Method, apparatus, medium, and computing device for determining media objects with potential for being a hit

Abstract

An embodiment of the present disclosure provides a method for determining a media object with hit potential, comprising: obtaining first behavior data of a target distribution user on a to-be-distributed media object in a first preset time period; in response to the first behavior data meeting a first preset condition, inserting the to-be-distributed media object into a list, and obtaining second behavior data of a user corresponding to the list on the to-be-distributed media object in a second preset time period; and in response to the second behavior data meeting a second preset condition, determining that the to-be-distributed media object is a media object with hit potential. The present disclosure only needs to obtain the behavior data of the user, has less demand for the play quantity or reading quantity of the media object, reduces the mining time of the media object with hit potential, improves the mining efficiency of the media object with hit potential, and quickly explores the potential ability of the media object. In addition, the present disclosure also provides a device for determining a media object with hit potential, a medium and a computing device.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the audio field, and more specifically, embodiments of this disclosure relate to methods, apparatus, media, and computing devices for determining media objects with blockbuster potential. Background Technology

[0002] This section is intended to provide background or context for embodiments of this disclosure. The description herein is not intended to imply that it is prior art simply because it is included in this section.

[0003] In the multimedia field, identifying media objects with the potential to become viral, such as e-books, audio, and video, is an important task in promoting these media objects.

[0004] In an exemplary technique, a vector of a media object is obtained by acquiring high-quality comment data of the media object. This vector is then compared with the media objects of historical viral media objects to calculate similarity, thereby determining whether the media object has the potential to become a viral media object based on the similarity.

[0005] However, high-quality comment data is obtained based on the media object having a large number of views or plays, and it takes a long time for a media object to have a large number of views or plays. In other words, it takes a long time to discover whether a media object has the potential to become a hit, resulting in low efficiency in discovering media objects with the potential to become hits. Summary of the Invention

[0006] This disclosure provides a method, apparatus, medium, and computing device for determining media objects with potential to become viral hits, in order to solve the problem of low efficiency in mining media objects with potential to become viral hits.

[0007] In a first aspect of the present disclosure, a method for determining a media object with the potential to become a hit is provided, comprising: distributing a media object to be distributed to a target distribution user, and obtaining first behavioral data of the target distribution user on the media object to be distributed in a first preset time period; in response to the first behavioral data satisfying a first preset condition, inserting the media object to be distributed into a list, and obtaining second behavioral data of the user corresponding to the list on the media object to be distributed in a second preset time period;

[0008] In response to the second behavioral data satisfying the second preset condition, it is determined that the media object to be distributed is a media object with the potential to become a hit.

[0009] In one embodiment of this disclosure, distributing the media object to be distributed to the target distribution user includes:

[0010] In response to the fact that the media object to be distributed is a cold-start media object, a first distribution user is determined through similar media objects of the media object to be distributed, and the media object to be distributed is distributed to the first distribution user;

[0011] The first red heart user among the first distribution users is identified, and users with similar behavior to the first red heart user are identified as second distribution users. The media object to be distributed is then distributed to the second distribution user. The first red heart user is the first distribution user who performs a preset operation on the media object to be distributed.

[0012] Users similar to the first and second "red heart" users are identified as third distribution users, and the media object to be distributed is distributed to the third distribution user. The target distribution user includes the first distribution user, the second distribution user, and the third distribution user. The second "red heart" user is the second distribution user who performs a preset operation on the media object to be distributed.

[0013] In another embodiment of this disclosure, determining the first distribution user through similar media objects of the media object to be distributed includes:

[0014] Determine the first similarity between the media object to be distributed and each target media object in the database;

[0015] Among the various first similarities, a target similarity greater than a first preset similarity is determined, and a first distribution user is determined based on the user of the target media object corresponding to the target similarity.

[0016] In another embodiment of this disclosure, determining the first similarity between the media object to be distributed and each target media object in the database includes:

[0017] Determine the first feature vector of the media object to be distributed;

[0018] Determine the inner product between the first feature vector and the second feature vector corresponding to each of the target media objects;

[0019] A first similarity between the media object to be distributed and the target media objects is determined based on each of the inner products.

[0020] In another embodiment of this disclosure, determining the first feature vector of the media object to be distributed includes:

[0021] Obtain the audio and text content of the media object to be distributed;

[0022] Determine the audio vector corresponding to the audio, and determine the text vector based on the text content;

[0023] The first feature vector of the media object to be distributed is determined based on the audio vector and the text vector.

[0024] In another embodiment of this disclosure, it further includes:

[0025] The vector index that matches the first feature vector is determined in the database, and the media object that matches the vector index is determined as the target media object.

[0026] In another embodiment of this disclosure, determining users who have similar behavior to the first "red heart" user includes:

[0027] Among the first red heart users and the third red heart users corresponding to the target media object, the same users are identified. The third red heart users are users who perform preset operations on the target media object.

[0028] Based on the number of identical users and the total number, a similarity score is determined between the target media object and the media object to be distributed, where the total number is the total number of the third red heart users;

[0029] The similarity score that is greater than the preset score is determined as the target similarity score;

[0030] The third red heart user of the target media object corresponding to the target similarity score is determined as a user who has similar behavior to the first red heart user.

[0031] In another embodiment of this disclosure, determining users similar to the first and second "red heart" users includes:

[0032] Acquire first interaction information between the target user, the target media object, and the creator of the target media object, wherein the target user includes the first red heart user and the second red heart user;

[0033] Obtain second interaction information between the user to be identified, the target media object, and the creator;

[0034] Based on the first interaction information and the second interaction information, a second similarity between the user to be determined and the target user is determined;

[0035] Users whose similarity score is greater than the second preset similarity score are identified as users similar to the target user.

[0036] In another embodiment of this disclosure, determining the second similarity between the user to be determined and the target user based on the first interaction information and the second interaction information includes:

[0037] The third feature vector of the target user is determined based on the first interaction information, and the fourth feature vector of the user to be determined is determined based on the second interaction information.

[0038] Based on the third feature vector and the fourth feature vector, a second similarity between the user to be determined and the target user is determined.

[0039] In another embodiment of this disclosure, before distributing the media object to be distributed to the target distribution user, the method further includes:

[0040] Distribute the media objects to be distributed to the users who wish to view them;

[0041] Obtain the third behavioral data of the appreciating user on the media object to be distributed during the third preset time period;

[0042] In response to the third set of preset conditions being met by the third set of data, the media object to be distributed is distributed to the target distribution user.

[0043] In a second aspect of this disclosure, an apparatus for determining a media object with blockbuster potential is also provided, comprising:

[0044] The sending module is used to distribute the media object to be distributed to the target distribution user and to obtain the first behavior data of the target distribution user on the media object to be distributed within a first preset time period.

[0045] An insertion module is used to insert the media object to be distributed into the list in response to the first behavior data meeting the first preset condition, and to obtain the second behavior data of the user corresponding to the list on the media object to be distributed during the second preset time period.

[0046] The determination module is used to determine, in response to the second behavior data satisfying the second preset condition, that the media object to be distributed is a media object with the potential to become a hit.

[0047] In one embodiment of this disclosure, it further includes:

[0048] The determining module is further configured to, in response to the fact that the media object to be distributed is a cold-start media object, determine a first distribution user through similar media objects of the media object to be distributed, and distribute the media object to be distributed to the first distribution user;

[0049] The determining module is further configured to determine the first red heart user among the first distribution users, and determine users with similar behavior to the first red heart user as second distribution users, and distribute the media object to be distributed to the second distribution user, wherein the first red heart user is the first distribution user who performs a preset operation on the media object to be distributed;

[0050] The determining module is further configured to determine users similar to the first and second heart users as third distribution users, and distribute the media object to be distributed to the third distribution user. The target distribution user includes the first distribution user, the second distribution user, and the third distribution user. The second heart user is the second distribution user who performs a preset operation on the media object to be distributed.

[0051] In another embodiment of this disclosure, it further includes:

[0052] The determining module is further configured to determine the first similarity between the media object to be distributed and each target media object in the database;

[0053] The determining module is further configured to determine a target similarity greater than a first preset similarity among each of the first similarities, and to determine a first distribution user based on the user of the target media object corresponding to the target similarity.

[0054] In another embodiment of this disclosure, it further includes:

[0055] The determining module is further configured to determine the first feature vector of the media object to be distributed;

[0056] The determining module is further configured to determine the inner product between the first feature vector and the second feature vector corresponding to each of the target media objects;

[0057] The determining module is further configured to determine a first similarity between the media object to be distributed and each of the target media objects based on each of the inner products.

[0058] In another embodiment of this disclosure, it further includes:

[0059] The first acquisition module is used to acquire the audio and text content of the media object to be distributed;

[0060] The determining module is further configured to determine the audio vector corresponding to the audio, and to determine the text vector based on the text content;

[0061] The determining module is further configured to determine a first feature vector of the media object to be distributed based on the audio vector and the text vector.

[0062] In another embodiment of this disclosure, it further includes:

[0063] The determining module is configured to determine a vector index that matches the first feature vector in the database, and determine the media object that matches the vector index as the target media object.

[0064] In another embodiment of this disclosure, it further includes:

[0065] The determining module is further configured to determine the same user among the first red heart user and the third red heart user corresponding to the target media object, wherein the third red heart user is the user who performs the preset operation on the target media object;

[0066] The determining module is further configured to determine a similarity score between the target media object and the media object to be distributed based on the number of identical users and the total number, wherein the total number is the total number of the third red heart users;

[0067] The determining module is further configured to determine the similarity score that is greater than a preset score as the target similarity score;

[0068] The determining module is further configured to determine the third red heart user of the target media object corresponding to the target similarity score as a user with similar behavior to the first red heart user.

[0069] In another embodiment of this disclosure, it further includes:

[0070] The second acquisition module is used to acquire first interaction information between the target user, the target media object, and the creator of the target media object, wherein the target user includes the first red heart user and the second red heart user;

[0071] The second acquisition module is further configured to acquire second interaction information between the user to be determined, the target media object, and the creator;

[0072] The determining module is further configured to determine a second similarity between the user to be determined and the target user based on the first interaction information and the second interaction information;

[0073] The determining module is further configured to determine the user to be determined corresponding to the second similarity that is greater than the second preset similarity as a user similar to the target user.

[0074] In another embodiment of this disclosure, it further includes:

[0075] The determining module is further configured to determine the third feature vector of the target user based on the first interaction information, and to determine the fourth feature vector of the user to be determined based on the second interaction information;

[0076] The determining module is further configured to determine a second similarity between the user to be determined and the target user based on the third feature vector and the fourth feature vector.

[0077] In another embodiment of this disclosure, it further includes:

[0078] The sending module is also used to distribute the media object to be distributed to the viewing users;

[0079] The third acquisition module is also used to acquire the third behavioral data of the appreciating user on the media object to be distributed during a third preset time period;

[0080] The sending module is further configured to distribute the media object to be distributed to the target distribution user in response to the third line data satisfying the third preset condition.

[0081] In a third aspect of this disclosure, a medium is also provided, comprising: computer execution instructions, which, when executed by a processor, are used to determine a media object with the potential to become a hit as described above.

[0082] In a third aspect of this disclosure, a computing device is also provided, comprising:

[0083] Memory and processor;

[0084] The memory stores computer-executed instructions;

[0085] The processor executes computer execution instructions stored in the memory, causing the processor to perform the method for determining media objects with blockbuster potential as described above.

[0086] In this embodiment, when the potential of a media object is initially determined by the behavioral data of users in the target distribution of the media object, the media object is inserted into the list. Then, the behavioral data of users in the list on the media object is used to determine whether the media object has the potential to become a hit. That is, the method for mining media objects with the potential to become hits provided by this disclosure only needs to obtain user behavioral data, and the requirements for the playback volume or reading volume of the media object are relatively small. This reduces the mining time of media objects with the potential to become hits, improves the mining efficiency of media objects with the potential to become hits, and quickly tests the potential of the media object. Attached Figure Description

[0087] The above and other objects, features, and advantages of this disclosure will become readily apparent from the following detailed description of exemplary embodiments, taken in conjunction with the accompanying drawings. Several embodiments of this disclosure are illustrated in the drawings by way of example and not limitation, in which:

[0088] Figure 1 The illustration shows an application scenario diagram of the method for determining the potential of media objects according to embodiments of the present disclosure;

[0089] Figure 2 A schematic flowchart according to an embodiment of the present disclosure is shown;

[0090] Figure 3 A schematic flowchart according to another embodiment of the present disclosure is shown;

[0091] Figure 4 A schematic flowchart according to yet another embodiment of the present disclosure is shown;

[0092] Figure 5 A schematic flowchart according to another embodiment of the present disclosure is shown;

[0093] Figure 6 A schematic flowchart according to another embodiment of the present disclosure is shown;

[0094] Figure 7 A schematic flowchart according to yet another embodiment of the present disclosure is shown;

[0095] Figure 8 A schematic diagram of a program product provided according to an embodiment of the present disclosure is shown.

[0096] Figure 9 A schematic diagram of the structure of a device for determining the potential of a media object to become a hit, according to an embodiment of the present disclosure, is shown.

[0097] Figure 10 A schematic diagram of the structure of a computing device provided according to an embodiment of the present disclosure is shown.

[0098] In the accompanying drawings, the same or corresponding reference numerals indicate the same or corresponding parts. Detailed Implementation

[0099] The principles and spirit of this disclosure will now be described with reference to several exemplary embodiments. It should be understood that these embodiments are given merely to enable those skilled in the art to better understand and implement this disclosure, and are not intended to limit the scope of this disclosure in any way. Rather, these embodiments are provided to make this disclosure more thorough and complete, and to fully convey the scope of this disclosure to those skilled in the art.

[0100] Those skilled in the art will recognize that embodiments of this disclosure can be implemented as a system, apparatus, device, method, or computer program product. Therefore, this disclosure can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0101] According to embodiments of this disclosure, a method, apparatus, medium, and computing device for determining media objects with blockbuster potential are proposed.

[0102] Furthermore, the number of any elements in the accompanying drawings is for illustrative purposes only and not for limitation, and any naming is for distinction only and has no limiting meaning.

[0103] In addition, the data involved in this disclosure may be data authorized by the user or fully authorized by all parties. The collection, dissemination and use of the data shall comply with the requirements of relevant national laws and regulations. The implementation methods / executives of this disclosure may be combined with each other.

[0104] The principles and spirit of this disclosure will be explained in detail below with reference to several representative embodiments. Invention Overview

[0106] In the multimedia field, identifying media objects such as ebooks, audio, and video with the potential to become viral is a crucial task in promoting media objects. In one exemplary technique, a vector of the media object is obtained by acquiring high-quality comment data. This vector is then compared with historical viral media objects to calculate similarity, thereby determining whether the media object has the potential to become a viral hit.

[0107] The inventors of this patent discovered that high-quality comment data is obtained based on the media object having a large number of views or plays. However, it takes a long time for a media object to have a large number of views or plays, which means that it takes a long time to discover whether a media object has the potential to become a hit, resulting in low efficiency in discovering media objects with the potential to become hits.

[0108] The inventor of this patent therefore conceived of a method that distributes media objects to users and then collects user behavior data on those media objects. Based on this behavior data, the inventor can determine whether a media object has the potential to become a viral hit. This method requires less attention to the number of plays or reads of the media object, reduces the time spent identifying media objects with viral potential, improves the efficiency of identifying media objects with viral potential, and quickly probes the potential capabilities of the media object.

[0109] Application Scenarios Overview

[0110] First refer to Figure 1 , Figure 1This is a schematic diagram illustrating an application scenario of the method for determining media objects with blockbuster potential according to an embodiment of this disclosure. The device 100 for determining media objects with blockbuster potential distributes media objects to the terminals 200 of various target users. The media objects to be distributed can be e-books, audio, or video. Distributing to the terminal 200 means sending the playback link or reading link of the media object to the terminal 200 of the target user. Target users generate behavioral data based on the media objects on their terminals 200. Behavioral data includes, for example, data such as collecting, liking, and commenting on the media objects. The device 100 for determining media objects with blockbuster potential collects first behavioral data of each target user on the media object through the terminal 200. If the media object is determined to have potential based on the first behavioral data, the device 100 inserts the media object into a ranking list. All users 300 can participate in ranking the media objects in the ranking list. Therefore, the device 100 for determining media objects with blockbuster potential obtains second behavioral data of the users on the ranking list on the media object. If the second behavioral data meets preset conditions, the media object can be determined to have blockbuster potential.

[0111] Exemplary methods

[0112] The following is combined with Figure 1 Application scenarios, refer to Figures 2-7 This describes a method for determining media objects with blockbuster potential according to exemplary embodiments of the present disclosure. It should be noted that the above application scenarios are shown only to facilitate understanding of the spirit and principles of the present disclosure, and the embodiments of the present disclosure are not limited in any way. Rather, the embodiments of the present disclosure can be applied to any applicable scenario.

[0113] For example, the device for determining media objects with blockbuster potential involved in the embodiments of this disclosure can be any device with data processing capabilities. For example, the device for determining media objects with blockbuster potential can be a computer, a server, etc.

[0114] Reference Figure 2 , Figure 2 An exemplary flowchart is shown for one embodiment of a method for determining media objects with blockbuster potential according to an embodiment of the present disclosure. The method for determining media objects with blockbuster potential includes:

[0115] Step S201: Distribute the media object to be distributed to the target distribution user and obtain the first behavior data of the target distribution user on the media object to be distributed within a first preset time period.

[0116] In this embodiment, the execution subject is the device for determining media objects with blockbuster potential. For ease of description, the device will be referred to as the device for determining media objects with blockbuster potential in the following description.

[0117] In the multimedia field, identifying media objects with the potential to become viral hits is a crucial task. Media objects with viral potential refer to those that can be loved by a broad audience. These media objects can include audio, e-books, videos, TV series, movies, and more.

[0118] Shortly after a media object is created, its potential can be analyzed to determine if it has the potential to become a viral hit. In this embodiment, the media object for potential analysis is defined as a media object to be distributed.

[0119] The device first identifies the target distribution users, and then distributes the media object to be distributed to each target distribution user. In other words, the device pushes the media object to be distributed to the terminals of each target distribution user so that the target distribution user can play or read the media object. For example, if the media object to be distributed is an e-book, the target distribution user will read the e-book; if the media object to be distributed is audio, video, a TV series, or a movie, the target distribution user will play the media object.

[0120] The device can determine whether a user can be identified as a target distribution user based on their user information. For example, if the number of media objects a user has played or read exceeds 1,000, then that user can be identified as a target distribution user. As another example, if a user has saved more than 100 media objects with potential to become viral, and this saving occurred before the media objects were identified as having viral potential, then that user has high discernment and can be identified as a target distribution user.

[0121] When a target user plays or reads a media object to be distributed, they generate first-action data for that media object. For example, when a target user plays audio to be distributed, they may perform actions such as liking, saving, downloading, and commenting on the audio; that is, the first-action data includes at least one of liking, saving, downloading, and commenting. Similarly, when a target user reads an ebook to be distributed, they may perform actions such as commenting to request updates, saving, voting, and tipping; in this case, the first-action data includes at least one of commenting, tipping, saving, and voting.

[0122] In order to quickly determine whether the media object to be distributed has the potential to become a hit, the device acquires the first behavior data of the target distribution user towards the media object within a first preset time period. That is, it collects the user's behavior data towards the media object within a time period, avoiding the need to spend a long time acquiring the first behavior data.

[0123] In step S202, in response to the first behavior data meeting the first preset condition, the media object to be distributed is inserted into the list, and the second behavior data of the user corresponding to the list for the media object to be distributed is obtained in the second preset time period.

[0124] The device collects the first behavioral data of each target distribution user. If the first behavioral data meets the first preset condition, it is preliminarily determined that the media object to be distributed is a media object with the potential to become a hit.

[0125] In one example, the device can determine whether a target distribution user likes the media object to be distributed based on first behavioral data. If the first behavioral data records at least one behavioral operation by the target distribution user in the media object to be distributed, such as liking, favorites, voting, tipping, positive comments, or downloading, then it can be determined that the target distribution user likes the media object to be distributed. The device determines the number of target distribution users who express liking for the media object to be distributed. If the ratio of this number to the total number of target distribution users is greater than a first preset ratio, then it can be determined that the first behavioral data satisfies a first preset condition.

[0126] In another example, the device can score the actions of target distribution users on media objects recorded in the first set of behavioral data. For example, actions such as liking, disliking, playback duration, reading duration, favorites, voting, tipping, and downloading each correspond to a score. The device determines the total score for each target distribution object, and then determines the number of target distribution users whose total score is greater than a preset score. If the ratio of this number to the total number of target distribution users is greater than a second preset ratio, then the first set of behavioral data can be determined to satisfy a first preset condition.

[0127] The target audience for distribution is a specific group of people. The fact that the media to be distributed is accepted by this target audience only indicates good conversion rates within that specific demographic; it doesn't mean the media will be popular with the general public. Therefore, behavioral data on the media's behavior towards it from all user groups is needed to verify its ability to expand beyond its initial target audience. In other words, it's necessary to verify whether the media is liked by all types of users.

[0128] The leaderboard is a centralized traffic source where all users can see the media objects within it, making it indiscriminate for each user. Therefore, the device inserts the media objects to be distributed into the leaderboard to obtain secondary behavioral data on how users within the leaderboard treat the distributed media objects.

[0129] To quickly determine whether a media object to be distributed has the potential to become a hit, the device acquires the second behavioral data of users corresponding to the ranking list regarding the media object to be distributed within a second preset time period. This means collecting behavioral data on users on the ranking list regarding the media object to be distributed within a specific time period, avoiding the need to spend excessive time acquiring this second behavioral data. The ranking list can be a leaderboard for the media object. Furthermore, the device can insert the media object to be distributed into any centralized scenario, which refers to a scenario where all types of users can play or read the media object to be distributed.

[0130] Step S203: In response to the second behavior data satisfying the second preset condition, determine that the media object to be distributed is a media object with the potential to become a hit.

[0131] In this embodiment, the list can be a trending list, which focuses on the growth rate of media objects. Therefore, the second row of data can be the change in the ranking of the media object to be distributed in the trending list within a second preset time period. If the change in ranking is positive, then the change in ranking is greater than a preset value, which means that the second row of data meets the second preset condition, that is, it is determined that the media object to be distributed is a media object with the potential to become a hit.

[0132] Furthermore, the device can determine the effective heart rate of the media object to be distributed by users on the leaderboard based on the second set of data. The effective heart rate is the number of hearts for the media object to be distributed divided by the number of valid plays for the media object. The number of hearts refers to the number of users on the leaderboard who have favorited the media object to be distributed, and valid plays refer to the duration of continuous playback of the media object to be distributed exceeding a preset duration. If the effective heart rate is greater than a preset threshold, it can be determined that the second set of data meets the second preset condition, that is, it can be determined that the media object to be distributed has the potential to become a hit.

[0133] In this embodiment, when the potential of a media object is initially determined by the behavioral data of users in the target distribution, the media object is inserted into the ranking list. Then, the behavioral data of users in the ranking list is used to determine whether the media object has the potential to become a hit. That is, the method for mining media objects with the potential to become hits provided in this disclosure only needs to obtain user behavioral data, and the requirements for the playback volume or reading volume of the media object are relatively small. This reduces the mining time of media objects with the potential to become hits, improves the mining efficiency of media objects with the potential to become hits, and quickly explores the potential capabilities of media objects.

[0134] Reference Figure 3 , Figure 3 An exemplary flowchart illustrates another embodiment of the method for determining media objects with blockbuster potential according to embodiments of this disclosure, based on... Figure 1 In the embodiment shown, step S201 includes:

[0135] Step S301: In response to the fact that the media object to be distributed is a cold-started media object, the first distribution user is determined by similar media objects of the media object to be distributed, and the media object to be distributed is distributed to the first distribution user.

[0136] In this embodiment, the media object to be distributed is a cold-start media object, which means that the media object to be distributed has a small user base, that is, only a small number of users have played or read the media object to be distributed.

[0137] Because the user base of the media object to be distributed is small, the number of users to whom the media object will be distributed is also small. Therefore, the device identifies similar media objects corresponding to the media object to be distributed. The users corresponding to these similar media objects are defined as the first distribution users. The device then distributes the media object to be distributed to these first distribution users, who are included in the target distribution list. Users corresponding to similar media objects can be those who play or read similar media objects, or those who have saved similar media objects to their favorites.

[0138] Similar media objects refer to media objects that are similar to the media object to be distributed. For example, media objects with the same musical style as the media object to be distributed are identified as similar media objects; musical styles could be upbeat or sad. Another example is media objects with the same type as the media object to be distributed; types could be rock, traditional Chinese style, classical, etc.

[0139] Step S302: Determine the first red heart user among the first distribution users, and determine users with similar behavior to the first red heart user as second distribution users, and distribute the media object to be distributed to the second distribution user. The first red heart user is the first distribution user who performs preset operations on the media object to be distributed.

[0140] The device can determine a second distribution user based on a first distribution user, and then send the media object to be distributed to the second distribution user. The target distribution user also includes the second distribution user. Specifically, the device determines the first "heart" user among the first distribution users. The first "heart" user is the first distribution user who performs a preset operation on the media object to be distributed. The preset operation refers to one or more of the following actions: liking, favorite, voting, and tipping.

[0141] Users who exhibit similar behavior to the first red heart user can be considered to be in the same category as the first red heart user. Therefore, the device can identify users who exhibit similar behavior to the first red heart user as second distribution users.

[0142] The device determines the common preferences of each first-heart user based on their preferences, and then identifies users with the same preferences as those who exhibit similar behavior to the first-heart users. For example, if the preference of each first-heart user is to play more than 100 rock songs, then users who play more than 100 rock songs are considered to have similar behavior to the first-heart users.

[0143] Step S303: Determine users similar to the first and second heart users as third distribution users, and distribute the media object to be distributed to the third distribution user. The target distribution users include the first distribution user, the second distribution user, and the third distribution user. The second heart user is the second distribution user who performs a preset operation on the media object to be distributed.

[0144] The device can determine a third distribution user based on the first and second heart users, and then distribute the media object to be distributed to the third distribution user. The target distribution user also includes the third distribution user.

[0145] The second "red heart" user is the second distribution user who performs preset operations on the media to be distributed. The method for determining the second "red heart" user is the same as that for determining the first "red heart" user, and will not be repeated here.

[0146] The device identifies users similar to the first "red heart" user and users similar to the second "red heart" user as third-party recipients. For example, based on the user information of each first "red heart" user, the device determines that each first "red heart" user is a student aged 18-23 who likes listening to music. The device then filters out students aged 18-23 who like listening to music from all users as similar users to the first "red heart" users. As another example, based on the user information of each second "red heart" user, the device determines that each second "red heart" user likes audio created by artist A and that each second "red heart" user's attention duration with artist A exceeds a preset duration. The device then filters out users from all users whose attention duration with artist A exceeds the preset duration as similar users to the second "red heart" users.

[0147] It should be noted that when the media object to be distributed is not a cold-start media object, meaning that the media object to be distributed has a large number of users playing it, the device does not need to determine the first distribution user through similar media. Instead, it can determine the first "heart user" by the users playing the media object to be distributed, then determine the second distribution user by users similar to the first heart user, and finally determine the third distribution user by combining the first and second heart users. Furthermore, the device does not need to determine the second heart user; it can directly determine the third distribution user by identifying users similar to the first heart user.

[0148] In this embodiment, the device expands the distribution users of the media object to be distributed by means of similar media objects, similar behaviors, and similar users, so that the device has enough data to accurately determine whether the media object to be distributed is a media object with the potential to become a hit.

[0149] Reference Figure 4 , Figure 4 An exemplary flowchart is shown for yet another embodiment of the method for determining media objects with blockbuster potential according to embodiments of the present disclosure, based on... Figure 3 In the embodiment shown, step S301 includes:

[0150] Step S401: Determine the first similarity between the media object to be distributed and each target media object in the database.

[0151] In this embodiment, the database stores multiple target media objects, and the type of the target media objects is the same as the type of the media objects to be distributed. For example, if the media object to be distributed is audio, then the target media object is also audio.

[0152] The device calculates the first similarity between the media object to be distributed and each target media object.

[0153] In one example, the media object to be distributed includes multiple tags, and the target media object also includes multiple tags. The ratio of the number of tags that the media object to be distributed shares with the target media object to the total number of tags can be used as the first similarity. The total number of tags can be the number of tags of the target media object. Tags could be, for example, singer A, composer B, lyricist C, rock, positive energy, etc.

[0154] In another example, the device determines a first feature vector of the media object to be distributed, and then determines the inner product between the first feature vector and the second feature vector of the target media object. The first similarity between the media object to be distributed and the target media object can be determined by the inner product. It should be noted that the device needs to normalize the first and second feature vectors, and then perform an inner product between the first and second feature vectors to obtain the first similarity. The value of the first similarity can range from 0 to 1. The device can use a similarity calculation model to calculate the first similarity. The samples used for training the similarity calculation model are the accumulated user-media object interaction data in the media object recommendation scenario. The media object recommendation scenario refers to the scenario of recommending media objects to users. Taking audio as an example, for audio A and audio B, if the heart rate and / or completion rate of recommending audio B based on audio A exceeds the average value, then audio A is considered similar to audio B. That is, the data corresponding to audio A and audio B can be used as positive samples for training the similarity calculation model, that is, as positive samples of similar audio for training the similarity calculation model. The "red heart rate" refers to the effective red heart rate, while the completion rate is the ratio of the number of audio tracks played completely to the total number of audio tracks. The negative samples used in the similarity calculation model training are randomly sampled, meaning a pair of audio tracks is randomly selected as dissimilar negative samples, while ensuring that the selected audio pairs (two audio tracks) are not positive samples. For positive samples, the score of the similarity calculation model training (the score being the similarity between the two audio tracks in the positive sample) is as close to 1 as possible; for negative samples, the score of the similarity calculation model training (the score being the similarity between the two audio tracks in the negative sample) is as close to 0 as possible.

[0155] The first feature vector can be determined from the audio and text content of the media object to be distributed. Specifically, the device acquires the audio and text content of the media object to be distributed, determines the text vector based on the text content, and determines the audio vector corresponding to the audio, that is, vectorizes the text content and audio to obtain audio vector and text vector, and then determines the first feature vector of the media object to be distributed based on the audio vector and text vector.

[0156] The device can use a model to determine the first feature vector. Specifically, for audio vectors, the open-source model YAMNet can be used to directly calculate the audio vector representation from the original audio signal of the media object to be distributed; that is, the audio vector of the media object to be distributed is directly obtained based on the open-source model. For the text content of the media object to be distributed, such as the lyrics of an audio file, the classic text processing algorithm word2vec can be used to calculate word vectors, and finally the text content sequence is converted into a vector sequence to obtain the text vector. After obtaining the audio vector and text vector, a deep learning model is used to calculate and learn the final vector representation of the media object, and then a deep learning model is used to transform the audio vector and text vector to obtain the first feature vector.

[0157] Furthermore, the device can utilize a deep learning model to calculate feature vectors for each media object in the database, thereby constructing a feature vector for each media object. Based on these feature vectors, a vector index is built for the database, essentially classifying the feature vectors and constructing a vector index corresponding to each class of feature vectors. After determining the first feature vector of the media object to be distributed, the device identifies the vector index in the database that matches the first feature image and determines the media object matching that vector index as the target media object. In other words, the media objects corresponding to each feature vector under the vector index are identified as target media objects. The device can calculate the similarity between the first feature vector and the vector index. If the similarity is greater than a preset threshold, the vector index is considered a match for the first feature vector. In this way, target media objects similar to the media object to be distributed can be initially filtered from a massive number of media objects.

[0158] Step S402: Determine the target similarity that is greater than the first preset similarity among the various first similarities, and determine the first distribution user based on the user of the target media object corresponding to the target similarity.

[0159] After determining the first similarity between the media object to be distributed and each target media object, the device determines a target similarity among the first similarities. The target similarity is a first similarity greater than a first preset similarity. The device determines the first distribution user based on the user of the target media object corresponding to the target similarity. For example, the user of the target media object corresponding to the target similarity can be a user who plays or reads the target media object, or a user who has saved the target media object.

[0160] In this embodiment, the device determines similar media objects to the media object to be distributed by calculating the similarity between the media object to be distributed and each target media object, thereby identifying the users of the similar media objects as the distribution users of the media object to be distributed, and accurately locating the distribution users of the media object to be distributed.

[0161] Reference Figure 5 , Figure 5 An exemplary flowchart is shown for another embodiment of the method for determining media objects with blockbuster potential according to embodiments of the present disclosure, based on... Figure 3 or Figure 4 In the embodiment shown, step S302 includes:

[0162] Step S501: Among the first red heart user and the third red heart user corresponding to the target media object, identify the same user. The third red heart user is the user who performs the preset operation on the target media object.

[0163] In this embodiment, the device expands the number of distribution users for the media object to be distributed by using the first red heart user.

[0164] Specifically, the device determines the third-heart user corresponding to the target media object. The target media object can be any media object, and the third-heart user is the user who performs a preset operation on the target media object. The preset operation is described above and will not be repeated here. Each first-heart user is considered as a group S1, and each third-heart user is considered as a group S2. S1 corresponds to the media object to be distributed, and S2 corresponds to the target media object. The device calculates the intersection of S1 and S2, that is, it identifies the same user among the first-heart users and the third-heart users.

[0165] Step S502: Based on the number of identical users and the total number, determine the similarity score between the target media object and the media object to be distributed. The total number is the total number of the third-red-heart users.

[0166] The device calculates a similarity score between the target media object and the media object to be distributed, based on the number of identical users and the total number of users. The total number is the total number of users with the third most popular heart.

[0167] For example, the device uses TF to represent the number of identical users and RU(S2) to represent the total number. Then, the proportion of identical users to the third heart users is TGI = TF / RU(S2), and the similarity score between the media object to be distributed and the target media object is TFTGI = TF*TGI.

[0168] Using the above method, the device calculates the similarity score between the media object to be distributed and each target media object.

[0169] Step S503: The similarity score that is greater than the preset score is determined as the target similarity score.

[0170] Step S504: The third red heart user of the target media object corresponding to the target similarity score is identified as a user with similar behavior to the first red heart user.

[0171] The device has a preset score, which can be any suitable value. The device determines the similarity score as the target similarity score if the similarity score is greater than the preset score. The device then considers each third-party user corresponding to the target media object with the target similarity score as a second distribution user, that is, these third-party users are users who have similar behaviors to the first-party users.

[0172] The method for identifying users with similar behavior to the first "red heart" user essentially involves calculating similar media objects to be distributed based on the user's interaction behavior with the media object. User interaction behavior with the media object can be "red heart" behavior, which refers to the act of adding a media object to one's favorites.

[0173] In this embodiment, based on the existing "red heart" group of the media object to be distributed, the "red heart" group of other media objects is calculated. By calculating the overlap and difference of the groups, similar media objects to be distributed are calculated, thereby identifying the "red heart" group of similar media objects as users with similar behaviors to the first "red heart" user.

[0174] Reference Figure 6 , Figure 6 An exemplary flowchart is shown in another embodiment of the method for determining media objects with blockbuster potential according to embodiments of the present disclosure, based on... Figures 3-5 In any of the embodiments shown, step S303 includes:

[0175] Step S601: Obtain the first interaction information between the target user, the target media object, and the creator of the target media object. The target user includes the first red heart user and the second red heart user.

[0176] In this embodiment, the number of distribution users for the media object to be distributed can be increased by using similar users. Specifically, after the media object to be distributed has accumulated a batch of "red heart" users (first red heart user and second red heart user), similar users of this batch of red heart users are identified.

[0177] To this end, the device acquires first interaction information between the target user, the target media object, and the creator of the target media object. The target user includes a first "like" user and a second "like" user. The target media object can be any media object. The first interaction information includes the interaction between the target user and the target media object, for example, the target user commenting on the target media object; the first interaction information also includes the interaction between the creator and the target user through the target media object, for example, the creator replying to the target user's comment in the comment section of the target media object.

[0178] Step S602: Obtain the second interaction information between the user to be identified, the target media object, and the creator.

[0179] The device further acquires second interaction information between the user to be identified, the target media object, and the creator. This second interaction information includes interactions between the user to be identified and the target media object, such as the user commenting on the target media object; it also includes interactions between the creator and the user to be identified through the target media object, such as the creator replying to a comment from the user to be identified in the target media object's comment section. The user to be identified refers to any user other than the target user.

[0180] Step S603: Determine the second similarity between the user to be determined and the target user based on the first interaction information and the second interaction information.

[0181] The device determines a second similarity between the user to be identified and the target user based on first and second interaction information. Specifically, the device determines a third feature vector of the target user based on the first interaction information and a fourth feature vector of the user to be identified based on the second interaction information, and then determines the second similarity between the user to be identified and the target user by comparing the third and fourth feature vectors.

[0182] The device can use a computational model to calculate the second similarity. This computational model employs a heterogeneous information network. The heterogeneous information network requires converting the first interaction information into a heterogeneous graph. The target user, target media object, and creator in the first interaction information are treated as nodes in the heterogeneous graph. If the target user interacts with the target media object, this interaction can be represented in the graph using edges between the target user node and the target media object. Similarly, the connection between the creator and the target media object can be represented by edges. Therefore, user-to-user paths can be determined from the graph. By training the graph network to perform random walks within the graph structure, the user's vector representation can be obtained. This means that the third feature vector can be obtained based on the first interaction information, and the fourth feature vector can be obtained through the second interaction information. The existence of a correlation between the target user and the user to be determined is determined by calculating the third and fourth feature vectors. A correlation refers to the target user and the user to be determined interacting with the same media object or creator.

[0183] Step S604: The user to be determined corresponding to the second similarity greater than the second preset similarity is determined as a user similar to the target user.

[0184] The device identifies the user to be determined whose second similarity score is greater than the second preset similarity score as a user similar to the target user, that is, identifies the user to be determined as the third distribution user.

[0185] In this embodiment, similar users are identified by using the existing "red heart" users of the media object to be distributed, thereby expanding the number of users to whom the media object is distributed, and thus obtaining enough data to accurately determine whether the media object to be distributed has the potential to become a hit.

[0186] Reference Figure 7 , Figure 7 An exemplary flowchart is shown for yet another embodiment of the method for determining media objects with blockbuster potential according to embodiments of the present disclosure, based on... Figures 3-6 In any of the embodiments shown, prior to step S201, the method further includes:

[0187] Step S701: Distribute the media object to be distributed to the viewing users.

[0188] In this embodiment, users have different preferences for media objects, but there is a group of forward-thinking users who discover media objects long before the general public discovers them. These users are defined as connoisseurs.

[0189] For a high-quality media object, the effective heart rate of the appreciating user is significantly higher than the average user level. To this end, the device can calculate the effective heart rate for each user based on their user information. The definition and calculation method of the effective heart rate are explained above and will not be repeated here.

[0190] If a user's effective heart rate is greater than the average effective heart rate, then that user is identified as a spectator. The average effective heart rate is the average of the effective heart rates of all users. The device then distributes the media object to be distributed to the terminal corresponding to each spectator.

[0191] Step S702: Obtain the third behavior data of the appreciating user towards the media to be distributed during the third preset time period.

[0192] In order to quickly determine whether the media object to be distributed has the potential to become a hit, the device acquires the third behavioral data of the viewing user towards the media object within a third preset time period. That is, it collects the behavioral data of the viewing user towards the media object within a time period, avoiding the need to spend a long time to acquire the third behavioral data.

[0193] In step S703, in response to the third line data satisfying the third preset condition, the media object to be distributed is distributed to the target distribution user.

[0194] The device collects third-behavior data from each viewing user. If the third-behavior data meets the third preset condition, the media object to be distributed is distributed to the target distribution user.

[0195] In one example, the device can determine whether a user likes a media object to be distributed based on third-line data. If the third-line data records at least one user action among liking, favorited, voting, tipping, positive commenting, and downloading the media object, then it can be determined that the user likes the media object. The device determines the number of users who express liking for the media object. If the ratio of this number to the total number of users is greater than a third preset ratio, then it can be determined that the third-line data satisfies a third preset condition.

[0196] In another example, the device can score the actions of appreciating users on media objects based on the data recorded in the third behavioral data. For example, actions such as liking, disliking, playback duration, reading duration, favorites, voting, tipping, and downloading each correspond to a score. The device determines the total score corresponding to each target distribution object, and then determines the number of appreciating users whose total score is greater than a preset score. If the ratio of this number to the total number of appreciating users is greater than a fourth preset ratio, then the third behavioral data can be determined to meet the third preset condition.

[0197] In this embodiment, the media object to be distributed is first inspected by a forward-looking user group. After the inspection is passed, the media object to be distributed is sent to the target distribution user.

[0198] Exemplary media

[0199] After introducing the methods of exemplary embodiments of this disclosure, the following references are made. Figure 8 The storage medium of the exemplary embodiments of this disclosure will be described.

[0200] refer to Figure 8 As shown, the storage medium 80 stores a program product for implementing the above-described method according to an embodiment of the present disclosure. This program product may be a portable compact disc read-only memory (CD-ROM) and includes computer-executable instructions for causing a computing device to execute the method for determining a media object with blockbuster potential provided in this disclosure. However, the program product of this disclosure is not limited thereto.

[0201] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0202] A readable signal medium may include data signals propagated in baseband or as part of a carrier wave, carrying computer-executed instructions. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium.

[0203] Computer-executable instructions for performing the operations disclosed herein can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The computer-executable instructions can be executed entirely on the user's computing device, partially on the user's computing device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing devices can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN).

[0204] Exemplary device

[0205] Having introduced the medium of exemplary embodiments of this disclosure, the following references are made to... Figure 9 The apparatus for determining media objects with blockbuster potential according to exemplary embodiments of the present disclosure will be described. The apparatus for determining media objects with blockbuster potential is used to implement the method in any of the above method embodiments, and its implementation principle and technical effect are similar.

[0206] refer to Figure 9 , Figure 9 A schematic diagram of the structure of a device for determining the potential of a media object to become a hit, according to an embodiment of the present disclosure, is shown.

[0207] like Figure 9 As shown, the device for determining media objects with blockbuster potential includes: a sending module 910, used to distribute the media object to be distributed to a target distribution user and obtain first behavioral data of the target distribution user on the media object to be distributed within a first preset time period; an insertion module 920, used to insert the media object to be distributed into a list in response to the first behavioral data meeting a first preset condition and obtain second behavioral data of the user on the list corresponding to the media object to be distributed in a second preset time period; and a determining module 930, used to determine that the media object to be distributed is a media object with blockbuster potential in response to the second behavioral data meeting a second preset condition.

[0208] In one embodiment of this disclosure, the method further includes: a determining module 930, further configured to, in response to the media object to be distributed being a cold-started media object, determine a first distribution user through similar media objects of the media object to be distributed, and distribute the media object to be distributed to the first distribution user; the determining module 930 is further configured to, determine a first "heart" user among the first distribution users, and determine a user with similar behavior to the first "heart" user as a second distribution user, and distribute the media object to be distributed to the second distribution user, wherein the first "heart" user is the first distribution user who performs a preset operation on the media object to be distributed; the determining module 930 is further configured to, determine a user similar to the first "heart" user and the second "heart" user as a third distribution user, and distribute the media object to be distributed to the third distribution user, wherein the target distribution user includes the first distribution user, the second distribution user, and the third distribution user, and the second "heart" user is the second distribution user who performs a preset operation on the media object to be distributed.

[0209] In another embodiment of this disclosure, it further includes: a determining module 930, which is further configured to determine a first similarity between the media object to be distributed and each target media object in the database; the determining module 930 is further configured to determine a target similarity greater than a first preset similarity among the first similarities, and determine a first distribution user based on the user of the target media object corresponding to the target similarity.

[0210] In another embodiment of this disclosure, it further includes: a determining module 930, which is further configured to determine a first feature vector of the media object to be distributed; the determining module 930 is further configured to determine the inner product between the first feature vector and the second feature vector corresponding to each target media object; and the determining module 930 is further configured to determine a first similarity between the media object to be distributed and each target media object based on the inner products.

[0211] In another embodiment of this disclosure, it further includes: a first acquisition module, configured to acquire the audio and text content of the media object to be distributed; a determination module 930, configured to determine the audio vector corresponding to the audio and determine the text vector according to the text content; and a determination module 930, configured to determine the first feature vector of the media object to be distributed according to the audio vector and the text vector.

[0212] In another embodiment of this disclosure, it further includes: a determination module 930, configured to determine a vector index that matches the first feature vector in a database, and determine the media object that matches the vector index as the target media object.

[0213] In another embodiment of this disclosure, the method further includes: a determining module 930, which is further configured to determine identical users among the first red-heart users and the third red-heart users corresponding to the target media object, wherein the third red-heart users are users who perform preset operations on the target media object; the determining module 930 is further configured to determine a similarity score between the target media object and the media object to be distributed based on the number of identical users and the total number, wherein the total number is the total number of third red-heart users; the determining module 930 is further configured to determine a similarity score greater than a preset score as a target similarity score; and the determining module 930 is further configured to determine the third red-heart users of the target media object corresponding to the target similarity score as users who have similar behaviors to the first red-heart users.

[0214] In another embodiment of this disclosure, the method further includes: a second acquisition module, configured to acquire first interaction information between a target user, a target media object, and the creator of the target media object, wherein the target user includes a first heart user and a second heart user; the second acquisition module is further configured to acquire second interaction information between the user to be determined, the target media object, and the creator; the determination module 930 is further configured to determine a second similarity between the user to be determined and the target user based on the first interaction information and the second interaction information; the determination module 930 is further configured to determine the user to be determined corresponding to the second similarity greater than the second preset similarity as a user similar to the target user.

[0215] In another embodiment of this disclosure, the method further includes: a determining module 930, which is further configured to determine a third feature vector of the target user based on the first interaction information, and to determine a fourth feature vector of the user to be determined based on the second interaction information; the determining module 930 is further configured to determine a second similarity between the user to be determined and the target user based on the third feature vector and the fourth feature vector.

[0216] In another embodiment of this disclosure, it further includes: a sending module 910, which is further configured to distribute the media object to be distributed to the viewing user; a third acquisition module, which is further configured to acquire third behavioral data of the media object to be distributed by the viewing user during a third preset time period; and the sending module 910, which is further configured to distribute the media object to be distributed to the target distribution user in response to the third behavioral data meeting a third preset condition.

[0217] Exemplary computing device

[0218] Having described the methods, media, and apparatus of exemplary embodiments of this disclosure, the following references... Figure 10 A computing device according to an exemplary embodiment of the present disclosure will be described.

[0219] Figure 10 The computing device 100 shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein. Figure 10 As shown, the computing device 100 is presented in the form of a general-purpose computing device. The components of the computing device 100 may include, but are not limited to: at least one processing unit 1001, at least one storage unit 1002, and a bus 1003 connecting different system components (including the processing unit 1001 and the storage unit 1002). The at least one storage unit 1002 stores computer-executable instructions; the at least one processing unit 1001 includes a processor that executes the computer-executable instructions to implement the methods described above.

[0220] Bus 1003 includes a data bus, a control bus, and an address bus.

[0221] Storage unit 1002 may include readable media in the form of volatile memory, such as random access memory (RAM) 10021 and / or cache memory 10022, and may further include readable media in the form of non-volatile memory, such as read-only memory (ROM) 10023.

[0222] Storage unit 1002 may also include a program / utility 10025 having a set (at least one) program module 10024, such program module 10024 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0223] The computing device 100 can also communicate with one or more external devices 1004 (e.g., keyboard, pointing device, etc.). This communication can be performed via the input / output (I / O) interface 1005. Furthermore, the computing device 100 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via a network adapter 1006. Figure 10 As shown, network adapter 1006 communicates with other modules of computing device 100 via bus 1003. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with computing device 100, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0224] It should be noted that although several units / modules or sub-units / modules of the device for determining media objects with blockbuster potential have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0225] Furthermore, although the operations of the methods disclosed herein are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all of the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0226] While the spirit and principles of this disclosure have been described with reference to several specific embodiments, it should be understood that this disclosure is not limited to the disclosed specific embodiments, and the division of aspects does not imply that features in these aspects cannot be combined for benefit; such division is merely for convenience of expression. This disclosure is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A method for determining media objects with blockbuster potential, characterized in that, include: Distribute the media object to be distributed to the target distribution user, and obtain the first behavior data of the target distribution user on the media object to be distributed in a first preset time period; In response to the first behavioral data satisfying the first preset condition, the media object to be distributed is inserted into the list, and the second behavioral data of the user corresponding to the list on the media object to be distributed is obtained in the second preset time period. In response to the second behavioral data satisfying the second preset condition, it is determined that the media object to be distributed is a media object with the potential to become a hit; The step of distributing the media object to be distributed to the target distribution user includes: In response to the fact that the media object to be distributed is a cold-start media object, a first distribution user is determined through similar media objects of the media object to be distributed, and the media object to be distributed is distributed to the first distribution user; The first red heart user among the first distribution users is identified, and users with similar behavior to the first red heart user are identified as second distribution users. The media object to be distributed is then distributed to the second distribution user. The first red heart user is the first distribution user who performs a preset operation on the media object to be distributed. Users similar to the first and second "red heart" users are identified as third distribution users, and the media object to be distributed is distributed to the third distribution user. The target distribution user includes the first distribution user, the second distribution user, and the third distribution user. The second "red heart" user is the second distribution user who performs a preset operation on the media object to be distributed.

2. The method for determining media objects with blockbuster potential according to claim 1, characterized in that, The step of determining the first distribution user through similar media objects of the media object to be distributed includes: Determine the first similarity between the media object to be distributed and each target media object in the database; Among the various first similarities, a target similarity greater than a first preset similarity is determined, and a first distribution user is determined based on the user of the target media object corresponding to the target similarity.

3. The method for determining media objects with blockbuster potential according to claim 2, characterized in that, Determining the first similarity between the media object to be distributed and each target media object in the database includes: Determine the first feature vector of the media object to be distributed; Determine the inner product between the first feature vector and the second feature vector corresponding to each of the target media objects; A first similarity between the media object to be distributed and the target media objects is determined based on each of the inner products.

4. The method for determining media objects with blockbuster potential according to claim 3, characterized in that, Determining the first feature vector of the media object to be distributed includes: Obtain the audio and text content of the media object to be distributed; Determine the audio vector corresponding to the audio, and determine the text vector based on the text content; The first feature vector of the media object to be distributed is determined based on the audio vector and the text vector.

5. The method for determining media objects with blockbuster potential according to claim 3, characterized in that, Also includes: The vector index that matches the first feature vector is determined in the database, and the media object that matches the vector index is determined as the target media object.

6. The method for determining media objects with blockbuster potential according to claim 1, characterized in that, The determination of users who have similar behavior to the first "red heart" user includes: Among the first red heart users and the third red heart users corresponding to the target media object, the same users are identified. The third red heart users are users who perform preset operations on the target media object. Based on the number of identical users and the total number, a similarity score is determined between the target media object and the media object to be distributed, where the total number is the total number of the third red heart users; The similarity score that is greater than the preset score is determined as the target similarity score; The third red heart user of the target media object corresponding to the target similarity score is determined as a user who has similar behavior to the first red heart user.

7. The method for determining media objects with blockbuster potential according to claim 1, characterized in that, The process of identifying users similar to the first and second "red heart" users includes: Acquire first interaction information between the target user, the target media object, and the creator of the target media object, wherein the target user includes the first red heart user and the second red heart user; Obtain second interaction information between the user to be identified, the target media object, and the creator; Based on the first interaction information and the second interaction information, a second similarity between the user to be determined and the target user is determined; Users whose similarity score is greater than the second preset similarity score are identified as users similar to the target user.

8. The method for determining media objects with blockbuster potential according to claim 7, characterized in that, Determining the second similarity between the user to be determined and the target user based on the first interaction information and the second interaction information includes: The third feature vector of the target user is determined based on the first interaction information, and the fourth feature vector of the user to be determined is determined based on the second interaction information. Based on the third feature vector and the fourth feature vector, a second similarity between the user to be determined and the target user is determined.

9. The method for determining media objects with blockbuster potential according to any one of claims 1-8, characterized in that, Before distributing the media object to be distributed to the target distribution user, the process also includes: Distribute the media objects to be distributed to the users who wish to view them; Obtain the third behavioral data of the appreciating user on the media object to be distributed during the third preset time period; In response to the third set of preset conditions being met by the third set of data, the media object to be distributed is distributed to the target distribution user.

10. A device for determining media objects with blockbuster potential, characterized in that, include: The sending module is used to distribute the media object to be distributed to the target distribution user and to obtain the first behavior data of the target distribution user on the media object to be distributed within a first preset time period. An insertion module is used to insert the media object to be distributed into the list in response to the first behavior data meeting the first preset condition, and to obtain the second behavior data of the user corresponding to the list on the media object to be distributed during the second preset time period. The determination module is used to determine, in response to the second behavior data satisfying the second preset condition, that the media object to be distributed is a media object with the potential to become a hit. The determining module is further configured to, in response to the fact that the media object to be distributed is a cold-start media object, determine a first distribution user through similar media objects of the media object to be distributed, and distribute the media object to be distributed to the first distribution user; The determining module is further configured to determine the first red heart user among the first distribution users, and determine users with similar behavior to the first red heart user as second distribution users, and distribute the media object to be distributed to the second distribution user, wherein the first red heart user is the first distribution user who performs a preset operation on the media object to be distributed; The determining module is further configured to determine users similar to the first and second heart users as third distribution users, and distribute the media object to be distributed to the third distribution user. The target distribution user includes the first distribution user, the second distribution user, and the third distribution user. The second heart user is the second distribution user who performs a preset operation on the media object to be distributed.

11. The device for determining media objects with blockbuster potential according to claim 10, characterized in that, Also includes: The determining module is further configured to determine the first similarity between the media object to be distributed and each target media object in the database; The determining module is further configured to determine a target similarity greater than a first preset similarity among each of the first similarities, and to determine a first distribution user based on the user of the target media object corresponding to the target similarity.

12. The device for determining media objects with blockbuster potential according to claim 11, characterized in that, Also includes: The determining module is further configured to determine the first feature vector of the media object to be distributed; The determining module is further configured to determine the inner product between the first feature vector and the second feature vector corresponding to each of the target media objects; The determining module is further configured to determine a first similarity between the media object to be distributed and each of the target media objects based on each of the inner products.

13. The device for determining media objects with blockbuster potential according to claim 12, characterized in that, Also includes: The first acquisition module is used to acquire the audio and text content of the media object to be distributed; The determining module is further configured to determine the audio vector corresponding to the audio, and to determine the text vector based on the text content; The determining module is further configured to determine a first feature vector of the media object to be distributed based on the audio vector and the text vector.

14. The device for determining media objects with blockbuster potential according to claim 12, characterized in that, Also includes: The determining module is configured to determine a vector index that matches the first feature vector in the database, and determine the media object that matches the vector index as the target media object.

15. The device for determining media objects with blockbuster potential according to claim 10, characterized in that, Also includes: The determining module is further configured to determine the same user among the first red heart user and the third red heart user corresponding to the target media object, wherein the third red heart user is the user who performs the preset operation on the target media object; The determining module is further configured to determine a similarity score between the target media object and the media object to be distributed based on the number of identical users and the total number, wherein the total number is the total number of the third red heart users; The determining module is further configured to determine the similarity score that is greater than a preset score as the target similarity score; The determining module is further configured to determine the third red heart user of the target media object corresponding to the target similarity score as a user with similar behavior to the first red heart user.

16. The device for determining media objects with blockbuster potential according to claim 10, characterized in that, Also includes: The second acquisition module is used to acquire first interaction information between the target user, the target media object, and the creator of the target media object, wherein the target user includes the first red heart user and the second red heart user; The second acquisition module is further configured to acquire second interaction information between the user to be determined, the target media object, and the creator; The determining module is further configured to determine a second similarity between the user to be determined and the target user based on the first interaction information and the second interaction information; The determining module is further configured to determine the user to be determined corresponding to the second similarity that is greater than the second preset similarity as a user similar to the target user.

17. The device for determining media objects with blockbuster potential according to claim 16, characterized in that, Also includes: The determining module is further configured to determine the third feature vector of the target user based on the first interaction information, and to determine the fourth feature vector of the user to be determined based on the second interaction information; The determining module is further configured to determine a second similarity between the user to be determined and the target user based on the third feature vector and the fourth feature vector.

18. The apparatus for determining the potential of a media subject to become a hit according to any one of claims 10-17, characterized in that, Also includes: The sending module is also used to distribute the media object to be distributed to the viewing users; The third acquisition module is also used to acquire the third behavioral data of the appreciating user on the media object to be distributed during a third preset time period; The sending module is further configured to distribute the media object to be distributed to the target distribution user in response to the third line data satisfying the third preset condition.

19. A computer-readable storage medium, characterized in that, include: Computer execution instructions, when executed by a processor, are used to implement the method for determining a media object with blockbuster potential as described in any one of claims 1 to 9.

20. A computing device, characterized in that, include: Memory and processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method for determining a media object with blockbuster potential as described in any one of claims 1 to 9.

Images