Content recommendation method and device based on large model, electronic equipment and storage medium

Abstract

The invention provides a content recommendation method and device based on a large model, electronic equipment and a storage medium, and relates to the technical field of computers, in particular to the technical field of artificial intelligence such as deep learning, intelligent recommendation and large models. According to the specific scheme, firstly, historical consumption resources of a user in a first recommendation service and a second recommendation service installed in equipment of the user are obtained, and then the historical consumption resources are recognized based on a first recognition model associated with the first recommendation service and a second recognition model associated with the second recommendation service; the method comprises the steps of selecting a candidate resource list according to historical consumption resources to determine a target recommendation service associated with the historical consumption resources, then traversing the candidate resource list based on the target recommendation service to determine a target resource associated with the target recommendation service, finally reconfiguring a recommendation weight of the target resource, and recommending content to a user based on the reconfigured recommendation weight.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to the field of artificial intelligence technology such as deep learning, intelligent recommendation, and big models, and specifically to a content recommendation method, device, electronic device, and storage medium based on big models. Background Art

[0002] Current content recommendation systems are usually based on resource content, such as determining the type of video to be recommended based on the characters, environment, background music, etc. in the video, and then recommending content. However, this method of content recommendation will result in a lack of diversity in the types of content recommended to users. Summary of the invention

[0003] The present disclosure aims to solve one of the technical problems in the related art at least to some extent.

[0004] The first embodiment of the present disclosure proposes a content recommendation method based on a large model, including:

[0005] Obtaining historical consumption resources of a user in a first recommended service and a second recommended service installed in a device of the user;

[0006] Based on a first recognition model associated with the first recommendation service and a second recognition model associated with the second recommendation service, the historical consumption resource is identified to determine a target recommendation service associated with the historical consumption resource;

[0007] Based on the target recommendation service, traverse the candidate resource list to determine the target resource associated with the target recommendation service;

[0008] reconfiguring the recommendation weight of the target resource, wherein the recommendation weight after the reconfiguration is greater than the recommendation weight before the reconfiguration;

[0009] Based on the reconfigured recommendation weights, content is recommended to the user.

[0010] The second aspect of the present disclosure provides a content recommendation device based on a large model, including:

[0011] An acquisition module, used to acquire the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the device of the user;

[0012] an identification module, configured to identify the historical consumption resource based on a first identification model associated with the first recommendation service and a second identification model associated with the second recommendation service, so as to determine a target recommendation service associated with the historical consumption resource;

[0013] A traversal module, used to traverse the candidate resource list based on the target recommendation service to determine the target resource associated with the target recommendation service;

[0014] A configuration module, used to reconfigure the recommendation weight of the target resource, wherein the recommendation weight after reconfiguration is greater than the recommendation weight before reconfiguration;

[0015] A recommendation module is used to recommend content to the user based on the reconfigured recommendation weight.

[0016] The third aspect embodiment of the present disclosure proposes a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the large model-based content recommendation method proposed in the first aspect embodiment of the present disclosure is implemented.

[0017] The fourth aspect of the present disclosure provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the large model-based content recommendation method provided in the first aspect of the present disclosure is implemented.

[0018] The fifth aspect of the present disclosure provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the large model-based content recommendation method as provided in the first aspect of the present disclosure.

[0019] The content recommendation method, device, electronic device and storage medium based on a large model provided by the present disclosure have the following beneficial effects:

[0020] In the disclosed embodiment, the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device are first obtained, and then the historical consumption resources are identified based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service, respectively, to determine the target recommendation service associated with the historical consumption resources, and then based on the target recommendation service, the candidate resource list is traversed to determine the target resource associated with the target recommendation service, and finally the recommendation weight of the target resource is reconfigured, and content is recommended to the user based on the reconfigured recommendation weight. Thus, by identifying the historical consumption resources of the user based on the recognition models associated with different recommendation services, the target recommendation service is determined, and based on the target recommendation service, the candidate resource list is traversed to determine the target resource to be recommended to the user, and the recommendation weight of the target resource is reconfigured, and based on the reconfigured recommendation weight, the target resource is recommended to the user, thereby achieving content recommendation in a personalized atmosphere for the user and improving the diversity and reliability of content recommendation.

[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0023] Figure 1 A flowchart of a content recommendation method based on a large model provided by an embodiment of the present disclosure;

[0024] Figure 2 A flowchart of a content recommendation method based on a large model provided by an embodiment of the present disclosure;

[0025] Figure 3 A flowchart of a content recommendation method based on a large model provided by an embodiment of the present disclosure;

[0026] Figure 4 A flowchart of a content recommendation method based on a large model provided by an embodiment of the present disclosure;

[0027] Figure 5 A flowchart of a content recommendation method based on a large model provided by an embodiment of the present disclosure;

[0028] Figure 6 A flowchart of a content recommendation method based on a large model provided in an embodiment of the present disclosure;

[0029] Figure 7 A schematic diagram of the structure of a content recommendation device based on a large model provided in an embodiment of the present disclosure;

[0030] Figure 8 A block diagram of an exemplary computer device suitable for implementing embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0031] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0032] The present disclosure relates to artificial intelligence technology fields such as deep learning, intelligent recommendation, and large models.

[0033] Artificial Intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence.

[0034] Deep Learning (DL) is the process of learning the inherent laws and representation levels of sample data. The information obtained in the learning process is very helpful for interpreting data such as text, images, and sounds. The ultimate goal of deep learning is to enable machines to have analytical learning capabilities like humans and to recognize data such as text, images, and sounds.

[0035] Intelligent recommendation, as an important means of information filtering and personalized services, is centered on integrating multi-dimensional information such as users' historical behavior data, interest preferences, and contextual environment, and using algorithm models to automatically analyze and predict users' potential needs or interests, and then recommend content, products, services, or social relationships that may interest users.

[0036] The large model can also be called the Foundation Model. The model extracts knowledge through billions of corpora or images, learns and then produces a large model with billions of parameters.

[0037] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0038] The following describes a large model-based content recommendation method, device, electronic device, and storage medium according to an embodiment of the present disclosure with reference to the accompanying drawings.

[0039] Figure 1 A flowchart of a content recommendation method based on a large model provided in an embodiment of the present disclosure.

[0040] like Figure 1 As shown, the content recommendation method based on the large model may include the following steps:

[0041] Step 101: Obtain the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device.

[0042] The first recommendation service may be a service for content recommendation, a service for recommending any content, or any type of recommendation service. For example, the first recommendation service may be a graphic content recommendation service, or a video content recommendation service, etc. For another example, the first recommendation service may be a recommendation service platform for content recommendation, or a recommendation service community, etc., and the present disclosure does not limit this.

[0043] It should be noted that different first recommendation services may have different corresponding historical consumption resources of users. For example, when the first recommendation service is a picture and text recommendation service, the corresponding historical consumption resources of users may be picture and text consumption resources, and when the first recommendation service is a video recommendation service, the corresponding historical consumption resources of users may be video consumption resources, etc. This disclosure does not limit this.

[0044] It should be noted that the second recommended service may be any program with a recommended service installed in the user's device, and the present disclosure does not limit this.

[0045] In the present disclosure, before using the first recommendation service to recommend content to the user, in order to improve the diversity and accuracy of the content recommended to the user, the user's historical consumption resources in the first recommendation service and the second recommendation service installed in the user's device can be first obtained, thereby providing a data basis for recommending content to the user.

[0046] Step 102 : Identify the historical consumption resources based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service to determine the target recommendation service associated with the historical consumption resources.

[0047] The first recognition model and the second recognition model can both be models for identifying the type of content resources, and can be pre-set models of any type and structure. For example, the first recognition model and the second recognition model can both be models composed of a deep neural network and an LLM fine-tuning model, and the present disclosure does not limit this.

[0048] Among them, LLM is the abbreviation of Large Language Model (LLM).

[0049] It should be noted that, if the first recommendation service is different, the associated first recognition model may be different; if the second recommendation service is different, the associated second recognition model may be different; the first recognition model may be different from the second recognition model, and the present disclosure does not limit this.

[0050] The target recommendation service may be a recommendation service that matches the type of historical consumption resources of the user.

[0051] In the present disclosure, after obtaining the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device, the historical consumption resources can be identified based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service, respectively, to determine the target recommendation service associated with the historical consumption resources, thereby providing conditions for realizing content recommendations in a personalized atmosphere for the user.

[0052] Step 103: Based on the target recommendation service, the candidate resource list is traversed to determine the target resource associated with the target recommendation service.

[0053] The candidate resource list may be a resource list composed of candidate resources to be recommended to the user in the first recommendation service, which may be pre-set and may include any number of candidate resources, which is not limited in the present disclosure.

[0054] The target resource may be a resource recommended to the user.

[0055] In the present disclosure, after determining the target recommendation service associated with the historical consumption resource, the candidate resource list can be traversed based on the target recommendation service to determine the target resource associated with the target recommendation service, thereby improving the accuracy and reliability of the determined target resource.

[0056] Step 104 , reconfiguring the recommendation weight of the target resource, wherein the recommendation weight after reconfiguration is greater than the recommendation weight before reconfiguration.

[0057] In the present disclosure, after determining the target resource associated with the target recommendation service, in order to recommend the target resource to the user, the recommendation weight of the target resource can be reconfigured so that the recommendation weight after reconfiguration is greater than the recommendation weight before reconfiguration, thereby increasing the probability of recommending the target resource to the user and improving the user experience.

[0058] It should be noted that when reconfiguring the recommended weight of the target resource, the reconfiguration of the recommended weight can be preset or determined according to actual conditions. For example, the recommended weight of the target resource can be increased by a preset value, etc., which is not limited in the present disclosure.

[0059] Step 105: Recommend content to the user based on the reconfigured recommendation weights.

[0060] In the present disclosure, after the recommendation weights of the target resources are reconfigured, content recommendations can be made to users based on the reconfigured recommendation weights, thereby improving the diversity and accuracy of content recommendations.

[0061] In the disclosed embodiment, the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device are first obtained, and then the historical consumption resources are identified based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service, respectively, to determine the target recommendation service associated with the historical consumption resources, and then based on the target recommendation service, the candidate resource list is traversed to determine the target resource associated with the target recommendation service, and finally the recommendation weight of the target resource is reconfigured, and content is recommended to the user based on the reconfigured recommendation weight. Thus, by identifying the historical consumption resources of the user based on the recognition models associated with different recommendation services, the target recommendation service is determined, and based on the target recommendation service, the candidate resource list is traversed to determine the target resource to be recommended to the user, the recommendation weight of the target resource is reconfigured, and content is recommended to the user based on the reconfigured recommendation weight, thereby realizing content recommendation for the user's personalized atmosphere and improving the diversity and reliability of content recommendation based on large models.

[0062] Figure 2 A flowchart of a content recommendation method based on a large model provided in one embodiment of the present disclosure.

[0063] like Figure 2 As shown, the content recommendation method based on the large model may include the following steps:

[0064] Step 201: Obtain the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device.

[0065] The specific implementation form of step 201 can refer to the detailed description in other embodiments of the present disclosure, and will not be described in detail here.

[0066] Step 202 , inputting historical consumption resources into a first recognition model associated with a first recommendation service and a second recognition model associated with a second recommendation service, respectively, to obtain a first probability output by the first recognition model and a second probability output by the second recognition model.

[0067] It should be noted that the second recognition model corresponding to the second recommendation service can be obtained by taking the first resource set associated with the first recommendation service as a negative sample and the second resource set associated with the second recommendation service as a positive sample, so that the second recognition model can identify the resource type in the second recommendation service associated with it, thereby improving the effect and accuracy of the second recognition model. The present disclosure does not limit this.

[0068] The first resource set may be a resource set composed of resources in the first recommendation service, which may be pre-set and may include any number of resources, which is not limited in the present disclosure.

[0069] The second resource set may be a resource set composed of resources in the second recommendation service, which may be pre-set and may include any number of resources, which is not limited in the present disclosure.

[0070] It should be noted that, if the second recommendation service is different, the second resource set associated therewith may be different, and this disclosure does not limit this.

[0071] The first probability may be the probability that the type of the historically consumed resource is the resource type of the first recommended service.

[0072] The second probability may be a probability that the type of the historical consumption resource is the resource type of the second recommended service associated with the second recognition model.

[0073] It should be noted that the first probability may be different from the second probability, or may be the same as the second probability, and the present disclosure does not limit this.

[0074] It should be noted that, for different second recommendation services, the second probabilities output by the associated second recognition models may be different, and this disclosure does not limit this.

[0075] In the present disclosure, after obtaining the user's historical consumption resources in a first recommendation service and a second recommendation service installed in the user's device, in order to identify the type of the user's historical consumption resources and determine the recommended service preferred by the user, the historical consumption resources can be input into a first recognition model associated with the first recommendation service and a second recognition model associated with the second recommendation service, respectively, to obtain a first probability output by the first recognition model and a second probability output by the second recognition model, thereby providing a data basis for determining the target recommendation service associated with the historical consumption resources.

[0076] Step 203: determine the recommended service associated with the maximum value of the first probability and the second probability as the target recommended service.

[0077] In the present disclosure, after obtaining the first probability output by the first recognition model and the second probability output by the second recognition model, the recommended service associated with the maximum value of the first probability and the second probability can be determined as the target recommended service, thereby improving the accuracy of the determined target recommended service.

[0078] Optionally, when the number of historical consumption resources of the user in the first recommendation service obtained is at least two, each historical consumption resource can be input into the first recognition model and the second recognition model respectively to obtain a first probability output by the first recognition model and a second probability output by the second recognition model. Since for any historical consumption resource, the probabilities output by the recognition models associated with different recommendation services may be the same, or the number of maximum values ​​of the first probability and the second probability is multiple, at this time, at least one candidate recommendation service corresponding to each historical consumption resource can be determined according to the probabilities output by the recognition models associated with different recommendation services, and at least one candidate recommendation service with the highest frequency among the candidate recommendation services corresponding to at least two historical consumption resources can be determined as the target recommendation service, thereby improving the reliability and diversity of the determined target recommendation services.

[0079] Step 204: based on the target recommendation service, traverse the candidate resource list to determine the target resource associated with the target recommendation service.

[0080] Step 205 , reconfiguring the recommendation weight of the target resource, wherein the recommendation weight after reconfiguration is greater than the recommendation weight before reconfiguration.

[0081] Step 206: Recommend content to the user based on the reconfigured recommendation weights.

[0082] The specific implementation forms of steps 204 to 206 may refer to the detailed descriptions in other embodiments of the present disclosure, and will not be described in detail here.

[0083] In the disclosed embodiment, the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device are first obtained, and then the historical consumption resources are respectively input into the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service to obtain the first probability output by the first recognition model and the second probability output by the second recognition model, and then the recommendation service associated with the maximum value of the first probability and the second probability is determined as the target recommendation service, and finally based on the target recommendation service, the candidate resource list is traversed to determine the target resource associated with the target recommendation service, reconfigure the recommendation weight of the target resource, and recommend content to the user based on the reconfigured recommendation weight. Thus, by inputting the historical consumption resources of the user into the recognition models associated with different recommendation services, the probability output by each recognition model is obtained, the recommendation service associated with the maximum probability is determined as the target recommendation service corresponding to the historical consumption resources, and based on the target recommendation service and the candidate resource list, the target resource recommended to the user is determined, the recommendation weight of the target resource is reconfigured, and content is recommended to the user based on the reconfigured recommendation weight, thereby improving the diversity and accuracy of content recommendation.

[0084] Figure 3 A flowchart of a content recommendation method based on a large model provided in one embodiment of the present disclosure.

[0085] like Figure 3 As shown, the content recommendation method based on the large model may include the following steps:

[0086] Step 301: Obtain the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device.

[0087] Step 302 : Identify the historical consumption resources based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service to determine the target recommendation service associated with the historical consumption resources.

[0088] The specific implementation forms of step 301 to step 302 can refer to the detailed descriptions in other embodiments of the present disclosure, and will not be described in detail here.

[0089] Step 303: Match the target recommended service with the third recommended service associated with each candidate resource in the candidate resource list.

[0090] It should be noted that the third recommendation service may be the same as the first recommendation service or the second recommendation service, or may be different from the first recommendation service or the second recommendation service, and the present disclosure does not limit this.

[0091] In the present disclosure, after determining the target recommendation service associated with the user's historical consumption resources, the target recommendation service can be matched with the third recommendation service associated with each candidate resource in the candidate resource list, thereby providing conditions for recommending resources that meet the user's preferences to the user.

[0092] It should be noted that when matching the target recommended resource with the third recommended service associated with each candidate resource, each candidate resource can be input into the recognition model associated with different recommendation services offline or online to predetermine the third recommended service associated with each candidate resource. The present disclosure does not limit this.

[0093] Step 304: determine the candidate resource associated with the third recommendation service matching the target recommendation service as the target resource.

[0094] In the present disclosure, after matching the target recommendation service with the third recommendation service associated with each candidate resource in the candidate resource list, the candidate resources associated with the third recommendation service that matches the target recommendation service are determined as target resources recommended to the user, so that the type of the determined target resource is consistent with the resource type preferred by the user, thereby improving the user experience.

[0095] Step 305 , reconfigure the recommendation weight of the target resource, wherein the recommendation weight after reconfiguration is greater than the recommendation weight before reconfiguration.

[0096] Step 306: Recommend content to the user based on the reconfigured recommendation weights.

[0097] The specific implementation forms of step 305 to step 306 can refer to the detailed descriptions in other embodiments of the present disclosure, and will not be described in detail here.

[0098] In the disclosed embodiment, the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device are first obtained, and then the historical consumption resources are identified based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service to determine the target recommendation service associated with the historical consumption resources, and then the target recommendation service is matched with the third recommendation service associated with each candidate resource in the candidate resource list, and the candidate resources associated with the third recommendation service matched with the target recommendation service are determined as the target resources, and finally the recommendation weight of the target resource is reconfigured, and content is recommended to the user based on the reconfigured recommendation weight. Thus, after determining the target recommendation service associated with the user's historical consumption resources based on the recognition models associated with different recommendation services, the target recommendation service can be matched with the recommendation service associated with each candidate resource, and the candidate resources matched with the associated recommendation service and the target recommendation service can be determined to be the target resource recommended to the user, and the recommendation weight of the target resource can be reconfigured, and content is recommended to the user based on the reconfigured recommendation weight, thereby ensuring that the types of resources recommended to the user meet the user's preferences, and improving the diversity and accuracy of content recommendations.

[0099] Figure 4 A flowchart of a content recommendation method based on a large model provided in one embodiment of the present disclosure.

[0100] like Figure 4 As shown, the content recommendation method based on the large model may include the following steps:

[0101] Step 401: Obtain the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device.

[0102] Step 402 : Identify the historical consumption resources based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service to determine the target recommendation service associated with the historical consumption resources.

[0103] The specific implementation forms of step 401 to step 402 can refer to the detailed descriptions in other embodiments of the present disclosure, and will not be described in detail here.

[0104] Step 403: for each candidate resource in the candidate resource list, input it into each recognition model to obtain each third probability corresponding to the candidate resource.

[0105] The third probability may be the type of the candidate resource, which is the probability of the resource type of the recommended service associated with the identification model.

[0106] It should be noted that, due to different recognition models, the third probability corresponding to the same candidate resource may be different, or may be different, and the present disclosure does not limit this.

[0107] In the present disclosure, after determining the target recommendation service associated with the historical consumption resources, in order to determine the resources recommended to the user, each candidate resource in the candidate resource list can be first input into each recognition model respectively to obtain the third probabilities corresponding to the candidate resource, so that the recommendation service associated with each candidate resource can be determined, providing conditions for determining the resources recommended to the user.

[0108] It should be noted that after the candidate resources are input into each recognition model, each recognition model can determine the third probabilities corresponding to the candidate resources based on the title, cover image and other information of the candidate resources, and the present disclosure does not limit this.

[0109] Step 404: determine the recommended service associated with the recognition model corresponding to the maximum value among the third probabilities as the third recommended service associated with the candidate resource.

[0110] In the present disclosure, after obtaining the third probabilities corresponding to each candidate resource, the maximum value of the third probabilities corresponding to each candidate resource can be first determined, and then the recommended service associated with the recognition model corresponding to the maximum value can be determined as the third recommended service associated with the corresponding candidate resource, thereby improving the accuracy of the third recommended service associated with each candidate resource.

[0111] It should be noted that, based on the recommended service associated with the recognition model corresponding to the maximum value of each third probability corresponding to each candidate resource, the third recommended service associated with each candidate resource may be different or may be the same, and the present disclosure does not limit this.

[0112] It should be noted that the process of inputting each candidate resource into a recognition model respectively, obtaining the third probabilities corresponding to each candidate resource, and determining the third recommended service associated with each candidate resource based on the third probabilities corresponding to each candidate resource can be determined offline or online, and the present disclosure does not limit this.

[0113] Step 405: Match the target recommended service with the third recommended service associated with each candidate resource in the candidate resource list.

[0114] Step 406: Determine the candidate resource associated with the third recommendation service that matches the target recommendation service as the target resource.

[0115] Step 407, reconfiguring the recommendation weight of the target resource, wherein the recommendation weight after reconfiguration is greater than the recommendation weight before reconfiguration.

[0116] Step 408: Recommend content to the user based on the reconfigured recommendation weights.

[0117] The specific implementation forms of steps 405 to 408 can refer to the detailed descriptions in other embodiments of the present disclosure, and will not be described in detail here.

[0118] In the disclosed embodiment, the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device are first obtained, and the historical consumption resources are identified based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service, respectively, to determine the target recommendation service associated with the historical consumption resources. Then, for each candidate resource in the candidate resource list, each recognition model is inputted respectively to obtain the third probabilities corresponding to the candidate resource, and the recommendation service associated with the recognition model corresponding to the maximum value of each third probability is determined as the third recommendation service associated with the candidate resource. Then, the target recommendation service is matched with the third recommendation service associated with each candidate resource in the candidate resource list, and the candidate resource associated with the third recommendation service matched with the target recommendation service is determined as the target resource. Finally, the recommendation weight of the target resource is reconfigured, and content is recommended to the user based on the reconfigured recommendation weight. Therefore, after determining the target recommended service associated with the user's historical consumption resources based on the recognition models associated with different recommendation services, the probabilities corresponding to each candidate resource in the candidate resource list are determined based on each recognition model, and the recommendation service associated with the recognition model corresponding to the maximum value of each probability is determined as the recommendation service for the corresponding candidate resource management, the target recommendation service is matched with the recommendation service associated with each candidate resource, the candidate resources associated with the successfully matched recommendation service are determined as the target resources for recommendation to the user, and their recommendation weights are reconfigured. Based on the reconfigured recommendation weights, content is recommended to the user, thereby improving the reliability of content recommendations based on large models.

[0119] Figure 5 A flowchart of a content recommendation method based on a large model provided in one embodiment of the present disclosure.

[0120] like Figure 5As shown, the content recommendation method based on the large model may include the following steps:

[0121] Step 501: Obtain the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device.

[0122] Step 502 : Identify the historical consumption resources based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service to determine the target recommendation service associated with the historical consumption resources.

[0123] Step 503: for each candidate resource in the candidate resource column, input it into each recognition model to obtain each third probability corresponding to the candidate resource.

[0124] The specific implementation forms of steps 501 to 503 can refer to the detailed descriptions in other embodiments of the present disclosure, and will not be described in detail here.

[0125] Step 504: when at least two of the third probabilities are greater than a threshold, at least two recommended services respectively associated with the at least two third probabilities are determined as third recommended services associated with the candidate resource.

[0126] The threshold value may be a critical value of the third probability for determining whether the recommended service associated with the third probability can be determined as the third recommended service associated with the candidate resource, and may be preset, which is not limited in the present disclosure.

[0127] In the present disclosure, after obtaining the third probabilities corresponding to each candidate resource, when at least two of the third probabilities corresponding to the candidate resources are greater than a threshold, the type of the candidate resource can be determined, and the degree of similarity with the resource types of the recommended services respectively associated with the at least two third probabilities is relatively high. At this time, the at least two recommended services associated with the at least two third probabilities can be determined as the third recommended services associated with the candidate resource, thereby improving the accuracy of the determined third recommended services associated with the candidate resource.

[0128] Step 505: Match the target recommended service with the third recommended service associated with each candidate resource in the candidate resource list.

[0129] Step 506: Determine the candidate resource associated with the third recommendation service that matches the target recommendation service as the target resource.

[0130] The specific implementation forms of step 505 to step 506 can refer to the detailed descriptions in other embodiments of the present disclosure, and will not be described in detail here.

[0131] Step 507: for each target resource, determine the first weight adjustment value corresponding to the target resource according to its corresponding maximum third probability value.

[0132] The first weight adjustment value may be used to reconfigure the recommended weight value of the target resource, which may be preset and is not limited in the present disclosure.

[0133] In the present disclosure, after determining the target resources recommended to the user, in order to increase the probability of each target resource being recommended to the user, a first weight adjustment value corresponding to each target resource may be determined according to its corresponding maximum third probability value.

[0134] It should be noted that when determining the first weight adjustment value corresponding to each target resource based on the maximum third probability value corresponding to each target resource, the first weight adjustment value corresponding to each target resource may be different. For example, the larger the maximum third probability value corresponding to the target resource is, the more similar the type of the target resource can be to the resource type of the target recommendation service preferred by the user, and the higher the first weight adjustment value of the target resource may be, and the present disclosure does not limit this.

[0135] Step 508: Increase the recommendation weight of the target resource by the first weight adjustment value.

[0136] In the present disclosure, after determining the first weight adjustment value corresponding to the target resource, the recommendation weight of the target resource can be increased by the first weight adjustment value to complete the reconfiguration of the recommendation weight of the target resource, thereby increasing the probability of the target resource being successfully recommended to the user.

[0137] Optionally, after determining the target resources, it is also possible to determine, for each target resource, a second weight adjustment value corresponding to the target resource based on the number of third probability values ​​corresponding to the target resource that are greater than a threshold, and then increase the recommendation weight of the target resource by the second weight adjustment value to complete the reconfiguration of the recommendation weight of the target resource. For example, the more third probability values ​​greater than the threshold that correspond to the target resource, the higher the second weight adjustment value by which the recommendation weight of the target resource is increased, and so on, thereby increasing the probability of the target resource being recommended to the user and improving the reliability of content recommendation. The present disclosure does not limit this.

[0138] Optionally, after determining the target resources, the recommendation weight of each target resource can be directly increased by a preset fixed value for each target resource to complete the reconfiguration of the recommendation weight of each target resource, thereby increasing the probability of each target resource being recommended to the user. The present disclosure does not limit this.

[0139] Step 509: Recommend content to the user based on the reconfigured recommendation weights.

[0140] In the present disclosure, after the recommendation weight of the target resource is reconfigured by increasing the recommendation weight of the target resource by a first weight adjustment value, content can be recommended to the user based on the reconfigured recommendation weight, thereby improving the diversity of content recommendations.

[0141] The specific implementation of step 509 can refer to the detailed description in other embodiments of the present disclosure, and will not be described in detail here.

[0142] Combine the following Figure 6 , the content recommendation method based on the large model provided by the present disclosure is illustrated by way of example, Figure 6 A flowchart of a large model-based content recommendation method provided in an embodiment of the present disclosure.

[0143] like Figure 6 As shown, the content recommendation method based on the large model provided by the present disclosure can be divided into two parts, namely, an offline part and an online part. Among them, LLM is the abbreviation of Large Language Model (LLM).

[0144] In the offline part, the candidate resources may be first input into each recognition model to obtain each third probabilities corresponding to the candidate resources, and then based on each third probabilities of the candidate resources, the recommended services associated with the candidate resources may be determined to obtain a mapping table of candidate resources and recommended services.

[0145] In the online part, first obtain the user's historical consumption resources and the second recommendation service installed in his device, and determine the target recommendation service through the second recommendation service and the historical consumption resources. Then, determine the recommendation service to which the candidate resource belongs through a mapping table based on candidate resources and recommendation services. When the recommendation service to which the candidate resource belongs matches the target recommendation service, the candidate resource is weighted (reconfigured recommendation weight) and recommended to the user.

[0146] Among them, the recognition model can be composed of a deep neural network and an LLM fine-tuning model, which is not limited in the present disclosure.

[0147] Among them, the recognition model can be trained through different recommendation services and their corresponding resources to obtain the recognition models associated with different recommendation services, which is not limited in the present disclosure.

[0148] In the disclosed embodiment, the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device are first obtained, and the historical consumption resources are identified based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service to determine the target recommendation service associated with the historical consumption resources. Then, for each candidate resource in the candidate resource column, each candidate resource is input into each recognition model to obtain each third probability corresponding to the candidate resource. When at least two third probabilities among the third probabilities are greater than a threshold, at least two recommendation services respectively associated with the at least two third probabilities are determined as the third recommendation service associated with the candidate resource. Then, the target recommendation service is matched with the third recommendation service associated with each candidate resource in the candidate resource list, and the candidate resource associated with the third recommendation service matched with the target recommendation service is determined as the target resource. Finally, for each target resource, the first weight adjustment value corresponding to the target resource is determined according to its corresponding maximum third probability value, and the recommendation weight of the target resource is increased by the first weight adjustment value. Based on the reconfigured recommendation weight, content is recommended to the user. Therefore, after determining the target recommendation service associated with the user's historical consumption resources through identification models associated with different recommendation services, each identification model is used to determine the probabilities corresponding to each candidate resource, and the recommendation service associated with a probability greater than a threshold is determined as the recommendation service associated with the corresponding candidate resource. The target resource recommended to the user is determined by matching the target recommendation service with the recommendation service associated with each candidate resource, and the weight adjustment value corresponding to the target resource is determined based on the maximum third probability value corresponding to the target resource. The recommendation weight of the target resource is reconfigured, and content is recommended to the user based on the reconfigured recommendation weight, thereby improving the diversity and reliability of content recommendations.

[0149] In order to implement the above embodiment, the present disclosure also proposes a content recommendation device based on a large model.

[0150] Figure 7 A schematic diagram of the structure of a large model-based content recommendation device provided in an embodiment of the present disclosure.

[0151] like Figure 7 As shown, the large model-based content recommendation device 700 includes: an acquisition module 701, an identification module 702, a traversal module 703, a configuration module 704, and a recommendation module 705.

[0152] An acquisition module 701 is used to acquire the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the device of the user;

[0153] An identification module 702, configured to identify the historical consumption resource based on a first identification model associated with the first recommendation service and a second identification model associated with the second recommendation service, so as to determine a target recommendation service associated with the historical consumption resource;

[0154] A traversal module 703, configured to traverse a candidate resource list based on the target recommendation service to determine a target resource associated with the target recommendation service;

[0155] A configuration module 704, configured to reconfigure the recommendation weight of the target resource, wherein the recommendation weight after reconfiguration is greater than the recommendation weight before reconfiguration;

[0156] The recommendation module 705 is used to recommend content to the user based on the reconfigured recommendation weights.

[0157] In a possible implementation of the present disclosure, the recognition model corresponding to the second recommendation service is obtained by taking the first resource set associated with the first recommendation service as a negative sample and taking the second resource set associated with the second recommendation service as a positive sample.

[0158] In a possible implementation of the present disclosure, the identification module 702 is specifically configured to:

[0159] Inputting the historical consumption resources into the first recognition model and the second recognition model respectively to obtain a first probability output by the first recognition model and a second probability output by the second recognition model;

[0160] The recommended service associated with the maximum value of the first probability and the second probability is determined as the target recommended service.

[0161] In a possible implementation of the present disclosure, the identification module 702 is further configured to:

[0162] Determine at least one candidate recommended service corresponding to each of the historical consumption resources;

[0163] At least one candidate recommended service with the highest appearance frequency among the candidate recommended services respectively corresponding to the at least two historical consumption resources is determined as the target recommended service.

[0164] In a possible implementation of the present disclosure, the traversal module 703 is specifically used to:

[0165] Matching the target recommended service with a third recommended service associated with each candidate resource in the candidate resource list;

[0166] The candidate resource associated with the third recommendation service matching the target recommendation service is determined as the target resource.

[0167] In a possible implementation of the present disclosure, the traversal module 703 is further used to:

[0168] For each of the candidate resources, input it into each recognition model to obtain each third probability corresponding to the candidate resource;

[0169] The recommended service associated with the recognition model corresponding to the maximum value among the third probabilities is determined as the third recommended service associated with the candidate resource.

[0170] In a possible implementation of the present disclosure, the traversal module 703 is further used to:

[0171] When at least two of the third probabilities are greater than a threshold, at least two recommended services respectively associated with the at least two third probabilities are determined as third recommended services associated with the candidate resource.

[0172] In a possible implementation of the present disclosure, the traversal module 703 is further used to:

[0173] For each of the target resources, determine a first weight adjustment value corresponding to the target resource according to its corresponding maximum third probability value;

[0174] The recommended weight of the target resource is increased by the first weight adjustment value.

[0175] In a possible implementation of the present disclosure, the traversal module 703 is further used to:

[0176] For each of the target resources, determining a second weight adjustment value corresponding to the target resource according to the number of third probability values ​​corresponding to the target resource that are greater than a threshold;

[0177] The recommendation weight of the target resource is increased by the second weight adjustment value.

[0178] The functions and specific implementation principles of the above modules in the embodiments of the present disclosure can be referred to the above method embodiments, and will not be repeated here.

[0179] In the disclosed embodiment, the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the user's device are first obtained, and then the historical consumption resources are identified based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service, respectively, to determine the target recommendation service associated with the historical consumption resources, and then based on the target recommendation service, the candidate resource list is traversed to determine the target resource associated with the target recommendation service, and finally the recommendation weight of the target resource is reconfigured, and content is recommended to the user based on the reconfigured recommendation weight. Thus, by identifying the historical consumption resources of the user based on the recognition models associated with different recommendation services, determining the target recommendation service, and based on the target recommendation service, traversing the candidate resource list to determine the target resource to be recommended to the user, reconfiguring the recommendation weight of the target resource, and based on the reconfigured recommendation weight, content is recommended to the user, thereby realizing content recommendation in a personalized atmosphere for the user and improving the diversity and reliability of content recommendation.

[0180] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0181] Figure 8 A schematic block diagram of an example electronic device 800 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0182] like Figure 8 As shown, the device 800 includes a computing unit 801, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the device 800 can also be stored. The computing unit 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0183] A number of components in the device 800 are connected to the I / O interface 805, including: an input unit 806, such as a keyboard, a mouse, etc.; an output unit 807, such as various types of displays, speakers, etc.; a storage unit 808, such as a disk, an optical disk, etc.; and a communication unit 809, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 809 allows the device 800 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0184] The computing unit 801 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 801 performs the various methods and processes described above, such as a content recommendation method based on a large model. For example, in some embodiments, the content recommendation method based on a large model may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 808. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of the content recommendation method based on the large model described above may be performed. Alternatively, in other embodiments, the computing unit 801 may be configured to perform a content recommendation method based on a large model in any other appropriate manner (e.g., by means of firmware).

[0185] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0186] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable large-model-based content recommendation device, so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0187] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0188] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0189] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), the Internet, and a blockchain network.

[0190] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server of a distributed system, or a server combined with a blockchain.

[0191] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0192] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. In the description of the present disclosure, the words "if" and "if" used can be interpreted as "at the time of" or "when" or "in response to determination" or "under the circumstances of".

[0193] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A content recommendation method based on a large model, comprising: Obtaining historical consumption resources of a user in a first recommendation service and a second recommendation service installed in a device of the user; Based on a first recognition model associated with the first recommendation service and a second recognition model associated with the second recommendation service, the historical consumption resource is identified to determine a target recommendation service associated with the historical consumption resource; Based on the target recommendation service, traverse the candidate resource list to determine the target resource associated with the target recommendation service; reconfiguring the recommendation weight of the target resource, wherein the recommendation weight after the reconfiguration is greater than the recommendation weight before the reconfiguration; Based on the reconfigured recommendation weights, content is recommended to the user.

2. The method of claim 1, wherein: The recognition model corresponding to the second recommendation service is obtained by taking the first resource set associated with the first recommendation service as a negative sample and taking the second resource set associated with the second recommendation service as a positive sample.

3. The method of claim 1, wherein: The identifying the historical consumption resource based on the first recognition model associated with the first recommendation service and the second recognition model associated with the second recommendation service to determine the target recommendation service associated with the historical consumption resource includes: Inputting the historical consumption resources into the first recognition model and the second recognition model respectively to obtain a first probability output by the first recognition model and a second probability output by the second recognition model; The recommended service associated with the maximum value of the first probability and the second probability is determined as the target recommended service.

4. The method of claim 3, wherein: The number of the historical consumption resources is at least two, and the determining of the target recommendation service associated with the historical consumption resources includes: Determine at least one candidate recommended service corresponding to each of the historical consumption resources; At least one candidate recommended service with the highest appearance frequency among the candidate recommended services respectively corresponding to the at least two historical consumption resources is determined as the target recommended service.

5. The method of claim 1, wherein: Based on the target recommendation service, a candidate resource list is traversed to determine a target resource associated with the target recommendation service, including: Matching the target recommended service with a third recommended service associated with each candidate resource in the candidate resource list; The candidate resource associated with the third recommendation service matching the target recommendation service is determined as the target resource.

6. The method of claim 5, wherein: Before traversing the candidate resource list based on the target recommendation service, the method further includes: For each of the candidate resources, input it into each recognition model to obtain each third probability corresponding to the candidate resource; The recommended service associated with the recognition model corresponding to the maximum value among the third probabilities is determined as the third recommended service associated with the candidate resource.

7. The method of claim 6, wherein: After obtaining the third probabilities corresponding to the candidate resources, the method further includes: When at least two of the third probabilities are greater than a threshold, at least two recommended services respectively associated with the at least two third probabilities are determined as third recommended services associated with the candidate resource.

8. The method of claim 7, wherein: The reconfiguring the recommended weight of the target resource includes: For each of the target resources, determine a first weight adjustment value corresponding to the target resource according to its corresponding maximum third probability value; The recommended weight of the target resource is increased by the first weight adjustment value.

9. The method of claim 7, wherein: The reconfiguring the recommended weight of the target resource includes: For each of the target resources, determining a second weight adjustment value corresponding to the target resource according to the number of third probability values ​​corresponding to the target resource that are greater than a threshold; The recommendation weight of the target resource is increased by the second weight adjustment value.

10. A content recommendation device based on a large model, wherein: The device comprises: An acquisition module, used to acquire the historical consumption resources of the user in the first recommendation service and the second recommendation service installed in the device of the user; an identification module, configured to identify the historical consumption resource based on a first identification model associated with the first recommendation service and a second identification model associated with the second recommendation service, so as to determine a target recommendation service associated with the historical consumption resource; A traversal module, used to traverse the candidate resource list based on the target recommendation service to determine the target resource associated with the target recommendation service; A configuration module, used to reconfigure the recommendation weight of the target resource, wherein the recommendation weight after reconfiguration is greater than the recommendation weight before reconfiguration; A recommendation module is used to recommend content to the user based on the reconfigured recommendation weight.

11. The device of claim 10, wherein: The recognition model corresponding to the second recommendation service is obtained by taking the first resource set associated with the first recommendation service as a negative sample and taking the second resource set associated with the second recommendation service as a positive sample.

12. The device of claim 10, wherein: The identification module is specifically used for: Inputting the historical consumption resources into the first recognition model and the second recognition model respectively to obtain a first probability output by the first recognition model and a second probability output by the second recognition model; The recommended service associated with the maximum value of the first probability and the second probability is determined as the target recommended service.

13. The device of claim 12, wherein: The identification module is further used for: Determine at least one candidate recommended service corresponding to each of the historical consumption resources; At least one candidate recommended service with the highest appearance frequency among the candidate recommended services respectively corresponding to the at least two historical consumption resources is determined as the target recommended service.

14. The device of claim 10, wherein: The traversal module is specifically used for: Matching the target recommended service with a third recommended service associated with each candidate resource in the candidate resource list; The candidate resource associated with the third recommendation service matching the target recommendation service is determined as the target resource.

15. The method of claim 14, wherein: The traversal module is also used to: For each of the candidate resources, input it into each recognition model to obtain each third probability corresponding to the candidate resource; The recommended service associated with the recognition model corresponding to the maximum value among the third probabilities is determined as the third recommended service associated with the candidate resource.

16. The device of claim 15, wherein: The traversal module is also used to: When at least two of the third probabilities are greater than a threshold, at least two recommended services respectively associated with the at least two third probabilities are determined as third recommended services associated with the candidate resource.

17. The device of claim 16, wherein: The traversal module is also used to: For each of the target resources, determine a first weight adjustment value corresponding to the target resource according to its corresponding maximum third probability value; The recommended weight of the target resource is increased by the first weight adjustment value.

18. The device of claim 16, wherein: The traversal module is also used to: For each of the target resources, determining a second weight adjustment value corresponding to the target resource according to the number of third probability values ​​corresponding to the target resource that are greater than a threshold; The recommendation weight of the target resource is increased by the second weight adjustment value.

19. An electronic device, characterized in that: include: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores instructions that may be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1 to 9.

20. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-9.

21. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 9.

Images