file uploaded by the target user.

[0057] In this step, after determining the user tag of the target user and multiple target resources to be recommended through S110, in order to meet the target user's needs for obtaining personalized resources, the present application can also obtain image features corresponding to the images uploaded by the target user and file features corresponding to the files uploaded by the target user, and recommend more accurate target resources to the target user through user tags, image features and file features. Moreover, when there is no user tag or the user tag is not updated in time, intelligent resource recommendations can also be made in a timely and accurate manner.

[0058] Among them, the image features of this application refer to the feature data stored in the vector database after the image features are extracted and converted into vector representation during the vectorization process of the image uploaded by the target user. Each element in the feature data represents a feature of the image, such as shape, color distribution, texture, etc.; the file features of this application refer to the file names and file types of all files uploaded by the target user, including various file types such as pictures, videos, audios, words, etc.

[0059] S130: Calculate the total recommendation value of each target resource based on the user tag, image features, and file features.

[0060] In this step, after obtaining the user tag, picture features and file features of the target user through S110 and S120, the present application can calculate the total recommendation value of each target resource based on the user tag, picture features and file features, so that the resources finally recommended to the target user can be determined by the total recommendation value of each target resource.

[0061] For example, the present application can first calculate the recommendation value of each target resource based on user tags, then calculate the recommendation value of each target resource based on image features, and calculate the recommendation value of each target resource based on file features. Then, the present application can combine the recommendation values ​​of the same target resource under user tags, image features, and file features to determine the total recommendation value of each target resource, and then determine the final resource recommended to the target user.

[0062] S140: Determine the final resource recommended to the target user according to the total recommendation value of each target resource.

[0063] In this step, after the total recommendation value of each target resource is calculated by S130, the present application can determine the final resource recommended to the target user according to the total recommendation value of each target resource. For example, after the present application determines the total recommendation value of each target resource, the final resource recommended to the target user can be screened out according to the size of each total recommendation value and recommendation rules, thereby effectively improving the accuracy of resource recommendation.

[0064] In a specific implementation, the target user who logs into the business system in this application, after uploading the file, enters the resource plaza module, the business system can automatically trigger the above resource recommendation method to recommend resources, and the recommended resources are displayed in the recommendation module. If the user does not have the target resource that meets the recommendation, the module recommendation is empty, and the system default resource is displayed. Of course, the user can also actively trigger the resource recommendation function of the system, and push client messages to users with recommended resources, and the user can open the recommended resources through the pushed message.

[0065] In addition, the present application can also expand the application of the above-mentioned algorithm. When the target resource is not a series of resources but a text of user characteristics, such as music hobbies, the system algorithm calculation is performed at this time, and users who meet the target resource form a list, that is, the users are classified. This scenario can be applied to user diversion or A / B testing of new product launches, etc.

[0066] In the above embodiment, when recommending resources to the target user, the user tag of the target user and the multiple target resources to be recommended can be determined first, and then the image features corresponding to the images uploaded by the target user and the file features corresponding to the files uploaded by the target user can be obtained. In this way, the total recommendation value of each target resource can be calculated based on the user tag, image features and file features, and finally the final resource recommended to the target user can be determined based on the total recommendation value of each target resource. This process does not require manual labeling of users or labeling of target resources to be pushed. The recommended resources can be determined directly based on the image features and text features uploaded by the user, thereby realizing intelligent recommendations for users based on their preferences, meeting the needs of users to obtain personalized resources, protecting user rights, and increasing push open rates and resource subscription rates, thereby increasing the possibility of conversion and ordering, promoting marketing, and increasing revenue. At the same time, it can promote user activity and achieve the purpose of ensuring user stickiness.

[0067] In one embodiment, Figure 2 As shown, Figure 2 A schematic diagram of a process for calculating the total recommendation value of each target resource provided in an embodiment of the present application; in S130, calculating the total recommendation value of each target resource based on the user tag, the image feature, and the file feature may include:

[0068] S131: Determine a first recommendation value of the user tag relative to each target resource according to the similarity between the user tag and each target resource.

[0069] S132: Determine a second recommendation value of the image feature relative to each target resource according to the similarity between the image feature and each target resource.

[0070] S133: Determine a third recommendation value of the file feature relative to each target resource according to the similarity between the file feature and each target resource.

[0071] S134: Calculate the total recommendation value of each target resource based on the first recommendation value, the second recommendation value, and the third recommendation value corresponding to each target resource.

[0072] In this embodiment, when recommending resources to target users in the resource plaza of the business system, preference recommendations can be made based on similarity features such as the user's files and pictures, and resources of interest to the user can be pushed, thereby increasing the push open rate and resource subscription rate, thereby increasing the possibility of conversion and subscription.

[0073] Specifically, when calculating the total recommendation value of each target resource based on the user tags, picture features and file features of the target user, the present application can first calculate the recommendation values ​​of the user tags, picture features and file features relative to each target resource respectively, and then calculate the total recommendation value of each target resource.

[0074] For example, when calculating the first recommendation value of a user tag relative to each target resource, the present application may first calculate the similarity between the user tag and each target resource, and then determine the first recommendation value of the user tag relative to each target resource based on each similarity; when calculating the second recommendation value of an image feature relative to each target resource, the present application may first calculate the similarity between the image feature and each target resource, and then determine the second recommendation value of the image feature relative to each target resource based on each similarity; when calculating the third recommendation value of a file feature relative to each target resource, the present application may first calculate the similarity between the file feature and each target resource, and then determine the third recommendation value of the file feature relative to each target resource based on each similarity; after obtaining the first recommendation value, the second recommendation value and the third recommendation value of each target resource, the total recommendation value of each target resource may be calculated based on the first recommendation value, the second recommendation value and the third recommendation value corresponding to each target resource.

[0075] The above-mentioned embodiment does not need to obtain privacy information such as user browsing history and preferences. Instead, it uses a vectorized similarity recommendation algorithm to perform system analysis and recommendation of image features and file features with user authorization, helping users to quickly select and view resources of interest. The recommended resources can dynamically change with the file and image information uploaded by the user. There is no human intervention throughout the process, and the recommended resource information type can only be viewed by the user himself, thereby protecting the user's rights and interests.

[0076] In one embodiment, when there are multiple user tags, determining the first recommendation value of the user tag relative to each target resource according to the similarity between the user tag and each target resource in S131 may include:

[0077] S1311: For each target resource, calculate the similarity between each user tag and the target resource, and select at least one user tag whose similarity is greater than a first similarity threshold from among the user tags as an initial user tag.

[0078] S1312: sorting the initial user tags according to the similarity, and selecting a first preset number of initial user tags from the initial user tags as target user tags according to the first sorting result.

[0079] S1313: Calculate a first recommendation value of the user tag relative to the target resource according to the similarity between each target user tag and the target resource.

[0080] In this embodiment, when determining the first recommendation value of the user tag relative to each target resource according to the similarity between the user tag and each target resource, the user tag and each target resource may be calculated one by one to determine the first recommendation value of each target resource.

[0081] To elaborate, for each target resource, the present application can first calculate the similarity between each user tag and the target resource, and select at least one user tag whose similarity is greater than a first similarity threshold from each user tag as the initial user tag, and then sort the initial user tags according to the similarity, and select a first preset number of initial user tags from each initial user tag as the target user tags according to the first sorting result, and finally calculate the first recommendation value of the user tag relative to the target resource based on the similarity between each target user tag and the target resource.

[0082] In a specific implementation, the application can predefine that the number of user tags is unlimited, the minimum number can be 0, and the tags in the user tag (S1) are respectively , … , where each contains one element, namely the tag name, and the number of tags n can be 0 to infinite; the target resources (T) in this application are , … .

[0083] Next, the present application can calculate the similarity between each user tag of the target user and the target resource based on the cosine similarity algorithm (hereinafter referred to as Cosine), and then set the first similarity threshold of the user tag to , sort by the similarity between the user tag and the target resource, from large to small, and take the first H similarity values ​​that meet the threshold condition, and then calculate, then the first recommended value of the user tag is:

[0084]

[0085] The first recommendation value of the user tag relative to each target resource can be calculated by the above formula.

[0086] In one embodiment, when there are multiple image features, determining the second recommendation value of the image feature relative to each target resource according to the similarity between the image feature and each target resource in S132 may include:

[0087] S1321: For each target resource, calculate the similarity between each image feature and the target resource, and select at least one image feature whose similarity is greater than a second similarity threshold from among the image features as an initial image feature.

[0088] S1322: sorting the initial picture features according to the similarity, and selecting a second preset number of initial picture features from the initial picture features as target picture features according to the second sorting result.

[0089] S1323: Calculate a second recommendation value of each target image feature relative to the target resource according to the similarity between each target image feature and the target resource.

[0090] In this embodiment, when determining the second recommendation value of the image feature relative to each target resource according to the similarity between the image feature and each target resource, the image feature and each target resource may be calculated one by one to determine the second recommendation value of each target resource.

[0091] To elaborate, for each target resource, the present application can first calculate the similarity between each image feature and the target resource, and select at least one image feature whose similarity is greater than a second similarity threshold from each image feature as the initial image feature, and then sort each initial image feature according to the size of the similarity, and select a second preset number of initial image features from each initial image feature as the target image features according to the second sorting result, and finally calculate the second recommendation value of the image feature relative to the target resource based on the similarity between each target image feature and the target resource.

[0092] In a specific implementation, the present application may set the picture vectors in the picture feature (S2) to be , … , where each contains multiple elements, such as shape, color distribution, texture, etc. The type of element can be determined according to the degree of recognition of the feature value captured by the image vectorization. It can be set from 1 to multiple, and the number of images can be 0 to infinite. Next, the application can first calculate the similarity between each image feature of the target user and the target resource based on the cosine similarity algorithm, and then set the second similarity threshold of the single image feature to , take the first M similarity values ​​that meet the threshold condition, and then calculate, the recommended value of the image feature is:

[0093]

[0094] In addition, considering the large number of user files, the system here specifically takes the first M file values ​​that meet the second similarity threshold in order, which can improve the calculation efficiency of the system while also achieving the accuracy of the recommendation.

[0095] In one embodiment, when there are multiple file features, determining the third recommendation value of the file feature relative to each target resource according to the similarity between the file feature and each target resource in S133 may include:

[0096] S1331: For each target resource, calculate the similarity between each file feature and the target resource, and select at least one file feature whose similarity is greater than a third similarity threshold from among the file features as an initial file feature.

[0097] S1332: sorting the initial file features according to the similarity, and selecting a third preset number of initial file features from the initial file features as target file features according to the third sorting result.

[0098] S1333: Calculate a third recommendation value of each target file feature relative to the target resource according to the similarity between each target file feature and the target resource.

[0099] In this embodiment, when determining the third recommendation value of the file feature relative to each target resource according to the similarity between the file feature and each target resource, the file feature and each target resource may be calculated one by one to determine the third recommendation value of each target resource.

[0100] To elaborate, for each target resource, the present application can first calculate the similarity between each file feature and the target resource, and select at least one file feature whose similarity is greater than a third similarity threshold from each user tag as the initial file feature, and then sort the initial file features according to the size of the similarity, and select a third preset number of initial file features from each initial file feature as the target file features according to the third sorting result, and finally calculate the third recommendation value of the file feature relative to the target resource based on the similarity between each target file feature and the target resource.

[0101] In a specific implementation, the present application may set the files in the file feature (S3) to be , … .in, The file name and file type are included, and the number of user files can be 0 to unlimited. Next, the application can first calculate the similarity between each image feature of the target user and the target resource based on the cosine similarity algorithm, and then set the third similarity threshold of the image feature to , take the first N similarity values ​​that meet the threshold condition, and then calculate again, then the third recommended value of the file feature is:

[0102]

[0103] In addition, considering the large number of user files, the system here specifically takes the first N file values ​​that meet the similarity threshold conditions in order, which can improve the system's calculation efficiency while also achieving the accuracy of recommendations.

[0104] In one embodiment, calculating the total recommendation value of each target resource based on the first recommendation value, the second recommendation value, and the third recommendation value corresponding to each target resource in S134 may include:

[0105] S1341: Determine a first weight corresponding to the user tag, a second weight corresponding to the image feature, and a third weight corresponding to the text feature.

[0106] S1342: For each target resource, calculate the total recommendation value of the target resource based on the first recommendation value and first weight of the user tag relative to the target resource, the second recommendation value and second weight of the image feature relative to the target resource, and the third recommendation value and third weight of the text feature relative to the target resource.

[0107] In this embodiment, when calculating the total recommendation value of each target resource, the first weight corresponding to the user tag, the second weight corresponding to the image feature, and the third weight corresponding to the text feature can be determined first. In this way, for each target resource, the total recommendation value of the target resource can be calculated based on the first recommendation value and first weight of the user tag relative to the target resource, the second recommendation value and second weight of the image feature relative to the target resource, and the third recommendation value and third weight of the text feature relative to the target resource.

[0108] In a specific implementation, when calculating the total recommendation value of each target resource, the application can first set the weight of the user tag to , the weight of the image feature is , the weight of the file feature is , then set the threshold for target resource recommendation to , set the algorithm in this paper to be abbreviated as Recom, then the threshold is met through the system algorithm The target resources are all the resources that can be recommended to users, namely:

[0109]

[0110] The total recommendation value of each target resource can be calculated through the above calculation formula, and then the final resource recommended to the target user can be determined according to the total recommendation value of each target resource.

[0111] In one embodiment, determining the final resource recommended to the target user according to the total recommendation value of each target resource in S140 may include:

[0112] S141: Select at least one target resource whose total recommendation value is greater than a preset recommendation threshold from various target resources as a final resource recommended to the target user.

[0113] S143: Alternatively, at least one target resource whose total recommendation value is greater than a preset recommendation threshold is selected from each target resource as a recommendable resource.

[0114] S143: sorting the recommendable resources according to the total recommendation value, and selecting a fourth preset number of recommendable resources from the recommendable resources according to the fourth sorting result as final resources recommended to the target user.

[0115] In this embodiment, when determining the final resource to be recommended to the target user, at least one target resource whose total recommendation value is greater than a preset recommendation threshold can be selected from each target resource as the final resource recommended to the target user, or at least one target resource whose total recommendation value is greater than the preset recommendation threshold can be selected from each target resource as a recommendable resource, and then each recommendable resource is sorted according to the size of the total recommendation value, and a fourth preset number of recommendable resources are selected from each recommendable resource according to the fourth sorting result as the final resource recommended to the target user.

[0116] For example, when the application sorts the recommendable resources according to the total recommendation value, it can be calculated according to the following formula:

[0117]

[0118] The above formula can be used to calculate all the resources that can be recommended to the target user. At the same time, the top K resources can be obtained according to the ranking for priority recommendation, thereby meeting the personalized recommendation needs.

[0119] The resource recommendation device provided in an embodiment of the present application is described below. The resource recommendation device described below and the resource recommendation method described above can refer to each other.

[0120] In one embodiment, Figure 3 As shown, Figure 3 A schematic diagram of the structure of a resource recommendation device provided in an embodiment of the present application; the present application also provides a resource recommendation device, including a data acquisition module 210, a feature acquisition module 220, a recommendation value calculation module 230, and a resource recommendation module 240, which specifically includes the following:

[0121] The data acquisition module 210 is used to determine the user tag of the target user and a plurality of target resources to be recommended when recommending resources to the target user.

[0122] The feature acquisition module 220 is used to acquire the image features corresponding to the image uploaded by the target user and the file features corresponding to the file uploaded by the target user.

[0123] The recommendation value calculation module 230 is used to calculate the total recommendation value of each target resource based on the user tag, the image feature and the file feature.

[0124] The resource recommendation module 240 is used to determine the final resource recommended to the target user according to the total recommendation value of each target resource.

[0125] In the above embodiment, when recommending resources to the target user, the user tag of the target user and the multiple target resources to be recommended can be determined first, and then the image features corresponding to the images uploaded by the target user and the file features corresponding to the files uploaded by the target user can be obtained. In this way, the total recommendation value of each target resource can be calculated based on the user tag, image features and file features, and finally the final resource recommended to the target user can be determined based on the total recommendation value of each target resource. This process does not require manual labeling of users or labeling of target resources to be pushed. The recommended resources can be determined directly based on the image features and text features uploaded by the user, thereby realizing intelligent recommendations for users based on their preferences, meeting the needs of users to obtain personalized resources, protecting user rights, and increasing push open rates and resource subscription rates, thereby increasing the possibility of conversion and ordering, promoting marketing, and increasing revenue. At the same time, it can promote user activity and achieve the purpose of ensuring user stickiness.

[0126] In one embodiment, the present application also provides a computer-readable storage medium, in which computer-readable instructions are stored. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the resource recommendation method described in any of the above embodiments.

[0127] In one embodiment, the present application also provides a computer device, including: one or more processors, and a memory.

[0128] The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the steps of the resource recommendation method described in any one of the above embodiments are performed.

[0129] Indicatively, if Figure 4 As shown, Figure 4This is a schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. The computer device 300 may be provided as a server. Figure 4 The computer device 300 includes a processing component 302, which further includes one or more processors, and a memory resource represented by a memory 301, for storing instructions executable by the processing component 302, such as an application. The application stored in the memory 301 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 302 is configured to execute instructions to perform the resource recommendation method of any of the above embodiments.

[0130] The computer device 300 may further include a power supply component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305. The computer device 300 may operate based on an operating system stored in the memory 301, such as Windows Server TM, Mac OS X TM, Unix TM, Linux TM, Free BSD TM, or the like.

[0131] Those skilled in the art will understand that Figure 4 The 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.

[0132] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0133] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can refer to each other.

[0134] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A resource recommendation method, characterized in that: The method comprises: When recommending resources to a target user, determining a user tag of the target user and a plurality of target resources to be recommended; Acquire image features corresponding to the image uploaded by the target user, and file features corresponding to the file uploaded by the target user; Calculate the total recommendation value of each target resource based on the user tag, the image feature and the file feature; The final resource recommended to the target user is determined according to the total recommendation value of each target resource.

2. The resource recommendation method according to claim 1, characterized in that: The calculating the total recommendation value of each target resource based on the user tag, the image feature and the file feature includes: Determining a first recommendation value of the user tag relative to each target resource according to the similarity between the user tag and each target resource; Determining, according to the similarity between the image feature and each target resource, a second recommendation value of the image feature relative to each target resource; Determining a third recommendation value of the file feature relative to each target resource according to the similarity between the file feature and each target resource; Based on the first recommendation value, the second recommendation value, and the third recommendation value corresponding to each target resource, the total recommendation value of each target resource is calculated.

3. The resource recommendation method according to claim 2, characterized in that: When there are multiple user tags, determining the first recommendation value of the user tag relative to each target resource according to the similarity between the user tag and each target resource includes: For each target resource, calculate the similarity between each user tag and the target resource, and select at least one user tag whose similarity is greater than a first similarity threshold from each user tag as an initial user tag; Sorting the initial user tags according to the similarity, and selecting a first preset number of initial user tags from the initial user tags as target user tags according to the first sorting result; According to the similarity between each target user tag and the target resource, a first recommendation value of the user tag relative to the target resource is calculated.

4. The resource recommendation method according to claim 2, characterized in that: When there are multiple image features, determining the second recommendation value of the image feature relative to each target resource according to the similarity between the image feature and each target resource includes: For each target resource, calculate the similarity between each image feature and the target resource, and select at least one image feature whose similarity is greater than a second similarity threshold from each image feature as an initial image feature; Sorting the initial picture features according to their similarity, and selecting a second preset number of initial picture features from the initial picture features as target picture features according to the second sorting result; According to the similarity between each target image feature and the target resource, a second recommendation value of the image feature relative to the target resource is calculated.

5. The resource recommendation method according to claim 2, characterized in that: When there are multiple file features, determining the third recommendation value of the file feature relative to each target resource according to the similarity between the file feature and each target resource includes: For each target resource, calculate the similarity between each file feature and the target resource, and select at least one file feature whose similarity is greater than a third similarity threshold from each file feature as an initial file feature; Sorting the initial file features according to the similarity, and selecting a third preset number of initial file features from the initial file features as target file features according to the third sorting result; According to the similarity between each target file feature and the target resource, a third recommendation value of the file feature relative to the target resource is calculated.

6. The resource recommendation method according to claim 2, characterized in that: The calculating the total recommendation value of each target resource based on the first recommendation value, the second recommendation value, and the third recommendation value corresponding to each target resource includes: Determine a first weight corresponding to the user tag, a second weight corresponding to the image feature, and a third weight corresponding to the text feature; For each target resource, the total recommendation value of the target resource is calculated based on the first recommendation value and first weight of the user tag relative to the target resource, the second recommendation value and second weight of the image feature relative to the target resource, and the third recommendation value and third weight of the text feature relative to the target resource.

7. The resource recommendation method according to any one of claims 1 to 6, characterized in that: The determining the final resource recommended to the target user according to the total recommendation value of each target resource includes: Selecting at least one target resource whose total recommendation value is greater than a preset recommendation threshold from each target resource as a final resource recommended to the target user; Alternatively, at least one target resource whose total recommendation value is greater than a preset recommendation threshold is selected from each target resource as a recommendable resource; The recommendable resources are sorted according to the total recommendation value, and a fourth preset number of recommendable resources are selected from the recommendable resources according to the fourth sorting result as final resources recommended to the target user.

8. A resource recommendation device, characterized in that: include: A data acquisition module, used to determine the user tag of the target user and a plurality of target resources to be recommended when recommending resources to the target user; A feature acquisition module, used to acquire image features corresponding to the image uploaded by the target user, and file features corresponding to the file uploaded by the target user; A recommendation value calculation module, used to calculate the total recommendation value of each target resource based on the user tag, the image feature and the file feature; The resource recommendation module is used to determine the final resource recommended to the target user according to the total recommendation value of each target resource.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the resource recommendation method according to any one of claims 1 to 7.

10. A computer device, characterized in that: include: one or more processors, and memory; The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the steps of the resource recommendation method according to any one of claims 1 to 7 are performed.

Images