Object recommendation method, device, electronic device and storage medium

Abstract

The present disclosure relates to an object recommendation method, device, electronic device and storage medium. The method includes: obtaining multiple candidate objects in a target application, a first user set and a second user set; the registration duration of users in the first user set is less than or equal to a first preset duration and greater than a second preset duration, the registration duration of users in the second user set is less than or equal to the second preset duration, the first preset duration is greater than the second preset duration, and the registration duration is the duration from the user's registration in the target application to the current time; randomly recommending multiple candidate objects to users in the first user set, and obtaining first interaction information between users in the first user set and multiple candidate objects; determining a target object from multiple candidate objects based on first business indicator information determined by the first interaction information; and recommending the target object to users in the second user set. According to the technical solution provided by the present disclosure, the recommendation effect of cold-starting new users can be improved.

Description

Technical Field

[0001] The present disclosure relates to the field of Internet application technology, and in particular to an object recommendation method, device, electronic device, and storage medium. Background Art

[0002] With the development of internet applications, recommendation services have attracted considerable attention, and a variety of recommendation methods have emerged. Related technologies typically leverage users' historical behavior on online platforms to identify their interests, thereby providing personalized recommendations based on the matching degree between their interests and objects on the online platforms. However, users on online platforms register and join the platform gradually, resulting in a continuous flow of cold-start users. These users lack historical behavior, have inaccurate static features, and suffer from incomplete coverage. Existing personalized recommendation methods, when used to recommend objects to cold-start users, are easily dominated by the object's unilateral features, resulting in a preponderance of popular objects. This leads to poor recommendation effectiveness and a failure to capture user interest. Summary of the Invention

[0003] The present disclosure provides an object recommendation method, device, electronic device, and storage medium to at least solve the problem of how to improve the recommendation effect of cold-start new users in the related art. The technical solution of the present disclosure is as follows:

[0004] According to a first aspect of an embodiment of the present disclosure, there is provided an object recommendation method, comprising:

[0005] Acquire multiple candidate objects, a first user set, and a second user set in a target application; the registration duration of users in the first user set is less than or equal to a first preset duration and greater than a second preset duration, the registration duration of users in the second user set is less than or equal to the second preset duration, the first preset duration is greater than the second preset duration, and the registration duration is the duration from when the user was registered in the target application to the current time;

[0006] Randomly recommending the multiple candidate objects to users in the first user set, and obtaining first interaction information between the users in the first user set and the multiple candidate objects;

[0007] Determining, based on the first interaction information, first business indicator information of each of the multiple candidate objects under multiple business indicators;

[0008] Determining a target object from the multiple candidate objects according to the first business indicator information;

[0009] The target object is recommended to users in the second user set.

[0010] In a possible implementation, the step of obtaining multiple candidate objects in the target application includes:

[0011] Acquire multiple objects and a third user set in a target application, wherein the registration duration of users in the third user set is greater than a first preset duration;

[0012] determining, based on second interaction information between users in the third user set and the multiple objects, second business indicator information of each of the multiple objects under multiple business indicators;

[0013] The plurality of candidate objects are determined from the plurality of objects according to the second business indicator information.

[0014] In a possible implementation, the step of determining a target object from the multiple candidate objects according to the first business indicator information includes:

[0015] Determining first recommendation parameter information for each of the plurality of candidate objects based on the first business indicator information;

[0016] Obtaining an object category of each of the plurality of candidate objects;

[0017] Based on the object categories, determining a subset of candidate objects under each object category;

[0018] Based on the first recommendation parameter information, the target objects are respectively screened out from the candidate object subsets under each object category.

[0019] In a possible implementation, the step of determining the multiple candidate objects from the multiple objects according to the second business indicator information includes:

[0020] Determining second recommendation parameter information for each of the plurality of candidate objects according to the second business indicator information;

[0021] Obtaining an object category of each of the plurality of objects;

[0022] determining, based on the respective object categories of the plurality of objects, a subset of objects under each object category;

[0023] Based on the second recommendation parameter information, the multiple candidate objects are screened out from the object subsets under each object category.

[0024] In a possible implementation, the second business indicator information represents a correlation between the number of times the plurality of objects execute preset actions within a preset time period and the exposure of the plurality of objects, and the step of obtaining the plurality of objects in the target application includes:

[0025] Obtaining the exposure corresponding to each object in the target application;

[0026] The objects whose exposure amounts are greater than the exposure amount threshold are regarded as the multiple objects.

[0027] In a possible implementation, the step of obtaining multiple objects in the target application includes:

[0028] Obtaining release duration information of each object in the target application;

[0029] The objects whose publishing duration information is less than a preset duration threshold are regarded as the multiple objects.

[0030] In a possible implementation, the step of randomly recommending the multiple candidate objects to users in the first user set includes:

[0031] Obtaining objects to be recommended that match each user in the first user set;

[0032] sorting the objects to be recommended and the multiple candidate objects matched by each user respectively to obtain a sorting result corresponding to each user in the first user set, wherein the multiple candidate objects in the sorting result are sorted at intervals within the objects to be recommended;

[0033] Based on the ranking result, the objects to be recommended and the candidate objects that match each user are recommended to each user in the first user set.

[0034] According to a second aspect of an embodiment of the present disclosure, there is provided an object recommendation device, comprising:

[0035] An acquisition module is configured to execute acquisition of a plurality of candidate objects, a first user set, and a second user set in a target application; the registration duration of users in the first user set is less than or equal to a first preset duration and greater than a second preset duration, the registration duration of users in the second user set is less than or equal to the second preset duration, the first preset duration is greater than the second preset duration, and the registration duration is the duration from when the user was registered in the target application to the current time;

[0036] A random recommendation module is configured to randomly recommend the multiple candidate objects to users in the first user set, and obtain first interaction information between the users in the first user set and the multiple candidate objects;

[0037] A first business indicator information determining module is configured to determine first business indicator information of each of the plurality of candidate objects under a plurality of business indicators based on the first interaction information;

[0038] a target object determining module, configured to determine a target object from the plurality of candidate objects according to the first business indicator information;

[0039] The recommendation module is configured to recommend the target object to users in the second user set.

[0040] In a possible implementation, the acquisition module includes:

[0041] An acquiring unit is configured to acquire a plurality of objects in a target application and a third user set, wherein a registration duration of users in the third user set is greater than a first preset duration;

[0042] A second business indicator information determining unit is configured to determine second business indicator information of each of the multiple objects under multiple business indicators based on second interaction information between users in the third user set and the multiple objects;

[0043] The candidate object determining unit is configured to determine the multiple candidate objects from the multiple objects according to the second business indicator information.

[0044] In a possible implementation, the target object determination module includes:

[0045] A first recommendation parameter information determining unit is configured to determine first recommendation parameter information of each of the plurality of candidate objects according to the first business indicator information;

[0046] an object category acquiring unit, configured to acquire the object category of each of the plurality of candidate objects;

[0047] a candidate object subset determining unit, configured to determine a candidate object subset under each object category based on the object category;

[0048] The target object determining unit is configured to filter out the target object from the candidate object subsets under each object category based on the first recommendation parameter information.

[0049] In a possible implementation, the candidate object determining unit includes:

[0050] A second recommendation parameter information determining subunit is configured to determine second recommendation parameter information of each of the plurality of candidate objects according to the second business indicator information;

[0051] an object category determination subunit, configured to obtain the object category of each of the plurality of objects;

[0052] an object subset determining subunit, configured to determine an object subset under each object category based on the respective object categories of the plurality of objects;

[0053] The candidate object determination subunit is configured to filter out the multiple candidate objects from the object subsets under each object category based on the second recommendation parameter information.

[0054] In a possible implementation, the second business indicator information represents a correlation between the number of times the plurality of objects perform preset actions within a preset time period and the exposure amounts of the plurality of objects, and the acquiring unit includes:

[0055] an exposure acquisition subunit, configured to acquire the exposure corresponding to each object in the target application;

[0056] The multiple object determination subunit is configured to determine the objects whose exposure amounts are greater than an exposure amount threshold as the multiple objects.

[0057] In a possible implementation, the acquiring unit includes:

[0058] A duration information acquisition subunit is configured to acquire release duration information of each object in the target application in the target application;

[0059] The multiple object acquisition subunits are configured to take the objects whose publishing duration information is less than a preset duration threshold as the multiple objects.

[0060] In a possible implementation, the random recommendation module includes:

[0061] a to-be-recommended object acquiring unit, configured to execute acquiring to-be-recommended objects that match each user in the first user set;

[0062] a sorting unit configured to sort the objects to be recommended and the multiple candidate objects matched by each user, respectively, to obtain a sorting result corresponding to each user in the first user set, wherein the multiple candidate objects are sorted at intervals within the objects to be recommended in the sorting result;

[0063] The random recommendation unit is configured to recommend the objects to be recommended and the candidate objects that match each user to each user in the first user set based on the ranking result.

[0064] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement a method as described in any one of the first aspects above.

[0065] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is capable of executing any method described in the first aspect of the embodiment of the present disclosure.

[0066] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided, comprising computer instructions, which, when executed by a processor, enable a computer to execute any one of the methods according to the first aspect of the embodiment of the present disclosure.

[0067] The technical solutions provided by the embodiments of the present disclosure bring at least the following beneficial effects:

[0068] By recommending candidate objects to non-cold-start new users in a random recommendation manner, an unbiased recommendation distribution environment is constructed; on this basis, the first interaction information between the candidate objects and non-cold-start new users is used to determine the first business indicator information of each candidate object under multiple business indicators, which can ensure the universality of the target objects determined based on the first business indicator information among users, and effectively alleviate the phenomenon of biased recommendations and popular recommendations caused by the dominance of unilateral features of objects and the lack of static features in the recommendations for cold-start new users, thereby improving the recommendation effect of cold-start new users and improving the experience and retention of cold-start new users; and target objects with better universality can be effectively used to explore and capture user interests, providing a guarantee for accurate recommendations for cold-start new users.

[0069] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] The accompanying drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description are used to explain the principles of the present disclosure, and do not constitute an improper limitation of the present disclosure.

[0071] Figure 1 It is a schematic diagram showing an application environment according to an exemplary embodiment.

[0072] Figure 2 The figure is a flowchart of a method for recommending an object according to an exemplary embodiment.

[0073] Figure 3 The present invention is a flowchart of a method for randomly recommending multiple candidate objects to users in a first user set according to an exemplary embodiment.

[0074] Figure 4 The present invention is a flowchart of a method for obtaining multiple candidate objects in a target application according to an exemplary embodiment.

[0075] Figure 5 The figure is a flow chart of a method for acquiring multiple objects in a target application according to an exemplary embodiment.

[0076] Figure 6 The present invention is a flowchart showing a method for determining a target object from multiple candidate objects based on first business indicator information according to an exemplary embodiment.

[0077] Figure 7 The present invention is a flowchart illustrating a method for determining multiple candidate objects from multiple objects based on second business indicator information according to an exemplary embodiment.

[0078] Figure 8 The figure is a block diagram of a device for recommending an object according to an exemplary embodiment.

[0079] Figure 9 It is a block diagram of an electronic device for object recommendation according to an exemplary embodiment.

[0080] Figure 10 is a block diagram of another electronic device for object recommendation according to an exemplary embodiment. DETAILED DESCRIPTION

[0081] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0082] It should be noted that the terms "first," "second," and the like in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the numbers used in this manner are interchangeable where appropriate so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure as detailed in the appended claims.

[0083] See also Figure 1 , Figure 1 is a schematic diagram showing an application environment according to an exemplary embodiment. Figure 1 As shown, the application environment may include a server 01 and a terminal 02 .

[0084] In an optional embodiment, server 01 can be used for object recommendation processing. Specifically, server 01 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0085] In an optional embodiment, terminal 02 can be used to display recommended objects. Specifically, terminal 02 may include, but is not limited to, electronic devices such as smartphones, desktop computers, tablet computers, laptop computers, smart speakers, digital assistants, augmented reality (AR) / virtual reality (VR) devices, and smart wearable devices. Optionally, the operating system running on the electronic device may include, but is not limited to, Android, iOS, Linux, and Windows.

[0086] In addition, it should be noted that Figure 1 What is shown is only one application environment of the image processing method provided by the present disclosure.

[0087] In the embodiments of this specification, the server 01 and the terminal 02 may be directly or indirectly connected via wired or wireless communication, which is not limited in this application.

[0088] It should be noted that the following figure shows a possible order of steps, which is not actually limited to this order. Some steps can be executed in parallel without relying on each other. The user information (including but not limited to user device information, user personal information, user behavior information, etc.) and data (including but not limited to data for display, data for training, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties.

[0089] Figure 2 FIG. 1 is a flow chart of an object recommendation method according to an exemplary embodiment. Figure 2 As shown, the following steps may be included.

[0090] In step S201, multiple candidate objects, a first user set, and a second user set in the target application are obtained; the registration duration of the users in the first user set can be less than or equal to the first preset duration and greater than the second preset duration, and the registration duration of the users in the second user set can be less than or equal to the second preset duration, the first preset duration is greater than the second preset duration, and the registration duration is the duration from the user's registration in the target application to the current time, that is, the duration from the user's registration time in the target application to the current time.

[0091] In actual applications, candidate objects may be all or part of the objects in the target application, and the objects in the target application may refer to objects that can be displayed in the target application, for example, they may include multimedia, advertisements, news, etc.; wherein, multimedia may include short videos, long videos, graphic information, etc., which is not limited in this disclosure. The users in the first user set may refer to non-cold-start new users, that is, new users in a non-cold-start state; the users in the second user set may refer to cold-start users, that is, new users in a cold-start state. In an example, the first preset time period may be 14 days and the second preset time period may be 10 hours, which is not limited in this disclosure. Accordingly, users whose registration time is longer than the first preset time period may be regarded as old users in the target application.

[0092] Based on the above-mentioned old users being in a non-cold start state, optionally, the users in the first user set may refer to old users in a non-cold start state and new users in a non-cold start state, that is, users whose registration time is greater than the second preset time.

[0093] In the embodiments of this specification, the registration duration of each user in the target application can be obtained, and thus a first user set and a second user set can be determined based on the relationship between the registration duration and the first preset duration and the second preset duration. Multiple candidate objects can also be obtained from the target application for object recommendation. For example, objects published in the target application within a recent period can be selected as multiple candidate objects, although this disclosure is not limited to this.

[0094] In step S203, a plurality of candidate objects are randomly recommended to users in the first user set, and first interaction information between the users in the first user set and the plurality of candidate objects is obtained.

[0095] In actual applications, multiple candidate objects can be randomly recommended to users in the first user set. For example, multiple candidate objects can be randomly recommended to each user in the first user set until multiple candidate objects are recommended once; or, multiple candidate objects can be randomly divided into a certain number of object sets, which can be the number of users in the first user set, so that objects in an object set can be sequentially recommended to a user in the first user set until objects in each object set are recommended once.

[0096] Optionally, in the random recommendation process, the multimedia to be recommended that is personalized recommended by each user in the first user set can be combined to achieve random recommendation of multiple candidate objects. In one example, Figure 3 As shown, step S203 may include:

[0097] In step S301, objects to be recommended that match each user in the first user set are obtained;

[0098] In step S303, the objects to be recommended and the multiple candidate objects matched by each user are sorted respectively to obtain a sorting result corresponding to each user in the first user set, in which the multiple candidate objects are sorted at intervals in the objects to be recommended;

[0099] In step S305 , based on the ranking result, the objects to be recommended and the candidate objects that match each user are recommended to each user in the first user set.

[0100] In actual applications, personalized recommendations can be made to each user in the first user set. For example, objects associated with the preferences of each user in the first user set can be obtained as objects to be recommended that match each user. This is not limited in the present disclosure. Thus, the objects to be recommended and multiple candidate objects can be sorted to obtain a sorting result, in which multiple candidate objects are sorted at intervals in the objects to be recommended. For example, a candidate object can be inserted between every 10 objects to be recommended. In this way, when recommending multimedia to be recommended and candidate multimedia to each user in the first user set based on the sorting result, it can be ensured that the objects recommended to non-cold-start new users include objects that match their own preferences, thereby achieving random recommendation of candidate objects. This can ensure the universality of the target object and the recommendation experience of candidate objects for non-cold-start new users. In addition, the interests of non-cold-start new users can be explored and mined through candidate objects, so that the interests of non-cold-start new users can be effectively explored.

[0101] In actual applications, after randomly recommending multiple candidate objects to users in the first user set, first interaction information between the users in the first user set and the multiple candidate objects can be obtained. The first interaction information can be interaction information within a first preset time period, the start time of the first preset time period can be the time corresponding to the random recommendation, and the length of the first preset time period can be pre-set, such as 10 minutes, etc., which is not limited in this disclosure. Among them, the first interaction information can refer to the associated information of the users in the first user set performing preset actions on the candidate objects within the first preset time period. The associated information here can refer to the number of times a candidate object is subjected to preset actions by users in the first user set. The preset actions here can include clicks, likes, follows, comments, and entering the author's personal page of the object, etc., which is not limited in this disclosure.

[0102] In step S205 , based on the first interaction information, first business indicator information of each of the multiple candidate objects under multiple business indicators is determined.

[0103] In the embodiments of this specification, a business indicator may correspond to a preset behavior. For example, a business indicator may represent the correlation between the number of times a preset behavior is performed and the exposure amount, such as a click indicator, a like indicator, a follow indicator, a comment indicator, and an indicator of the author's personal page of the object. Accordingly, the first business indicator information of a candidate object under a business indicator may represent the correlation between the number of times a preset behavior is performed on the candidate object and the exposure amount of the candidate object in the first user set. The correlation may be the ratio of the number of times to the exposure amount, such as the click rate, the like rate, the follow rate, the long broadcast rate, the comment rate, and the probability of entering the author's personal page of the object, etc. This disclosure is not limited to this. Among them, the above-mentioned preset behavior may correspond to a business indicator. It should be noted that, corresponding to the first interaction information, the exposure amount here is also the exposure amount within the first preset time period. The above-mentioned first business indicator information under multiple business indicators may refer to the first business indicator information corresponding to multiple business indicators respectively, that is, each business indicator has corresponding first business indicator information.

[0104] For example, taking click-through rate as an example, assuming that short video H is exposed to 100 users in the first user set, and 50 of these 100 users click on the short video H, it can be determined that the click-through rate is 50 / 100=50%.

[0105] In step S207, a target object is determined from a plurality of candidate objects according to the first business indicator information.

[0106] In practical applications, multiple candidate objects can be ranked according to the weighted sum of the first business indicator information, and the target object can be determined based on the ranking. For example, if the top ranking indicates better business indicator information, a preset number of candidate objects with the top ranking can be used as target objects.

[0107] In step S209 , the target object is recommended to users in the second user set.

[0108] In the embodiments of this specification, target objects can be randomly recommended to users in the second user set, and this disclosure is not limited to this. Optionally, target objects can also be sorted based on user attribute information of users in the second user set, so as to sequentially recommend target objects to users in the second user set.

[0109] By recommending candidate objects to non-cold-start new users in a random recommendation manner, an unbiased recommendation distribution environment is constructed; on this basis, the first interaction information between the candidate objects and non-cold-start new users is used to determine the first business indicator information of each candidate object under multiple business indicators, which can ensure the universality of the target objects determined based on the first business indicator information among users, and effectively alleviate the phenomenon of biased recommendations and popular recommendations caused by the dominance of unilateral features of objects and the lack of static features in the recommendations for cold-start new users, thereby improving the recommendation effect of cold-start new users and improving the experience and retention of cold-start new users; and target objects with better universality can be effectively used to explore and capture user interests, providing a guarantee for accurate recommendations for cold-start new users.

[0110] Figure 4 FIG. 1 is a flow chart showing a method for obtaining multiple candidate objects in a target application according to an exemplary embodiment. Figure 4 As shown, in a possible implementation, step S201 may include:

[0111] In step S401, a plurality of objects in a target application and a third user set are obtained, wherein the registration duration of users in the third user set is greater than a first preset duration.

[0112] In the embodiments of this specification, a third set of users in the target application can be obtained. For example, users whose registration duration in the target application is longer than a first preset duration can be obtained to form the third set of users. Users whose registration duration is longer than the first preset duration can be old users of the target application. As an example, the first preset duration can be 14 days, which is not limited in this disclosure.

[0113] This disclosure does not limit the method for obtaining multiple objects in the target application. In one possible implementation, the second business indicator information represents the correlation between the number of times the multiple objects perform preset actions within a preset time period and the exposure of the multiple objects. Accordingly, the step of obtaining multiple objects in the target application may include:

[0114] Get the exposure corresponding to each object in the target application;

[0115] Objects whose exposure is greater than the exposure threshold are counted as multiple objects.

[0116] In actual applications, the exposure corresponding to each object may refer to the total number of users to whom each object was displayed in the target application within a second preset time period. For example, if a short video H is displayed to 1,000 users, the exposure corresponding to short video H is 1,000. The second preset time period may refer to the period from a preset time point to the current time. The preset time point may be the registration time of the object in the target application, or may be a time point at a predetermined distance from the current time, such as a time point 48 hours from the current time. This disclosure is not limited to this.

[0117] As an example, the exposure corresponding to each object in the target application can be counted, so that objects with exposure greater than the exposure threshold can be counted as multiple objects to ensure the confidence of the second business indicator information, such as ensuring the confidence of the denominator of click-through rate, like rate, etc.

[0118] In another possible implementation, Figure 5 As shown, the steps of obtaining multiple objects in the target application may include:

[0119] In step S501, the publishing duration information of each object in the target application is obtained;

[0120] In step S503, the objects whose publishing duration information is less than a preset duration threshold are regarded as multiple objects.

[0121] In practical applications, considering that most objects in the target application are time-sensitive, such as short videos with strong timeliness and short life cycles, multiple objects can be obtained through time information. For example, the creation time or release time of each object in the target application can be obtained, so that the length of time from the creation time or release time to the current time can be determined as the release duration information, and objects whose release duration information is less than a preset duration threshold can be regarded as multiple objects, so as to obtain objects created in the most recent time period as multiple objects. Among them, the preset duration threshold can be set by means of statistical testing, and this disclosure does not limit this. By treating objects whose release duration information is less than the preset duration threshold as multiple objects, the timeliness of multiple objects is better, and the universality of the objects is further guaranteed.

[0122] In step S403, based on the second interaction information between the users in the third user set and the multiple objects, second business indicator information of each of the multiple objects under the multiple business indicators is determined.

[0123] The second interaction information may be interaction information within a third preset time period. The second interaction information may refer to associated information regarding preset actions performed on the multiple objects by users in the third user set within the third preset time period, such as the number of times the preset actions were performed on each of the multiple objects by users in the third user set within the third preset time period. The end time of the third preset time period may be the current time, and the duration of the third preset time period may be preset, which is not limited in this disclosure, for example, one month from the current time.

[0124] In step S405, multiple candidate objects are determined from the multiple objects according to the second business indicator information.

[0125] The implementation of the above steps S403 to S405 can refer to the above steps S205 to S207, which will not be repeated here.

[0126] Through the second interaction information between old users and objects in the target application, the second business indicator information of each object under multiple business indicators is determined, and based on the second business indicator information, multiple candidate objects are determined from multiple objects, thereby ensuring that the candidate objects themselves have better universality in the personalized recommendation for old users; and then when such candidate objects are used to make random recommendations to non-cold-start new users, the experience of random recommendations for non-cold-start new users can be guaranteed, that is, it can ensure unbiased recommendation prediction of candidate objects among non-cold-start new users to improve the recommendation effect for cold-start users, and can reduce the impact of random recommendations on the experience of non-cold-start new users.

[0127] Figure 6 This is a flow chart of a method for determining a target object from multiple candidate objects based on first business indicator information according to an exemplary embodiment. Figure 6 As shown, in a possible implementation, step S207 may include:

[0128] In step S601, first recommended parameter information of each of a plurality of candidate objects is determined according to first business indicator information.

[0129] In one example, taking the candidate object as a video, the weighted sum of the first business indicator information can be used as the first recommendation parameter information. For example, the following formula (1) can be used to determine the first recommendation parameter information Score of a candidate object:

[0130] Score=a*ctr+b*ltr+c*wtr+d*lvtr+e*cmtr+f*pptr(1)

[0131] Among them, CTR can refer to click-through rate, LTR can refer to like rate, WTR can refer to follow rate, LVTR can refer to long-play rate, CMTR can refer to comment rate, and PPTR can refer to the probability of entering a personal page. Long-play rate can refer to the ratio of the number of times a video is played for a duration greater than the preset play time to the video's exposure. This disclosure does not limit the preset play time.

[0132] The above-mentioned click rate, like rate, attention rate, comment rate and probability of entering the personal page can be the first business indicator information; it should be noted that the above is only an example and does not limit the present disclosure.

[0133] In step S603 , the object category of each of the plurality of candidate objects is obtained.

[0134] In the embodiments of this specification, the object category may refer to the label category of the object in the target application, such as sports, animals, food, etc., which is not limited in this disclosure. Based on this, the label category of the candidate object can be obtained, thereby obtaining the object category of each of the multiple candidate objects.

[0135] In step S605 , based on the object categories, a subset of candidate objects under each object category is determined.

[0136] In practical applications, to ensure that target objects cover a wide range of object categories, target objects can be obtained by object category. Based on this, multiple candidate objects can be divided into different object categories to form candidate object subsets within each object category. The candidate objects in each candidate object subset have the same object category. For example, in the candidate object subset under sports, all candidate objects have the same object category, thus achieving object category-based clustering of multiple candidate objects.

[0137] In step S607 , based on the first recommendation parameter information, target objects are screened out from the candidate object subsets under each object category.

[0138] In the embodiments of this specification, the candidate objects in each candidate subset can be sorted based on the first recommendation parameter information to obtain a sorting result, and the target object can be screened out based on the sorting result. As an example, based on the first recommendation parameter information, candidate objects that meet preset conditions can be screened out from the candidate object subsets under each object category as target objects. For example, candidate target objects that meet preset conditions can be screened out based on the sorting result. The preset conditions may include a recommendation parameter threshold, or may refer to a preset number of candidate objects that are ranked high in the sorting result, etc.

[0139] For example, the candidate subset for sports includes 10 candidates: H1-H10. Based on the first recommendation parameter information of these 10 candidates, the ranking results are: H5, H3, H6, H7, H8, H1, H4, H2, H10, H9. The top five candidates can be selected as the target objects: H5, H3, H6, H7, H8.

[0140] By filtering target objects from each object category separately, we can avoid the target objects from being concentrated in certain object categories that naturally have a higher recommendation priority, so that the target objects can cover a richer range of object categories, thereby further ensuring the universality of the target objects from the object category perspective; and, by recommending such target objects to cold-start new users, we can quickly and fully explore the interests of cold-start new users.

[0141] Figure 7 FIG. 1 is a flow chart showing a method for determining multiple candidate objects from multiple objects based on second business indicator information according to an exemplary embodiment. Figure 7 As shown, in a possible implementation, step S207 may include:

[0142] In step S701, second recommendation parameter information of each of a plurality of candidate objects is determined according to the second business indicator information;

[0143] In step S703, the object category of each of the multiple objects is obtained;

[0144] In step S705, based on the respective object categories of the plurality of objects, a subset of objects under each object category is determined;

[0145] In step S707 , based on the second recommendation parameter information, multiple candidate objects are screened out from the object subsets under each object category.

[0146] In one example, based on the second recommendation parameter information, objects that meet preset conditions may be screened out from the object subsets under each object category as multiple candidate objects.

[0147] The implementation of steps S701 to S707 can refer to steps S601 to S607, which will not be described in detail here. By obtaining multiple candidate objects under each object category, the object category diversity of the multiple candidate objects can be ensured, which facilitates the subsequent cold start of new user interest capture.

[0148] Figure 8 FIG. 1 is a block diagram of an object recommendation device according to an exemplary embodiment. Figure 8 , the apparatus may include:

[0149] Acquisition module 801 is configured to execute acquisition of multiple candidate objects, a first user set, and a second user set in a target application; the registration duration of users in the first user set is less than or equal to a first preset duration and greater than a second preset duration, the registration duration of users in the second user set is less than or equal to the second preset duration, the first preset duration is greater than the second preset duration, and the registration duration is the duration from when the user was registered in the target application to the current time;

[0150] The random recommendation module 803 is configured to randomly recommend multiple candidate objects to users in the first user set, and obtain first interaction information between the users in the first user set and the multiple candidate objects;

[0151] The first business indicator information determining module 805 is configured to determine first business indicator information of each of the plurality of candidate objects under a plurality of business indicators based on the first interaction information;

[0152] The target object determination module 807 is configured to determine a target object from a plurality of candidate objects according to the first business indicator information;

[0153] The recommendation module 809 is configured to recommend the target object to users in the second user set.

[0154] By recommending candidate objects to non-cold-start new users in a random recommendation manner, an unbiased recommendation distribution environment is constructed; on this basis, the first interaction information between the candidate objects and non-cold-start new users is used to determine the first business indicator information of each candidate object under multiple business indicators, which can ensure the universality of the target objects determined based on the first business indicator information among users, and effectively alleviate the phenomenon of biased recommendations and popular recommendations caused by the dominance of unilateral features of objects and the lack of static features in the recommendations for cold-start new users, thereby improving the recommendation effect of cold-start new users and improving the experience and retention of cold-start new users; and target objects with better universality can be effectively used to explore and capture user interests, providing a guarantee for accurate recommendations for cold-start new users.

[0155] In a possible implementation, the obtaining module 801 may include:

[0156] An acquiring unit is configured to acquire a plurality of objects in a target application and a third user set, wherein a registration duration of users in the third user set is greater than a first preset duration;

[0157] A second business indicator information determining unit is configured to determine second business indicator information of each of the multiple objects under multiple business indicators based on second interaction information between users in the third user set and the multiple objects;

[0158] The candidate object determining unit is configured to determine a plurality of candidate objects from a plurality of objects according to the second business indicator information.

[0159] In a possible implementation, the target object determination module 807 may include:

[0160] A first recommendation parameter information determining unit is configured to determine first recommendation parameter information of each of the plurality of candidate objects according to the first business indicator information;

[0161] an object category obtaining unit configured to obtain an object category of each of a plurality of candidate objects;

[0162] a candidate object subset determining unit, configured to determine a candidate object subset under each object category based on the object category;

[0163] The target object determining unit is configured to filter out target objects from the candidate object subsets under each object category based on the first recommendation parameter information.

[0164] In a possible implementation, the candidate object determining unit may include:

[0165] A second recommendation parameter information determining subunit is configured to determine second recommendation parameter information of each of the plurality of candidate objects according to the second business indicator information;

[0166] The object category determination subunit is configured to obtain the object category of each of the plurality of objects;

[0167] The object subset determination subunit is configured to determine an object subset under each object category based on the respective object categories of the plurality of objects;

[0168] The candidate object determination subunit is configured to screen out a plurality of candidate objects from the object subsets under each object category based on the second recommendation parameter information.

[0169] In a possible implementation, when the second business indicator information represents a correlation between the number of times a preset behavior is performed on multiple objects within a preset time period and the exposure of the multiple objects, the acquiring unit may include:

[0170] An exposure acquisition subunit is configured to acquire the exposure corresponding to each object in the target application;

[0171] The multiple object determination subunit is configured to determine objects having exposure amounts greater than an exposure amount threshold as multiple objects.

[0172] In a possible implementation, the acquiring unit may include:

[0173] The duration information acquisition subunit is configured to acquire the duration information of each object in the target application;

[0174] The multiple object acquisition subunit is configured to execute the process of treating objects whose duration information is less than a preset duration threshold as multiple objects.

[0175] In a possible implementation, the random recommendation module 803 may include:

[0176] a to-be-recommended object acquiring unit, configured to execute acquiring to-be-recommended objects that match each user in the first user set;

[0177] A sorting unit is configured to sort the object to be recommended and the plurality of candidate objects to obtain a sorting result, in which the plurality of candidate objects are sorted at intervals within the object to be recommended;

[0178] The random recommendation unit is configured to recommend objects to be recommended and candidate objects to users in the first user set in sequence based on the sorting result.

[0179] Regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0180] Figure 9 is a block diagram of an electronic device for object recommendation according to an exemplary embodiment. The electronic device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 9 As shown. The electronic device includes a processor, a memory, a network interface, a display screen and an input device connected via a system bus. The processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the electronic device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for object recommendation is implemented. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the electronic device, or an external keyboard, touchpad or mouse, etc.

[0181] Those skilled in the art will understand that Figure 9The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present disclosure, and does not constitute a limitation on the electronic device to which the scheme of the present disclosure is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0182] Figure 10 is a block diagram of another electronic device for object recommendation according to an exemplary embodiment. The electronic device may be a server, and its internal structure diagram may be as shown in FIG. Figure 10 As shown. The electronic device includes a processor, a memory, and a network interface connected via a system bus. The processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the electronic device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, it implements a method for object recommendation.

[0183] Those skilled in the art will understand that Figure 10 The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present disclosure, and does not constitute a limitation on the electronic device to which the scheme of the present disclosure is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0184] In an exemplary embodiment, an electronic device is further provided, including: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement the object recommendation method in the embodiment of the present disclosure.

[0185] In an exemplary embodiment, a computer-readable storage medium is also provided. When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the object recommendation method in the embodiment of the present disclosure. The computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.

[0186] In an exemplary embodiment, a computer program product containing instructions is also provided. When the computer program product is run on a computer, the computer is caused to execute the object recommendation method in the embodiment of the present disclosure.

[0187] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, which can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0188] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the present disclosure being indicated by the following claims.

[0189] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. An object recommendation method, characterized in that: include: Acquire multiple objects and a third user set in the target application, wherein the multiple objects are objects whose publishing duration information is less than a preset duration threshold, and the registration duration of users in the third user set is greater than a first preset duration; determining, based on second interaction information between users in the third user set and the multiple objects, second business indicator information of each of the multiple objects under multiple business indicators; determining a plurality of candidate objects from the plurality of objects according to the second business indicator information; Obtain a first user set and a second user set; the registration duration of users in the first user set is less than or equal to a first preset duration and greater than a second preset duration, the registration duration of users in the second user set is less than or equal to the second preset duration, the first preset duration is greater than the second preset duration, and the registration duration is the duration from when the user was registered in the target application to the current time; Randomly recommending the multiple candidate objects to users in the first user set, and obtaining first interaction information between the users in the first user set and the multiple candidate objects; Determining, based on the first interaction information, first business indicator information of each of the multiple candidate objects under multiple business indicators; Determining a target object from the multiple candidate objects according to the first business indicator information; The target object is randomly recommended to users in the second user set.

2. The object recommendation method according to claim 1, wherein: The step of determining a target object from the plurality of candidate objects according to the first business indicator information includes: Determining first recommendation parameter information for each of the plurality of candidate objects based on the first business indicator information; Obtaining an object category of each of the plurality of candidate objects; Based on the object categories, determining a subset of candidate objects under each object category; Based on the first recommendation parameter information, the target objects are respectively screened out from the candidate object subsets under each object category.

3. The object recommendation method according to claim 1 or 2, characterized in that: The step of determining the plurality of candidate objects from the plurality of objects according to the second business indicator information includes: Determining second recommendation parameter information for each of the plurality of candidate objects according to the second business indicator information; Obtaining an object category of each of the plurality of objects; determining, based on the respective object categories of the plurality of objects, a subset of objects under each object category; Based on the second recommendation parameter information, the multiple candidate objects are screened out from the object subsets under each object category.

4. The object recommendation method according to claim 1, wherein: The second business indicator information represents a correlation between the number of times the multiple objects perform preset actions within a preset time period and the exposure of the multiple objects. The step of obtaining the multiple objects in the target application includes: Obtaining the exposure corresponding to each object in the target application; The objects whose exposure amounts are greater than the exposure amount threshold are regarded as the multiple objects.

5. The object recommendation method according to claim 1, wherein: The step of obtaining multiple objects in the target application includes: Obtaining release duration information of each object in the target application; The objects whose publishing duration information is less than a preset duration threshold are regarded as the multiple objects.

6. The object recommendation method according to claim 1, wherein: The step of randomly recommending the plurality of candidate objects to users in the first user set comprises: Obtaining objects to be recommended that match each user in the first user set; sorting the objects to be recommended and the multiple candidate objects matched by each user respectively to obtain a sorting result corresponding to each user in the first user set, wherein the multiple candidate objects in the sorting result are sorted at intervals within the objects to be recommended; Based on the ranking result, the objects to be recommended and the candidate objects that match each user are recommended to each user in the first user set.

7. An object recommendation device, characterized in that: include: An acquisition module is configured to acquire a plurality of candidate objects, a first user set, and a second user set in a target application; The registration duration of users in the first user set is less than or equal to a first preset duration and greater than a second preset duration, and the registration duration of users in the second user set is less than or equal to the second preset duration, the first preset duration is greater than the second preset duration, and the registration duration is the duration from when the user was registered in the target application to the current time; A random recommendation module is configured to randomly recommend the multiple candidate objects to users in the first user set, and obtain first interaction information between the users in the first user set and the multiple candidate objects; A first business indicator information determining module is configured to determine first business indicator information of each of the plurality of candidate objects under a plurality of business indicators based on the first interaction information; a target object determining module, configured to determine a target object from the plurality of candidate objects according to the first business indicator information; A recommendation module, configured to randomly recommend the target object to users in the second user set; The acquisition module includes: an acquisition unit configured to execute acquisition of multiple objects and a third user set in the target application, wherein the multiple objects are objects whose publishing duration information is less than a preset duration threshold, and the registration duration of users in the third user set is greater than a first preset duration; A second business indicator information determining unit is configured to determine second business indicator information of each of the multiple objects under multiple business indicators based on second interaction information between users in the third user set and the multiple objects; The candidate object determining unit is configured to determine a plurality of candidate objects from the plurality of objects according to the second business indicator information.

8. The object recommendation device according to claim 7, wherein: The target object determination module includes: a first recommendation parameter information determination unit configured to determine first recommendation parameter information of each of the plurality of candidate objects according to the first business indicator information; an object category acquiring unit, configured to acquire the object category of each of the plurality of candidate objects; a candidate object subset determining unit, configured to determine a candidate object subset under each object category based on the object category; The target object determining unit is configured to filter out the target object from the candidate object subsets under each object category based on the first recommendation parameter information.

9. The object recommendation device according to claim 7 or 8, characterized in that The candidate object determination unit includes: A second recommendation parameter information determining subunit is configured to determine second recommendation parameter information of each of the plurality of candidate objects according to the second business indicator information; an object category determination subunit, configured to obtain the object category of each of the plurality of objects; an object subset determining subunit, configured to determine an object subset under each object category based on the respective object categories of the plurality of objects; The candidate object determination subunit is configured to filter out the multiple candidate objects from the object subsets under each object category based on the second recommendation parameter information.

10. The object recommendation device according to claim 7, wherein: The second business indicator information represents a correlation between the number of times the plurality of objects perform preset actions within a preset time period and the exposure amounts of the plurality of objects, and the acquiring unit includes: an exposure acquisition subunit, configured to acquire the exposure corresponding to each object in the target application; The multiple object determination subunit is configured to determine the objects whose exposure amounts are greater than an exposure amount threshold as the multiple objects.

11. The object recommendation device according to claim 7, wherein: The acquisition unit includes: A duration information acquisition subunit is configured to acquire release duration information of each object in the target application in the target application; The multiple object acquisition subunits are configured to take the objects whose publishing duration information is less than a preset duration threshold as the multiple objects.

12. The object recommendation device according to claim 7, wherein: The random recommendation module includes: a to-be-recommended object acquiring unit, configured to execute acquiring to-be-recommended objects that match each user in the first user set; a sorting unit configured to sort the objects to be recommended and the multiple candidate objects matched by each user, respectively, to obtain a sorting result corresponding to each user in the first user set, wherein the multiple candidate objects are sorted at intervals within the objects to be recommended in the sorting result; The random recommendation unit is configured to recommend the objects to be recommended and the candidate objects that match each user to each user in the first user set based on the ranking result.

13. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the object recommendation method according to any one of claims 1 to 6.

14. A computer-readable storage medium, characterized in that When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the object recommendation method according to any one of claims 1 to 6.

15. A computer program product comprising computer instructions, characterized in that When the computer instructions are executed by a processor, the object recommendation method according to any one of claims 1 to 6 is implemented.

Images