A method, device and equipment for identifying similarity of multimedia resource authors

Abstract

The present application discloses a method, device and equipment for identifying similarity of multimedia resource authors. The method comprises: firstly determining the effective audience of a target multimedia resource author within a preset time period, and forming a relationship pair between the target multimedia resource author and each of its effective audiences; then calculating the number of effective audiences overlapping between the target multimedia resource author and other authors according to the relationship pairs; then using the number of overlapping effective audiences and the total number of effective audiences, normalizing and de-extreme processing the audience overlap between the target multimedia resource author and other authors to obtain a normalized audience overlap value, and then splicing the normalized audience overlap values ​​between the target multimedia resource author and other authors into an overlap vector, and calculating the similarity between the overlap vectors corresponding to any two target multimedia resource authors, thereby improving the recognition efficiency and accuracy of the similarity.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device and equipment for identifying similarity of authors of multimedia resources. Background Art

[0002] With the rapid development of computer technology and mobile communication technology, various multimedia playback platforms based on electronic devices have been widely used, greatly enriching people's daily lives. For example, users can easily upload and watch various multimedia resources such as videos or audio through various short video playback platforms.

[0003] At present, Internet service providers will actively recommend multimedia resources such as short videos and music to users. Usually, these multimedia resources are provided by other authors and uploaded to the servers of Internet service providers. In order to provide users with multimedia resources that are more in line with their preferences, it is often necessary to refer to multimedia resources published by authoritative authors who are accepted or trusted by the user's group and have a greater influence on the group's attention behavior. However, the identification of these authoritative authors is often done through manual screening, based on indicators such as the labels and number of likes of the multimedia resource authors. On the one hand, the efficiency of such manual identification is relatively low, and on the other hand, it is difficult to guarantee the accuracy of the identification results. In addition, when a large number of authoritative authors with great influence and high similarity need to be identified at the same time, the cost of manual identification is also high. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to provide a method, device and equipment for identifying the similarity of multimedia resource authors, which can accurately identify the similarity between authors based on the situation of viewers watching the multimedia resources released by the authors, and then determine the authors with higher similarity.

[0005] The present application embodiment provides a method for identifying similarity of authors of multimedia resources, including:

[0006] Within a preset time period, determining the valid audience of the target multimedia resource author to be identified, and forming a relationship pair between the target multimedia resource author and each of its valid audiences;

[0007] According to the relationship pair, the number of valid viewers overlapping between the target multimedia resource author and other authors is calculated;

[0008] Using the number of overlapping effective viewers and the total number of effective viewers of the target multimedia resource author, normalize and de-extremely value the audience overlap between the target multimedia resource author and other authors to obtain a normalized audience overlap value between the target multimedia resource author and other authors;

[0009] The normalized audience overlap values ​​of the target multimedia resource author and other authors are concatenated into an overlap vector, and the similarity between any two overlap vectors corresponding to the target multimedia resource authors is calculated.

[0010] In a possible implementation, determining the effective audience of the author of the target multimedia resource to be identified within a preset time period includes:

[0011] In a preset time period, viewers who have effectively viewed, liked, commented on and shared the multimedia resources published by the target multimedia resource author are queried as effective viewers of the target multimedia resource author.

[0012] In a possible implementation, the effective viewing is viewing any multimedia resource published by the author of the target multimedia resource until a preset progress is reached.

[0013] In a possible implementation, after determining the valid audience of the target multimedia resource author to be identified within a preset time period and establishing a relationship pair between the target multimedia resource author and each of its valid audiences, the method further includes:

[0014] Deduplication is performed on the relationship pairs between the target multimedia resource author and each of its valid viewers obtained every day within a preset time period to obtain deduplicated relationship pairs;

[0015] The calculating, based on the relationship pair, the number of valid viewers overlapping between the target multimedia resource author and other authors includes:

[0016] The number of valid viewers overlapping between the target multimedia resource author and other authors is calculated based on the removed duplicate relationship pairs.

[0017] In a possible implementation, the method of using the number of overlapping effective viewers and the total number of effective viewers of the target multimedia resource author to normalize and de-extremely value the audience overlap between the target multimedia resource author and other authors to obtain a normalized audience overlap value between the target multimedia resource author and other authors includes:

[0018] The ratio of the number of overlapping effective viewers to the total number of effective viewers of the target multimedia resource author is calculated, and the ratio is de-extremely processed as the normalized audience overlap value corresponding to the target multimedia resource author and other authors.

[0019] In a possible implementation, the calculating the similarity between any two coincidence vectors corresponding to the authors of the target multimedia resources includes:

[0020] Calculate the cosine similarity between any two coincidence vectors corresponding to the authors of the target multimedia resources; or calculate the similarity between any two coincidence vectors corresponding to the authors of the target multimedia resources by using the Euclidean distance method.

[0021] In a possible implementation, the preset time period is 30 days before the current date.

[0022] The present application also provides a device for identifying similarities between authors of multimedia resources, the device comprising:

[0023] A determination unit, used to determine the valid audience of the target multimedia resource author to be identified within a preset time period, and form a relationship pair between the target multimedia resource author and each of its valid audiences;

[0024] A first calculation unit is used to calculate the number of valid viewers overlapping between the target multimedia resource author and other authors according to the relationship pair;

[0025] A processing unit, configured to perform normalization and de-extreme value processing on the audience overlap between the target multimedia resource author and other authors by using the number of overlapping effective viewers and the total number of effective viewers of the target multimedia resource author, so as to obtain a normalized audience overlap value between the target multimedia resource author and other authors;

[0026] The second calculation unit is used to splice the normalized audience overlap values ​​of the target multimedia resource author and other authors into a overlap vector, and calculate the similarity between any two overlap vectors corresponding to the target multimedia resource authors.

[0027] In a possible implementation manner, the determining unit is specifically configured to:

[0028] In a preset time period, viewers who have effectively viewed, liked, commented on and shared the multimedia resources published by the target multimedia resource author are queried as effective viewers of the target multimedia resource author.

[0029] In a possible implementation, the effective viewing is viewing any multimedia resource published by the author of the target multimedia resource until a preset progress is reached.

[0030] In a possible implementation manner, the device further includes:

[0031] A deduplication unit, configured to perform deduplication processing on the relationship pairs between the target multimedia resource author and each of its valid viewers obtained every day within a preset time period, to obtain deduplicated relationship pairs;

[0032] The first computing unit is specifically configured to:

[0033] The number of valid viewers overlapping between the target multimedia resource author and other authors is calculated based on the removed relationship pairs.

[0034] In a possible implementation, the processing unit is specifically configured to:

[0035] The ratio of the number of overlapping effective viewers to the total number of effective viewers of the target multimedia resource author is calculated, and the ratio is de-extremely processed as the normalized audience overlap value corresponding to the target multimedia resource author and other authors.

[0036] In a possible implementation manner, the second computing unit is specifically configured to:

[0037] Calculate the cosine similarity between any two coincidence vectors corresponding to the authors of the target multimedia resources; or calculate the similarity between any two coincidence vectors corresponding to the authors of the target multimedia resources by using the Euclidean distance method.

[0038] In a possible implementation, the preset time period is 30 days before the current date.

[0039] The embodiment of the present application also provides a device for identifying the similarity of multimedia resource authors, including: a processor, a memory, and a system bus;

[0040] The processor and the memory are connected via the system bus;

[0041] The memory is used to store one or more programs, and the one or more programs include instructions. When the instructions are executed by the processor, the processor executes any one of the implementations of the above-mentioned method for identifying the similarity of authors of multimedia resources.

[0042] The embodiment of the present application further provides a computer-readable storage medium, in which instructions are stored. When the instructions are executed on a terminal device, the terminal device executes any one of the implementations of the above-mentioned method for identifying the similarity of authors of multimedia resources.

[0043] The embodiment of the present application provides a method, device and equipment for identifying the similarity of multimedia resource authors. First, within a preset time period, the effective audience of the target multimedia resource author to be identified is determined, and the target multimedia resource author and each of its effective audience form a relationship pair, and then the number of effective audiences overlapping between the target multimedia resource author and other authors is calculated based on the relationship pair; then, the audience overlap between the target multimedia resource author and other authors is normalized and de-extreme processed using the number of overlapping effective audiences and the total number of effective audiences of the target multimedia resource author, so as to obtain the normalized audience overlap value between the target multimedia resource author and other authors, and then the normalized audience overlap value between the target multimedia resource author and other authors can be spliced ​​into a overlap vector, and the similarity between the overlap vectors corresponding to any two target multimedia resource authors is calculated. It can be seen that since the embodiment of the present application identifies the similarity between authors based on the situation of the audience watching the multimedia resources published by the author, so as to determine the author with higher similarity, the recognition efficiency and accuracy are greatly improved compared with the existing manual screening method. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 A flowchart of a method for identifying similarities of multimedia resource authors provided in an embodiment of the present application;

[0046] Figure 2 A schematic diagram of the composition of a device for identifying the similarity of authors of multimedia resources provided in an embodiment of the present application. DETAILED DESCRIPTION

[0047] At present, when identifying the similarity of multimedia resource authors, it is usually done through manual screening, based on the multimedia resource author's tags and likes, etc. However, this manual identification method relies heavily on the author's tags, which are automatically annotated by machines. On the one hand, there is a problem of accuracy, and on the other hand, the granularity of the tags that rely on manual identification is not objective and comprehensive enough. In addition, it is difficult to effectively compare authors with different numbers of followers (i.e. fans).

[0048] In order to solve the above defects, the present application provides a method for identifying the similarity of multimedia resource authors. First, within a preset time period, the effective audience of the target multimedia resource author to be identified is determined, and the target multimedia resource author and each of its effective audience form a relationship pair, and then according to the relationship pair, the number of effective audiences overlapping between the target multimedia resource author and other authors is calculated; then, using the number of overlapping effective audiences and the total number of effective audiences of the target multimedia resource author, the audience overlap between the target multimedia resource author and other authors is normalized and de-extremely processed to obtain the normalized audience overlap value between the target multimedia resource author and other authors, and then the normalized audience overlap value between the target multimedia resource author and other authors can be spliced ​​into a overlap vector, and the similarity between the overlap vectors corresponding to any two target multimedia resource authors is calculated. It can be seen that since the embodiment of the present application identifies the similarity between authors based on the situation of the audience watching the multimedia resources published by the author, so as to determine the author with higher similarity, compared with the existing manual screening method, the recognition efficiency and accuracy are greatly improved.

[0049] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0050] First embodiment

[0051] See also Figure 1 , is a flow chart of a method for identifying similarity of multimedia resource authors provided in this embodiment, the method comprising the following steps:

[0052] S101: within a preset time period, determining the valid audience of the target multimedia resource author to be identified, and forming a relationship pair between the target multimedia resource author and each of its valid audiences.

[0053] In this embodiment, any multimedia resource author for which similarity recognition is performed using this embodiment is defined as the target multimedia resource author. Furthermore, it should be noted that this embodiment does not limit the type of the target multimedia resource. For example, the target multimedia resource may be a short video or music released by its author on a short video playback platform. Furthermore, this embodiment does not limit the method for obtaining the target multimedia resource. The target multimedia resource may be obtained by downloading or the like according to actual needs.

[0054] In order to improve the recognition efficiency and accuracy of the similarity of multimedia resource authors, firstly, all valid audiences of each target multimedia resource author to be identified can be determined within a preset time period, and a relationship pair can be formed between the target multimedia resource author and each of its valid audiences to execute subsequent steps S102-S104, so as to realize accurate recognition of the similarity of the target multimedia resource author.

[0055] Specifically, an optional implementation method is that in order to more accurately judge the real-time influence of each multimedia resource author in his or her field, the present application proposes to query the audience who have effectively viewed, liked, commented on and shared the multimedia resources released by the target multimedia resource author within a preset time period (which can be set according to actual conditions, such as 30 days before the current date, etc.), as the effective audience of the target multimedia resource author.

[0056] Among them, effective viewing can refer to watching any multimedia resource published by the target multimedia resource author to reach a preset progress (which can be set according to actual conditions). For example, effective viewing can refer to watching a short video published by the target multimedia resource author to reach 70% of the short video's length. It can be seen that compared with the traditional concept of fans, effective viewers have a stronger real-time interaction with authors. Using effective viewers for identification also reduces the impact of "zombie fans", which is conducive to identifying multimedia resource authors in a rapid growth period, because although these authors may not have enough fans, some multimedia resources (such as short videos) are very popular and need to be identified in a timely manner.

[0057] In this implementation, within a preset time period (such as 30 days before the current date), all viewers who have had at least one valid behavior of viewing, liking, commenting or sharing any multimedia resource (such as a short video) published by each target multimedia resource author can be counted every day, and these viewers are regarded as the valid audience of the corresponding target multimedia resource author.

[0058] On this basis, for the target multimedia resource author, a one-to-one correspondence can be established between the author and each of its valid viewers, forming an "author-viewer" relationship pair for executing the subsequent step S102.

[0059] It should be noted that in actual applications, within a preset time period (such as 30 days before the current date), repeated "author-audience" relationship pairs may be counted on different days. For example, when counting yesterday, it was determined that audience A liked a short video released by author A, so the "author A-audience A" relationship pair was constructed, and when counting the day before yesterday, it was also determined that audience A commented on another short video released by author A, and the "author A-audience A" relationship pair was also constructed, which resulted in repeated relationship pairs "author A-audience A" in the statistical results of yesterday and the day before yesterday. Therefore, in order to facilitate subsequent data processing, it is necessary to deduplicate the relationship pairs formed by the target multimedia resource author and each of its valid audiences obtained every day within the preset time period to obtain the deduplicated "author-audience" relationship pairs for executing the subsequent step S102.

[0060] S102: Calculate the number of valid viewers overlapping between the target multimedia resource author and other authors based on the relationship pairs.

[0061] In this embodiment, after obtaining the "author-audience" relationship pairs through step S101, in order to improve the recognition efficiency and accuracy of the similarity of multimedia resource authors, it is further necessary to calculate the number of valid viewers overlapping between the target multimedia resource author and other authors based on the obtained "author-audience" relationship pairs, so as to execute the subsequent step S103.

[0062] For example, assuming that audience A is both a valid audience of target multimedia resource author A and a valid audience of target multimedia resource author B, and constitutes two relationship pairs of "author A-audience A" and "author B-audience A", then it can be determined that audience A is an overlapping valid audience between target multimedia resource authors A and B.

[0063] S103: using the number of overlapping effective viewers and the total number of effective viewers of the target multimedia resource author, normalize and de-extremely value the audience overlap between the target multimedia resource author and other authors to obtain a normalized audience overlap value between the target multimedia resource author and other authors.

[0064] It should be noted that when identifying the author of multimedia resources, in order to reduce the impact of the number of audiences, for example, in order to ensure that authors with hundreds of thousands of fans can be compared relatively fairly with authors with hundreds of millions of fans, after calculating the number of effective audiences that overlap between the target multimedia resource author and other authors through step S102, it is further necessary to use the number of effective audiences that overlap between the target multimedia resource author and other authors and the total number of effective audiences of the target multimedia resource author itself to normalize and de-extremely value the audience overlap between the target multimedia resource author and other authors (that is, delete some particularly large values ​​to ensure the balance of the data) to obtain the normalized audience overlap value between the target multimedia resource author and other authors for executing subsequent steps, such as executing subsequent step S104.

[0065] Specifically, an optional implementation method is that the implementation process in this step S103 may specifically include: first, calculating the ratio of the number of overlapping effective viewers to the total number of effective viewers of the target multimedia resource author, and then de-extreme processing the ratio as the normalized audience overlap value corresponding to the target multimedia resource author and other authors.

[0066] For example: Assuming that the number of valid viewers overlapped by author A and other authors is 2 million, and the total number of valid viewers corresponding to it is 10 million, then the normalized audience overlap value of author A and other authors is 0.2, that is, 200 / 1000 = 0.2. However, if it is calculated that the number of valid viewers overlapped by author M and other authors is 100,000, and the total number of valid viewers corresponding to it is also 100,000, then the normalized audience overlap value of author M and other authors is 1, that is, 10 / 10 = 1, then this maximum value can be deleted to ensure the balance of the data, that is, author M is no longer used as the target multimedia resource author for similarity recognition.

[0067] In this way, the influence of audience magnitude differences on the recognition results can be greatly reduced through normalization and de-extreme processing, thereby making the recognition results more accurate.

[0068] S104: concatenate the normalized audience overlap values ​​of the target multimedia resource author and other authors into a overlap vector, and calculate the similarity between the overlap vectors corresponding to any two target multimedia resource authors.

[0069] In this embodiment, after calculating the normalized audience overlap values ​​corresponding to all target multimedia resource authors and other authors through step S103, the normalized audience overlap values ​​of the target multimedia resource author and other authors can be further spliced ​​into an overlap vector, and the similarity between the overlap vectors corresponding to any two target multimedia resource authors can be calculated to improve the recognition efficiency and accuracy of the similarity.

[0070] Specifically, an optional implementation method is to concatenate the normalized audience overlap values ​​of the target multimedia resource author and other authors into a overlap vector, and then calculate the cosine similarity between the overlap vectors corresponding to any two target multimedia resource authors to characterize the similarity between the two authors. The specific calculation formula is as follows:

[0071]

[0072] Among them, A and B represent the coincidence vectors corresponding to the target multimedia resource author A and the target multimedia resource author B respectively; A i and B i Represent the components of vectors A and B respectively; similarity represents the cosine similarity between the target multimedia resource author A and the target multimedia resource author B.

[0073] Alternatively, after concatenating the normalized audience overlap values ​​of the target multimedia resource author and other authors into an overlap vector, the Euclidean distance method can be used to calculate the similarity between the overlap vectors corresponding to any two target multimedia resource authors. The specific implementation process will not be repeated here.

[0074] In this way, by using the above steps S101-S104, the similarity between authors can be identified based on the situation of the audience watching the multimedia resources released by the author, so as to determine the authors with higher similarity, thereby improving the recognition efficiency and accuracy of the similarity of the authors of multimedia resources. For example, when a group of viewers like author A, they are likely to like author B of the same type. As long as some authors are manually screened and recommended to users in the initial stage, after receiving feedback from users, several authors with better feedback effects can be selected from them, and then the similarity between each author can be calculated by executing the above steps S101-S104 as the similarity recognition result. In this way, the recognition result can be used to determine the authors who meet the conditions for subsequent product promotion in related fields to obtain better promotion effects.

[0075] In summary, the present embodiment provides a method for identifying the similarity of multimedia resource authors. First, within a preset time period, the effective audience of the target multimedia resource author to be identified is determined, and a relationship pair is formed between the target multimedia resource author and each of its effective audiences. Then, based on the relationship pair, the number of effective audiences that overlap between the target multimedia resource author and other authors is calculated; then, the audience overlap between the target multimedia resource author and other authors is normalized and de-extreme processed using the number of overlapping effective audiences and the total number of effective audiences of the target multimedia resource author, so as to obtain the normalized audience overlap value between the target multimedia resource author and other authors, and then the normalized audience overlap value between the target multimedia resource author and other authors can be spliced ​​into a overlap vector, and the similarity between the overlap vectors corresponding to any two target multimedia resource authors is calculated. It can be seen that since the present application embodiment identifies the similarity between authors based on the situation of the audience watching the multimedia resources published by the author, so as to determine the author with higher similarity, the recognition efficiency and accuracy are greatly improved compared with the existing manual screening method.

[0076] Second embodiment

[0077] This embodiment will introduce a device for identifying the similarity of authors of multimedia resources. For related content, please refer to the above method embodiment.

[0078] See also Figure 2 , is a schematic diagram of the composition of a device for identifying the similarity of multimedia resource authors provided in this embodiment, the device 200 includes:

[0079] A determination unit 201 is used to determine the valid audience of the target multimedia resource author to be identified within a preset time period, and form a relationship pair between the target multimedia resource author and each of its valid audiences;

[0080] A first calculation unit 202 is used to calculate the number of valid viewers overlapping between the target multimedia resource author and other authors according to the relationship pair;

[0081] The processing unit 203 is used to perform normalization and de-extreme value processing on the audience overlap between the target multimedia resource author and other authors by using the number of overlapping effective viewers and the total number of effective viewers of the target multimedia resource author, so as to obtain a normalized audience overlap value between the target multimedia resource author and other authors;

[0082] The second calculation unit 204 is used to concatenate the normalized audience overlap values ​​of the target multimedia resource author and other authors into a overlap vector, and calculate the similarity between any two overlap vectors corresponding to the target multimedia resource authors.

[0083] In an implementation of this embodiment, the determining unit 201 is specifically configured to:

[0084] In a preset time period, viewers who have effectively viewed, liked, commented on and shared the multimedia resources published by the target multimedia resource author are queried as effective viewers of the target multimedia resource author.

[0085] In an implementation of this embodiment, the effective viewing is viewing any multimedia resource published by the author of the target multimedia resource to reach a preset progress.

[0086] In an implementation of this embodiment, the device further includes:

[0087] A deduplication unit, configured to perform deduplication processing on the relationship pairs between the target multimedia resource author and each of its valid viewers obtained every day within a preset time period, to obtain deduplicated relationship pairs;

[0088] The first calculation unit 202 is specifically used for:

[0089] The number of valid viewers overlapping between the target multimedia resource author and other authors is calculated based on the removed duplicate relationship pairs.

[0090] In an implementation of this embodiment, the processing unit 203 is specifically configured to:

[0091] The ratio of the number of overlapping effective viewers to the total number of effective viewers of the target multimedia resource author is calculated, and the ratio is de-extremely processed as the normalized audience overlap value corresponding to the target multimedia resource author and other authors.

[0092] In an implementation of this embodiment, the second calculating unit 204 is specifically configured to:

[0093] Calculate the cosine similarity between any two coincidence vectors corresponding to the authors of the target multimedia resources; or calculate the similarity between any two coincidence vectors corresponding to the authors of the target multimedia resources by using the Euclidean distance method.

[0094] In one implementation of this embodiment, the preset time period is 30 days before the current date.

[0095] Furthermore, the embodiment of the present application also provides a device for identifying the similarity of multimedia resource authors, including: a processor, a memory, and a system bus;

[0096] The processor and the memory are connected via the system bus;

[0097] The memory is used to store one or more programs, and the one or more programs include instructions. When the instructions are executed by the processor, the processor executes any one of the implementation methods of the above-mentioned method for identifying the similarity of authors of multimedia resources.

[0098] Furthermore, an embodiment of the present application also provides a computer-readable storage medium, in which instructions are stored. When the instructions are executed on a terminal device, the terminal device executes any one of the implementation methods of the above-mentioned method for identifying the similarity of authors of multimedia resources.

[0099] It can be known from the description of the above implementation mode that those skilled in the art can clearly understand that all or part of the steps in the above-mentioned embodiment method can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., including several instructions for enabling a computer device (which can be a personal computer, a server, or a network communication device such as a media gateway, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0100] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0101] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0102] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying the similarity of multimedia resource authors, characterized in that: The method comprises: Within a preset time period, determine the valid audience of the target multimedia resource author to be identified, and form a relationship pair between the target multimedia resource author and each of its valid audiences, wherein the valid audience is an audience that has effectively viewed, liked, commented on, and shared the multimedia resource published by the target multimedia resource author; According to the relationship pair, the number of valid viewers overlapping between the target multimedia resource author and other authors is calculated; Deduplication is performed on the relationship pairs between the target multimedia resource author and each of its valid viewers obtained every day within a preset time period to obtain deduplicated relationship pairs; Using the number of overlapping effective viewers and the total number of effective viewers of the target multimedia resource author, normalize and de-extremely value the audience overlap between the target multimedia resource author and other authors to obtain a normalized audience overlap value between the target multimedia resource author and other authors; The normalized audience overlap values ​​of the target multimedia resource author and other authors are concatenated into an overlap vector, and the similarity between any two overlap vectors corresponding to the target multimedia resource authors is calculated.

2. The method according to claim 1, characterized in that Determining the effective audience of the target multimedia resource author to be identified within the preset time period includes: In a preset time period, viewers who have effectively viewed, liked, commented on and shared the multimedia resources published by the target multimedia resource author are queried as effective viewers of the target multimedia resource author.

3. The method according to claim 2, characterized in that The effective viewing is viewing any multimedia resource published by the author of the target multimedia resource to reach a preset progress.

4. The method according to claim 1, characterized in that The calculating, based on the relationship pair, the number of valid viewers overlapping between the target multimedia resource author and other authors includes: The number of valid viewers overlapping between the target multimedia resource author and other authors is calculated based on the removed relationship pairs.

5. The method according to claim 1, characterized in that The method of using the number of overlapping effective viewers and the total number of effective viewers of the target multimedia resource author to normalize and de-extremely value the audience overlap between the target multimedia resource author and other authors to obtain a normalized audience overlap value between the target multimedia resource author and other authors includes: The ratio of the number of overlapping effective viewers to the total number of effective viewers of the target multimedia resource author is calculated, and the ratio is de-extremely processed as the normalized audience overlap value corresponding to the target multimedia resource author and other authors.

6. The method according to claim 1, characterized in that The calculating the similarity between the coincidence vectors corresponding to any two authors of the target multimedia resources includes: Calculate the cosine similarity between any two coincidence vectors corresponding to the authors of the target multimedia resources; or calculate the similarity between any two coincidence vectors corresponding to the authors of the target multimedia resources by using the Euclidean distance method.

7. The method according to any one of claims 1 to 6, characterized in that: The preset time period is 30 days before the current date.

8. A device for identifying similarity of multimedia resource authors, characterized in that: The device comprises: A determination unit, configured to determine valid audiences of a target multimedia resource author to be identified within a preset time period, and form a relationship pair between the target multimedia resource author and each of its valid audiences, wherein the valid audiences are audiences who have effectively viewed, liked, commented on, and shared the multimedia resources published by the target multimedia resource author; A first calculation unit is used to calculate the number of valid viewers overlapping between the target multimedia resource author and other authors according to the relationship pair; A processing unit is used to perform normalization and de-extreme value processing on the audience overlap between the target multimedia resource author and other authors by using the number of overlapping effective viewers and the total number of effective viewers of the target multimedia resource author, so as to obtain the normalized audience overlap value between the target multimedia resource author and other authors; and to perform de-duplication processing on the relationship pairs between the target multimedia resource author and each of its effective viewers obtained every day within a preset time period, so as to obtain de-duplicated relationship pairs; The second calculation unit is used to splice the normalized audience overlap values ​​of the target multimedia resource author and other authors into a overlap vector, and calculate the similarity between any two overlap vectors corresponding to the target multimedia resource authors.

9. A device for identifying similarity of multimedia resource authors, characterized in that: include: Processor, memory, system bus; The processor and the memory are connected via the system bus; The memory is used to store one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by the processor, the processor executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the method according to any one of claims 1 to 7.

Images