Video Similarity Determination Method, Apparatus, Electronic Device, and Storage Medium

Abstract

An embodiment of the present invention provides a method, apparatus, electronic device, and storage medium for determining video similarity. The method includes: obtaining a video set, and for any two videos in the video set, determining the user overlap degree between the two videos and determining the first user overlap degree between the two videos; determining the second user overlap degree between the two videos that eliminates the influence of the user audience scale according to the user overlap degree and the first user overlap degree; determining the influence coefficient of the time interval between the two videos going online, and processing the second user overlap degree by using the influence coefficient to obtain the third user overlap degree between the two videos; determining that the third user overlap degree is the video similarity between the two videos. In this way, the third user overlap degree eliminates the influence of the user audience scale and the time interval between videos going online, which can ensure that the clustering result of subsequent video similarity clustering is reasonable and reliable.

Description

Technical Field

[0001] The present invention relates to the technical field of video processing, and in particular, to a method, apparatus, electronic device, and storage medium for determining video similarity. Background Art

[0002] Currently, videos such as TV dramas, movies, and variety shows on major video websites are diverse in categories, wide in coverage, and rich in content. However, the planning and management of these videos have always relied on the expert experience of relevant departments. However, the costs of video creation and procurement are huge, and the error tolerance rate of video planning and management is extremely low. There is an urgent need for scientific and objective mathematical tools for assistance. The essence of video planning and management is to perform video similarity clustering, and the best data tool is video similarity.

[0003] In related technologies, there are various video similarity algorithms, including text similarity based on video content, similarity of collaborative filtering recommendation algorithms based on user behavior, etc. Among various video similarity algorithms, the simplest and with the fewest usage constraints is user overlap. Here, user overlap is regarded as video similarity. Among them, there are also many user overlap algorithms, and the most commonly used one is the Jaccard index (union-intersection ratio: the numerator is the union set, and the denominator is the intersection set).

[0004] However, user overlap is extremely susceptible to the influence of the user audience scale and the time interval between the online times of videos. For example, the user overlap between two videos with a larger user audience scale is obviously higher than that between two videos with a smaller user audience scale, and the user overlap between two videos with a smaller time interval between the online times is obviously higher than that between two videos with a larger time interval between the online times. As a result, the clustering result of video similarity clustering is inaccurate. Summary of the Invention

[0005] The purpose of the embodiments of the present invention is to provide a method, apparatus, electronic device, and storage medium for determining video similarity to achieve the beneficial effect of reasonable and reliable video similarity clustering results. The specific technical solutions are as follows:

[0006] In the first aspect of the embodiments of the present invention, first, a method for determining video similarity is provided, which includes:

[0007] Obtain a video set, and for any two videos in the video set, determine the user overlap between the two videos and determine the first user overlap between the two videos;

[0008] According to the user overlap and the first user overlap, determine the second user overlap between the two videos that eliminates the influence of the user audience scale;

[0009] Determine the influence coefficient of the time interval between the launches of the two videos, and process the second user overlap degree using the influence coefficient to obtain the third user overlap degree between the two videos;

[0010] Determine that the third user overlap degree is the video similarity between the two videos.

[0011] In an alternative embodiment, after obtaining the video set, the method further includes:

[0012] Determine the users corresponding to any video in the video set to form the audience set of the video, where the users include those who have effectively played the video;

[0013] The determination of the user overlap degree between the two videos includes:

[0014] Determine the user overlap degree between the two videos according to the users corresponding to the two videos respectively.

[0015] In an alternative embodiment, the determination of the user overlap degree between the two videos according to the users corresponding to the two videos respectively includes:

[0016] Input the users corresponding to the two videos respectively into the user overlap degree calculation formula, where the user overlap degree calculation formula includes:

[0017]

[0018] The A and the B include the users corresponding to the two videos respectively, and the r ab includes the user overlap degree;

[0019] Obtain the user overlap degree between the two videos output by the user overlap degree calculation formula.

[0020] In an alternative embodiment, the determination of the first user overlap degree between the two videos includes:

[0021] Determine the total users of the video website corresponding to the video set, and determine the first user overlap degree between the two videos according to the users corresponding to the two videos respectively and the total users.

[0022] In an alternative embodiment, the determination of the first user overlap degree between the two videos according to the users corresponding to the two videos respectively and the total users includes:

[0023] Input the users corresponding to each of the two videos and the total users into the first user overlap degree calculation formula, where the first user overlap degree calculation formula includes:

[0024]

[0025] The r s Includes the first user overlap degree, the A and the B include the users corresponding to each of the two videos, and the All includes the total users;

[0026] Obtain the first user overlap degree between the two videos output by the first user overlap degree calculation formula.

[0027] In an optional implementation manner, the determining the second user overlap degree between the two videos that eliminates the influence of the user audience scale according to the user overlap degree and the first user overlap degree includes:

[0028] Input the user overlap degree and the first user overlap degree into the second user overlap degree calculation formula, where the second user overlap degree calculation formula includes:

[0029]

[0030] The r ab Includes the user overlap degree, the r s Includes the first user overlap degree, the s ab Includes the second user overlap degree, the All includes the total users, and the A and the B include the users corresponding to each of the two videos;

[0031] Obtain the second user overlap degree between the two videos that eliminates the influence of the user audience scale output by the second user overlap degree calculation formula.

[0032] In an optional implementation manner, the determining the influence coefficient of the time interval between the two videos going online includes:

[0033] Determine the influence coefficient of the time interval between the two videos going online according to the users corresponding to each of the two videos.

[0034] In an optional implementation manner, the determining the influence coefficient of the time interval between the two videos going online according to the users corresponding to each of the two videos includes:

[0035] Input the users corresponding to each of the two videos into the influence coefficient calculation formula, where the influence coefficient calculation formula includes:

[0036]

[0037] The A and the B include the users corresponding to the respective two videos, the e -k×dt including the influence coefficient, the dt includes the online time interval between the two videos, and the k includes a constant;

[0038] Obtain the influence coefficients of the online time intervals of the two videos output by the influence coefficient calculation formula.

[0039] In an alternative embodiment, the processing of the user second coincidence degree by using the influence coefficient to obtain the user third coincidence degree between the two videos includes:

[0040] Perform a mirror transformation process on the influence coefficient to obtain an exponential relationship between the total users and the online time interval between the two videos;

[0041] Replace the total users with the exponential relationship and input it into the user second coincidence degree calculation formula as follows;

[0042] The e k×dt including the exponential relationship, the s ab including the user third coincidence degree, the dt includes the online time interval between the two videos, and the k includes a constant;

[0043] Obtain the user third coincidence degree between the two videos output by the user second coincidence degree calculation formula.

[0044] In the second aspect of the embodiments of the present invention, there is also provided a video similarity determination device, the device includes:

[0045] A coincidence degree determination module, configured to obtain a video set, and for any two videos in the video set, determine the user coincidence degree between the two videos; and,

[0046] A first coincidence degree determination module, configured to determine the user first coincidence degree between the two videos;

[0047] A second coincidence degree determination module, configured to determine the user second coincidence degree that excludes the influence of the user audience scale between the two videos according to the user coincidence degree and the user first coincidence degree;

[0048] An influence coefficient determination module, configured to determine the influence coefficient of the online time interval between the two videos;

[0049] A third coincidence degree processing module, configured to process the user second coincidence degree by using the influence coefficient to obtain the user third coincidence degree between the two videos;

[0050] A video similarity determination module, configured to determine that the third coincidence degree of the user is the video similarity between the two videos.

[0051] In a third aspect of the embodiments of the present invention, an electronic device is further provided, including a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the communication bus;

[0052] The memory is used to store a computer program;

[0053] The processor is configured to implement the video similarity determination method described in any one of the first aspects above when executing the program stored on the memory.

[0054] In a fourth aspect of the embodiments of the present invention, a storage medium is further provided. Instructions are stored in the storage medium. When it runs on a computer, the computer is caused to execute the video similarity determination method described in any one of the first aspects above.

[0055] In a fifth aspect of the embodiments of the present invention, a computer program product containing instructions is further provided. When it runs on a computer, the computer is caused to execute the video similarity determination method described in any one of the first aspects above.

[0056] The technical solution provided by the embodiments of the present invention is to obtain a video set, for any two videos in the video set, determine the user coincidence degree between the two videos, and determine the first coincidence degree of the user between the two videos. According to the user coincidence degree and the first coincidence degree of the user, determine the second coincidence degree of the user between the two videos after removing the influence of the user audience scale, determine the influence coefficient of the time interval between the two videos going online, and use the influence coefficient to process the second coincidence degree of the user to obtain the third coincidence degree of the user between the two videos, and determine that the third coincidence degree of the user is the video similarity between the two videos. For any two videos, through the user coincidence degree between the two videos and the first coincidence degree of the user between the two videos, the second coincidence degree of the user between the two videos after removing the influence of the user audience scale can be determined, and then the influence coefficient of the time interval between the two videos going online is used to process the second coincidence degree of the user to obtain the third coincidence degree of the user between the two videos as the video similarity between the two videos. In this way, the third coincidence degree of the user removes the influence of the user audience scale and the time interval between the videos going online, which can ensure that the clustering result of the subsequent video similarity clustering is reasonable and reliable. Description of the Drawings

[0057] The drawings here are incorporated into the specification and constitute a part of this specification, showing the embodiments that conform to the present invention, and are used together with the specification to explain the principles of the present invention.

[0058] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0059] Figure 1 It is a schematic flowchart of the implementation process of a video similarity determination method shown in the embodiments of the present invention;

[0060] Figure 2 It is a schematic diagram of the relationship between two pieces of content in the embodiments of the present invention and dt;

[0061] Figure 3 It is a schematic structural diagram of a video similarity determination device shown in the embodiments of the present invention;

[0062] Figure 4 It is a schematic structural diagram of an electronic device shown in the embodiments of the present invention. Detailed implementation manners

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0064] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0065] Currently, the user overlap is extremely vulnerable to the scale of the user audience and the time interval between the online times of videos. For example, the user overlap between two videos with a larger user audience scale is obviously higher than that between two videos with a smaller user audience scale. For instance, if the penetration rates of two videos on a video website are both 90%, then the user overlap of these two videos is at least 80%.

[0066] The user overlap between two videos with a smaller time interval between the online times is obviously higher than that between two videos with a larger time interval between the online times. Because the larger the time interval between the online times, the more likely users are to be inactive, change devices, and churn, resulting in a decrease in the user overlap. Using the traditional user overlap (Jaccard index) for video similarity clustering leads to inaccurate clustering results of video similarity. For example, it may cluster videos with a large user audience scale but completely different video audiences and themes into one category, and it may also cluster videos with different online times but similar themes and the same video audiences into multiple categories.

[0067] From an algorithm perspective, the reason for the above clustering results is that the existing user overlap algorithm will amplify the influence of the user audience scale and the time interval between the online times of videos, while underestimating the true video similarity. Based on this, the solution is to calibrate the user overlap with respect to the user audience scale and the time interval between the online times of videos, so that the third clustering result is reasonable and reliable.

[0068] Based on this, for any two videos, through the user overlap between the two videos and the first user overlap between the two videos, the second user overlap between the two videos that eliminates the influence of the user audience scale can be determined. Then, continue to process the second user overlap using the influence coefficient of the time interval between the online times of the two videos to obtain the third user overlap between the two videos as the video similarity between the two videos. In this way, the third user overlap eliminates the influence of the user audience scale and the time interval between the online times of videos, which can ensure that the clustering result of subsequent video similarity clustering is reasonable and reliable.

[0069] Specifically, as Figure 1 shown, it is a schematic flowchart of an implementation process of a method for determining video similarity provided by an embodiment of the present invention. This method can be used in a processor and specifically may include the following steps:

[0070] S101, obtain a video set. For any two videos in the video set, determine the user overlap between the two videos, and determine the first user overlap between the two videos.

[0071] In the embodiments of the present invention, a video set can be prepared in advance according to actual needs. Among them, the video set includes at least two videos, and the videos in the video set need to be planned and managed. For example, the business side can prepare a video set in advance. For the video set, it can include Movie A, Movie B, Movie C,....

[0072] Based on this, the embodiments of the present invention can obtain the pre-prepared video set. For any two videos in the video set, determine the user overlap between these two videos, and determine the first user overlap between these two videos, that is, determine the user overlap between every two videos, and determine the first user overlap between every two videos. Here, the first user overlap can be regarded as the reference value of the user overlap.

[0073] For example, for the video set, it can include Movie A, Movie B, Movie C,.... Taking Movie A and Movie B as an example, determine the user overlap between Movie A and Movie B, and determine the first user overlap between Movie A and Movie B. The processing of any other two movies is similar, and the embodiments of the present invention will not elaborate here one by one. The final processing results are shown in Table 1 below.

[0074] Any two movies User overlap First user overlap Movies a and b <![CDATA[r ab > <![CDATA[r s1 > Movies a and c <![CDATA[r ac > <![CDATA[r s2 > …… …… ……

[0075] Table 1

[0076] Among them, in the embodiments of the present invention, after obtaining the video set, for any video in the video set, determine the users corresponding to this video to form the audience set of this video. Among them, the users here refer to the users who have effectively played this video.

[0077] For example, after obtaining the video set, for Movie A in the video set, determine the users who have effectively played Movie A during the popular period as the audience set of Movie A. The processing of the remaining movies is similar, and the embodiments of the present invention will not elaborate here one by one. In this way, each movie has a corresponding audience set.

[0078] Based on this, for any two videos in the video set, the user overlap between these two videos can be determined according to the users corresponding to these two videos respectively, so that the user overlap between any two videos can be determined. Among them, the Jaccard index calculation formula can be used to determine the user overlap between any two videos.

[0079] Specifically, for any two videos in the video set, the users corresponding to these two videos can be input into the user overlap calculation formula. Here, the user overlap calculation formula is the Jaccard index calculation formula, and thus the user overlap between these two videos output by the user overlap calculation formula can be obtained.

[0080] Among them, the calculation formula for user overlap, that is, the calculation formula for the Jaccard index, is as follows:

[0081]

[0082] Here, r ab represents the user overlap between any two videos. A and B include the users corresponding to the two videos respectively, A∩B represents the number of users who have watched both video A and video B (intersection), and A∪B represents the number of users who have watched video A or video B (union).

[0083] In addition, in the embodiments of the present invention, the total users of the video website corresponding to the video set can be determined. For any two videos in the video set, the first user overlap between the two videos can be determined according to the users corresponding to the two videos respectively and the total users of the video website. In this way, the first user overlap between any two videos can be determined. Among them, the TGI (target group index) method can be used to determine the first user overlap between any two videos.

[0084] Specifically, for any two videos in the video set, the users corresponding to the two videos respectively and the total users of the video website are input into the first user overlap calculation formula to obtain the first user overlap between the two videos output by the first user overlap calculation formula. Among them, the first user overlap calculation formula includes:

[0085]

[0086] The r s includes the first user overlap, the A and the B include the users corresponding to the two videos respectively, and the All includes the total users of the video website.

[0087] S102. Determine the second user overlap between the two videos after eliminating the influence of the user audience scale according to the user overlap and the first user overlap.

[0088] In the embodiments of the present invention, for any two videos in the video set, the second user overlap between the two videos after eliminating the influence of the user audience scale is determined according to the user overlap between the two videos and the first user overlap between the two videos. Here, the second user overlap can be considered as the intermediate value of the user overlap.

[0089] In this way, for the user overlap, the influence of the user audience scale can be eliminated first. Among them, the user audience scale generally refers to the number of users with effective plays. That is, for each video, the number of users who have effectively played the video is used as the user audience of the video. In this way, the user audience scales of each video are different.

[0090] Among them, for any two videos in the video set, the user overlap degree between the two videos and the first user overlap degree between the two videos are input into the second user overlap degree calculation formula to obtain the second user overlap degree between the two videos output by the second user overlap degree calculation formula after removing the influence of the user audience scale.

[0091] Among them, the second user overlap degree calculation formula is as follows, which means dividing r in Formula 1 ab by r in Formula 2 s The second user overlap degree between any two videos after removing the influence of the user audience scale can be calculated as follows:

[0092]

[0093] The r ab includes the user overlap degree, the r s includes the first user overlap degree, the s ab includes the second user overlap degree, the All includes the total users, and the A and the B include the users corresponding to the two videos respectively.

[0094] S103. Determine the influence coefficient of the time interval between the online times of the two videos, and process the second user overlap degree by using the influence coefficient to obtain the third user overlap degree between the two videos.

[0095] In the embodiment of the present invention, All in the above Formula 3 is actually a value that will change and is difficult to be replaced by a constant value. However, the basic assumption that the similarity between pairwise videos is independent of the time interval between the online times of pairwise videos can be used to fit All. Therefore, Formula 3 can be written as:

[0096]

[0097] Now it is known that s ab is independent of dt (the time interval between the online times of pairwise videos), but according to the actual calculation it will be found that has a negative exponential relationship with dt, as shown below. This relationship is stably reproduced and the slope does not change with the increase or decrease of the video samples. Here, the video can be regarded as content, then the relationship between pairwise content and dt is as Figure 2 shown.

[0098]

[0099] Therefore, in order to ensure s abIrrespective of dt, there must be an exponential relationship between All and dt. Based on this idea, the following processing is carried out:

[0100] In an embodiment of the present invention, for any two videos in the video set, an influence coefficient of the online time interval between the two videos is determined. Among them, for any two videos in the video set, according to the users corresponding to the two videos respectively, the influence coefficient of the online time interval between the two videos is determined.

[0101] Among them, for any two videos in the video set, the users corresponding to the two videos respectively are input into the influence coefficient calculation formula, and the influence coefficient of the online time interval between the two videos output by the influence coefficient calculation formula is obtained.

[0102] Among them, the influence coefficient calculation formula is the above formula 5, where A and B include the users corresponding to the two videos respectively, and e -k×dt includes the influence coefficient, dt includes the online time interval between the two videos, and k includes a constant.

[0103] Thus, using the influence coefficient of the online time interval between the two videos, the second user overlap degree between the two videos is processed to obtain the third user overlap degree between the two videos. Here, the third user overlap degree can be considered as the final value of the user overlap degree.

[0104] Specifically, the influence coefficient of the online time interval between the two videos is subjected to mirror conversion processing to obtain an exponential relationship between the total users of the video website and the online time interval between the two videos, that is, the exponential relationship between All and dt.

[0105] Replace the total users of the video website in the above formula 3 with the above exponential relationship, and re-enter it into the second user overlap degree calculation formula, that is, substitute the exponential relationship between All and dt into formula 3 to obtain formula 6 shown below, and obtain the second user overlap degree calculation formula, that is, the third user overlap degree between the two videos output by the following formula 6.

[0106] The e k×dt includes the exponential relationship, and s ab includes the third user overlap degree, dt includes the online time interval between the two videos, and k includes a constant.

[0107] S104, determine that the third user overlap degree is the video similarity between the two videos.

[0108] In an embodiment of the present invention, for any two videos in a video set, a third overlap degree between the two videos that excludes the influence of the user audience scale and the time interval between the online times of the videos can be determined, and this third overlap degree is determined as the video similarity between the two videos. Thus, for any two videos in the video set, there is a corresponding video similarity.

[0109] Based on the video similarity between any two videos in the video set, clustering is performed on the videos in the video set. Among them, the clustering algorithm used can be the K-Medoids clustering method, which uses the median value to determine the cluster center and minimizes the influence of outliers on the clustering result. In this way, the subsequent clustering result of the video similarity clustering can be ensured to be reasonable and reliable.

[0110] Through the above description of the technical solution provided by the embodiment of the present invention, a video set is obtained. For any two videos in the video set, the user overlap degree between the two videos is determined, and the first user overlap degree between the two videos is determined. According to the user overlap degree and the first user overlap degree, the second user overlap degree between the two videos that excludes the influence of the user audience scale is determined. The influence coefficient of the time interval between the two videos going online is determined, and the second user overlap degree is processed using the influence coefficient to obtain the third user overlap degree between the two videos, and the third user overlap degree is determined as the video similarity between the two videos.

[0111] For any two videos, through the user overlap degree between the two videos and the first user overlap degree between the two videos, the second user overlap degree between the two videos that excludes the influence of the user audience scale can be determined. Then, the second user overlap degree is further processed using the influence coefficient of the time interval between the two videos going online to obtain the third user overlap degree between the two videos as the video similarity between the two videos. In this way, the third user overlap degree excludes the influence of the user audience scale and the time interval between the videos, which can ensure that the subsequent clustering result of the video similarity clustering is reasonable and reliable.

[0112] Corresponding to the above method embodiment, an embodiment of the present invention further provides a video similarity determination device, as Figure 3 shown. The device may include: an overlap degree determination module 310, a first overlap degree determination module 320, a second overlap degree determination module 330, an influence coefficient determination module 340, a third overlap degree processing module 350, and a video similarity determination module 360.

[0113] The overlap degree determination module 310 is configured to obtain a video set and determine the user overlap degree between any two videos in the video set; and,

[0114] The first overlap degree determination module 320 is configured to determine the first user overlap degree between the two videos;

[0115] A second coincidence degree determination module 330, configured to determine a user second coincidence degree between the two videos after eliminating the influence of the user audience scale according to the user coincidence degree and the user first coincidence degree;

[0116] An influence coefficient determination module 340, configured to determine an influence coefficient of the time interval between the two videos going online;

[0117] A third coincidence degree processing module 350, configured to process the user second coincidence degree by using the influence coefficient to obtain a user third coincidence degree between the two videos;

[0118] A video similarity determination module 360, configured to determine that the user third coincidence degree is the video similarity between the two videos.

[0119] An embodiment of the present invention further provides an electronic device, as Figure 4 shown, including a processor 41, a communication interface 42, a memory 43, and a communication bus 44. Among them, the processor 41, the communication interface 42, and the memory 43 complete communication with each other through the communication bus 44,

[0120] The memory 43 is used to store a computer program;

[0121] The processor 41, when executing the program stored on the memory 43, implements the following steps:

[0122] Obtain a video set. For any two videos in the video set, determine the user coincidence degree between the two videos and determine the user first coincidence degree between the two videos; according to the user coincidence degree and the user first coincidence degree, determine the user second coincidence degree between the two videos after eliminating the influence of the user audience scale; determine the influence coefficient of the time interval between the two videos going online, and process the user second coincidence degree by using the influence coefficient to obtain the user third coincidence degree between the two videos; determine that the user third coincidence degree is the video similarity between the two videos.

[0123] The communication bus mentioned in the above electronic device may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity, only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus.

[0124] The communication interface is used for communication between the above-mentioned electronic device and other devices.

[0125] The memory may include a random access memory (RAM) and may also include a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0126] The above-mentioned processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0127] In another embodiment provided by the present invention, a storage medium is further provided. Instructions are stored in the storage medium, and when it runs on a computer, the computer is made to execute the video similarity determination method described in any one of the above embodiments.

[0128] In another embodiment provided by the present invention, a computer program product containing instructions is further provided. When it runs on a computer, the computer is made to execute the video similarity determination method described in any one of the above embodiments.

[0129] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a storage medium or transmitted from one storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.). The storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more integrated available media. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)).

[0130] It should be noted that, in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising a..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0131] Each embodiment in this specification is described in a related manner. The same or similar parts among the embodiments can be referred to each other, and the differences between each embodiment and other embodiments are emphasized. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiment.

[0132] The above is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A method for determining video similarity, characterized in that, The method includes: Obtaining a video set, determining the users corresponding to any video in the video set to form the audience set of the video, where the users include those who have effectively played the video; For any two videos in the video set, determining the user overlap between the two videos, including: determining the user overlap between the two videos according to the users corresponding to the two videos respectively, and determining the first user overlap between the two videos, including: determining the total users of the video website corresponding to the video set, and determining the first user overlap between the two videos according to the users corresponding to the two videos respectively and the total users; According to the user overlap and the first user overlap, determining the second user overlap between the two videos after eliminating the influence of the user audience scale, including: dividing the user overlap by the first user overlap to obtain the second user overlap; Determining the influence coefficient of the time interval between the two videos going online, and processing the second user overlap with the influence coefficient to obtain the third user overlap between the two videos, including: performing a mirror conversion process on the influence coefficient to obtain the exponential relationship between the total users and the time interval between the two videos going online; replacing the total users with the exponential relationship and inputting it into the second user overlap calculation formula; obtaining the third user overlap between the two videos output by the second user overlap calculation formula; Determining that the third user overlap is the video similarity between the two videos.

2. The method according to claim 1, wherein The determining the user overlap between the two videos according to the users corresponding to the two videos respectively includes: Inputting the users corresponding to the two videos respectively into the user overlap calculation formula, where the user overlap calculation formula includes: The A and the B include the users corresponding to the respective two videos, and the r ab includes the user overlap; Obtaining the user overlap between the two videos output by the user overlap calculation formula.

3. The method according to claim 1, characterized in that, The determining the first user overlap between the two videos according to the users corresponding to the two videos respectively and the total users includes: Inputting the users corresponding to the two videos respectively and the total users into the first user overlap calculation formula, where the first user overlap calculation formula includes: The said r s including the first coincidence degree of the said user, the said A and the said B include the said users corresponding to the respective two said videos, and the said All includes the said total users; Obtaining the first user overlap between the two videos output by the first user overlap calculation formula.

4. The method according to claim 1, wherein The determining the influence coefficient of the time interval between the two videos going online includes: Determining the influence coefficient of the time interval between the two videos going online according to the users corresponding to the two videos respectively.

5. The method according to claim 4, wherein The determining the influence coefficient of the time interval between the two videos going online according to the users corresponding to the two videos respectively includes: Inputting the users corresponding to the two videos respectively into the influence coefficient calculation formula, where the influence coefficient calculation formula includes: The A and the B include the users corresponding to the respective two videos, the e -k×dt include the influence coefficient, the dt includes the online time interval between the two videos, and the k includes a constant; Obtaining the influence coefficient of the time interval between the two videos going online output by the influence coefficient calculation formula.

6. The method according to claim 1, characterized in that, The processing the second user overlap with the influence coefficient to obtain the third user overlap between the two videos includes: Perform a mirror transformation process on the influence coefficient to obtain an exponential relationship between the total number of users and the time intervals between the launches of the two videos; Replace the total number of users with the exponential relationship and input it into the user second coincidence degree calculation formula as follows; The said e k×dt includes the said exponential relationship, the s ab includes the third coincidence degree of the user, the dt includes the online time interval between the two videos, and the k includes a constant; Obtain the user third coincidence degree between the two videos output by the user second coincidence degree calculation formula.

7. A video similarity determination device, characterized in that The device includes: A coincidence degree determination module, configured to obtain a video set, determine the users corresponding to any video in the video set, and form an audience set for the video, where the users include users who have effectively played the video; for any two videos in the video set, determine the user coincidence degree between the two videos, including: determining the user coincidence degree between the two videos according to the users corresponding to the two videos respectively; and A first coincidence degree determination module, configured to determine the user first coincidence degree between the two videos, including: determining the total number of users of the video website corresponding to the video set, and determining the user first coincidence degree between the two videos according to the users corresponding to the two videos respectively and the total number of users; A second coincidence degree determination module, configured to determine the user second coincidence degree between the two videos excluding the influence of the user audience scale according to the user coincidence degree and the user first coincidence degree, including: dividing the user coincidence degree by the user first coincidence degree to obtain the user second coincidence degree; An influence coefficient determination module, configured to determine the influence coefficient of the time interval between the launches of the two videos; A third coincidence degree processing module, configured to process the user second coincidence degree by using the influence coefficient to obtain the user third coincidence degree between the two videos, including: performing a mirror transformation process on the influence coefficient to obtain an exponential relationship between the total number of users and the time intervals between the launches of the two videos; replacing the total number of users with the exponential relationship and inputting it into the user second coincidence degree calculation formula; obtaining the user third coincidence degree between the two videos output by the user second coincidence degree calculation formula; A video similarity determination module, configured to determine that the user third coincidence degree is the video similarity between the two videos.

8. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the communication bus; The memory is used to store a computer program; The processor, when executing the program stored on the memory, implements the method steps described in any one of claims 1-6.

9. A storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method described in any one of claims 1-6.

Images