Cross-platform video promotion system and method based on artificial intelligence

Abstract

The invention relates to the field of video promotion, in particular to a cross-platform video promotion system and method based on artificial intelligence, and the system comprises a platform data module, a video association module, a cross-platform transmission module, a video recommendation module and an advertisement insertion module, the video association module is used for generating a joint influence function; the cross-platform transmission module is used for classifying user groups and carrying out classification data cross-platform through a cloud end; the video recommendation module is used for judging the group type of cross-platform users; the advertisement insertion module is used for inserting advertisements in videos and generating video streams; according to the method, the personal privacy of a single user can be protected, the intelligence and the safety of an algorithm are improved, meanwhile, the content strategy and the video pushing arrangement are intelligently adjusted, higher platform popularity is achieved, the exposure rate and the conversion rate of video advertisements are effectively improved, the popularization cost is saved, and benefits are increased.

Description

Technical Field

[0001] The present invention relates to the field of video promotion, and in particular to a cross-platform video promotion system and method based on artificial intelligence. Background Art

[0002] Video promotion is a promotion method that promotes products, services or brands to target audiences through online video media, and achieves product promotion goals by inserting advertisements in video content or playing advertisements on the platform. Since video promotion can be combined with various data analysis tools for effect tracking and evaluation, it can achieve accurate user group conversion.

[0003] There are many platforms that can play videos. If data is not exchanged between platforms, a lot of algorithm resources will be wasted. Therefore, cross-platform promotion algorithms came into being. Cross-platform promotion involves multiple platforms and data sources, but data synchronization and integration between different platforms are difficult to accurately judge users' preferences on different platforms. In addition, user preferences involve personal privacy, and there is also a risk of data leakage when transmitting privacy information across platforms.

[0004] In addition, ads affect the user's video viewing experience and need to be inserted evenly into the viewing process. However, users watch different videos for different lengths of time, making it difficult to accurately determine the ad delivery window. The current cross-platform video promotion algorithm may have the problem of over-delivery of ads, which affects the effective reach and conversion rate of ads, resulting in poor advertising results. Summary of the invention

[0005] The purpose of the present invention is to provide a cross-platform video promotion system and method based on artificial intelligence to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a cross-platform video promotion system based on artificial intelligence, comprising: a platform data module, a video association module, a cross-platform transmission module, a video recommendation module and an advertisement insertion module;

[0007] The platform data module is composed of a database of the video platform, platform software and a local server, and is used to record and store user information, video element information, interaction data and playback history, set a neural network model in the local server, and store the interaction information of a single user in a data pool of the server as a training library for the neural network model;

[0008] The video association module is used to extract the user's historical playback data with the user's username as the primary key, and store the user's historical playback data in a training pool of a local server for training a neural network model for a single user's video playback behavior to obtain a loss model. The loss model takes video elements as input and user interaction data as output, fits the association model between the interaction data and each video element, and generates a joint influence function between the video playback time and each video element;

[0009] The cross-platform transmission module is used to cut the joint influence function according to the preset definition domain interval of each video element, classify the user group according to the cutting result, and for each type of user group, package the joint influence function of all users in the group and upload it to the cloud. The cloud classifies and summarizes the data packets from each platform with the group label as the primary key, and receives the cross-platform data packet call;

[0010] The video recommendation module is used to parse the cloud data packet on the second platform, perform consistency fitting on the joint influence function in each type of group label data, and generate a typical influence function corresponding to the group label. Each time the user plays a video on the second platform, the video elements are input into each typical function, and a predicted sequence of user interaction data is output. The distance between the actual interaction data and each element of the sequence is used as the cumulative playback coefficient of the group label corresponding to the element;

[0011] The advertisement insertion module is used to extract videos from the video library under each video element classification according to the playback coefficient, and use the video playback time as a constraint condition to calculate the video arrangement when the length of each video advertisement is the longest, generate a video stream according to the video arrangement, and make automatic recommendations to users.

[0012] Further, the platform data module includes: a database unit and an artificial intelligence unit;

[0013] The database unit is used to store user information and video playback elements, the video playback elements include: video tags, duration, creator, picture quality and advertisement length, and interactive data include: click-through rate and video playback duration;

[0014] The artificial intelligence unit is set in a local database to analyze video playback data, and the neural network models used include: R-CNN model, DeepLab model, U-Net model and Watershed model.

[0015] Further, the video association module includes: a video classification unit, a loss model unit and a joint impact unit;

[0016] The video classification unit is used to extract the user's historical playback data and classify the historical playback data according to video elements;

[0017] The loss model unit is used to train a neural network model to fit the influence relationship of each video element on user interaction data;

[0018] The joint influence unit is used to express the fitting result in a functional form, and to generate a joint influence function with the interaction data as the dependent variable and each video element as the independent variable.

[0019] Further, the cross-platform transmission module includes: a group preference unit and a cloud communication unit;

[0020] The group preference unit is used to divide user groups according to the range of video elements and compress the joint influence functions corresponding to the groups into data packets;

[0021] The cloud communication unit is used to classify and aggregate data packets from various platforms according to group labels, and is responsible for cloud uploading and downloading of data packets.

[0022] Furthermore, the video recommendation module includes: a user identification unit, a group analysis unit and a play accumulation unit;

[0023] The user identification unit is used to fit the joint influence function in each data packet to generate a typical influence function, wherein the typical influence function represents the average preference level of the group label;

[0024] The group analysis unit is used to obtain the user's operation behavior on the second platform and determine the user's group label;

[0025] The playback accumulation unit is used to score the group label to which the user belongs according to the difference between the user's actual interaction data and the interaction data predicted by each typical influence function.

[0026] Further, the advertisement insertion module includes: a video extraction unit and a constraint conversion unit;

[0027] The video extraction unit is used to partition and store the video according to the intervals where the video elements are located, and extract videos from each partition according to the user's group label;

[0028] The constraint conversion unit is used to control the arrangement of the extracted videos using the video playback duration as a constraint condition so as to maximize the sum of the video advertisement lengths.

[0029] A cross-platform video promotion method based on artificial intelligence, comprising the following steps:

[0030] Step S1. Collect user interaction data and video playback elements in the first platform, and for each user ID in the platform, generate a data table in the local server with the interaction data as the table header and the video playback elements as the table items;

[0031] Step S2. Generate a training library from the data tables of all users in a preset period, train a neural network model with the data in the training library, and obtain a loss model. The loss model takes the video playback elements as input and the user's interaction data as output. The loss model is used to fit the influence of the video playback elements on the interaction data of each user to obtain a joint influence function;

[0032] Step S3. Classify the videos according to the video playback elements, substitute the video playback elements in each category into the joint influence function, and classify the joint influence function according to the video type where the joint influence function takes the maximum value;

[0033] Step S4. Pack the joint influence function according to the classification results, mark the obtained data packet with the video type and upload it to the cloud. The cloud classifies and summarizes the data packets from each platform with the video type as the keyword, and receives the cross-platform data packet retrieval;

[0034] Step S5. The second platform parses the data packet, performs consistency fitting on the joint influence function of each video type, obtains a typical function, substitutes the playback elements of the video watched by the user on the second platform into each typical function, accumulates scores for each group label according to the difference between the predicted results and the actual interaction data, extracts videos according to the score ratio, and generates a recommended video stream.

[0035] Further, step S1 includes:

[0036] Step S11. Collecting user interaction data and video playback elements on the first platform, wherein the interaction data is the video click-through rate or the video playback duration, and the video playback elements include: video tags, duration, creator, picture quality, and advertisement length;

[0037] Step S12. For each login behavior of each user ID, a data linked list is constructed, and the data linked list has n rows and m columns, where n is the total number of items of interactive data and video playback elements, and m is the total number of times the user plays videos on the platform during the current login. The interactive data is fixed as the header of the data linked list, and the video playback element is the table item of the data linked list. All the data linked lists generated under the user ID are stored in the local server.

[0038] Further, step S2 includes:

[0039] Step S21. The platform statistical cycle is preset, and the local server generates a new data space every other statistical cycle as the storage location of the newly generated data linked list within the statistical cycle;

[0040] Step S22. Using a neural network model to learn a training library consisting of all data linked lists under a single user ID;

[0041] The video playback elements in the training library are used as input items, and the user interaction data is used as output items. For each data linked list, the functional relationship between the output variable and the input variable is fitted. n-1 ), where x n-1 represents the n-1th input variable, and the neural network models include: R-CNN model, DeepLab model, U-Net model and Watershed model;

[0042] Step S23. According to the selected model category, use elastic network regression to fit the loss function:

[0043]

[0044] Where LOSS is the loss function, SE is the root mean square error of the model, Y is the regularization strength of the model, k is the number of data lists in the training library, wi represents the preset weight of the i-th data list in the training library, r is the proportion of the regularization term, Fi(x1,x2,…,x n-1 ) represents the functional relationship between the output variable and the input variable in the i-th data linked list, represents the function Fi(x1,x2,…,x n-1 )’s linear regression function;

[0045] Step S24. Calculate the regularization strength Y and the regularization term ratio r when the loss function LOSS takes the minimum value, and generate a positive definite matrix P according to the values ​​of Y and r, so that G(x1, x2, ..., x n-1 )=P·[x1,x2,…,x n-1 ] T , where T is the transpose symbol, G(x1,x2,…,x n-1 ) is the joint influence function.

[0046] Further, step S3 includes:

[0047] Step S31. Arrange and combine the playback elements of the video to obtain h video types, substitute the playback elements corresponding to each video type into the joint influence function, and the obtained function values ​​constitute a prediction set. Classify the joint influence function into the video type corresponding to the maximum value in the prediction set;

[0048] Step S32: Classify the joint influence function corresponding to each user ID, and hide the user ID in the classification result, and only keep the function formula.

[0049] Further, step S4 includes:

[0050] Step S41. Pack all joint influence functions contained in each video type and upload them to the cloud. The cloud obtains the data packets sent by each platform and classifies and summarizes the data packets using the video type as a keyword.

[0051] Step S42. When the user logs in to the second platform for the first time, the second platform downloads all data packets from the cloud and stores them in isolation in the local server according to the video type.

[0052] Further, step S5 includes:

[0053] Step S51. The second platform generates a typical function of each video type in the local server through a consistency equation:

[0054]

[0055] Among them, B(x1,x2,…,x n-1 ) represents the typical function of the video type, v represents the number of joint influence functions in the video type, Gj(x1,x2,…,x n-1 ) represents the jth joint influence function in the video type;

[0056] Solve the consistency equation to obtain the typical function corresponding to each video type;

[0057] Step S52. Obtain the video playback elements of the video that the user has played in the second platform, substitute the video playback elements into the typical functions of each video type, and record the obtained values ​​as the value set {A1, A2, ..., At, ..., Ah}, where At represents the function value of the typical function in the t-th video type;

[0058] Step S53. Obtain the actual interaction data A0 of the user for the played video, calculate the difference AS between each element in the value set and the actual interaction data, where AS = |At-A0|, and is the cumulative score UR of the video type corresponding to the element, where UR = 1 / AS;

[0059] Step S54: extract videos from the video library corresponding to each video type according to the accumulated scores, generate video streams, and push the video streams to the user.

[0060] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0061] 1. The present invention obtains various types of elements and interaction parameters of the video according to the user's video playback history on the platform. For the video playback behavior of a single user, the loss model is used to fit the association model between the interaction parameters and each video element, and the cumulative playback time and the joint influence function between various factors are obtained. The user's preferences and reactions to the video content can be understood, so as to better grasp the audience's needs, adjust the content strategy and video push arrangement, and achieve higher platform popularity.

[0062] 2. The present invention can cut the joint influence function with each video element as the definition domain, thereby classifying users according to the factor influence dimension, sending the classification results to the cloud via a cross-platform communication protocol, and performing cross-platform big data management with group preferences instead of individual preferences, thereby protecting the personal privacy of individual users and improving the intelligence and security of the algorithm.

[0063] 3. The present invention uses the actual viewing time and the distance between the set elements as the cumulative playback coefficient of each type, extracts videos according to the playback coefficient, and uses the interactive parameters as constraints to generate video streams in a video arrangement method when the length of each video advertisement is the largest. This can effectively improve the exposure and conversion rate of video advertisements, save promotion costs and increase benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0065] Figure 1 It is a structural schematic diagram of a cross-platform video promotion system based on artificial intelligence of the present invention;

[0066] Figure 2 It is a schematic diagram of the steps of an artificial intelligence-based cross-platform video promotion method of the present invention. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] See also Figure 1 ,The present invention provides a technical solution: an artificial intelligence-based cross-platform video promotion system, comprising: a platform data module, a video association module, a cross-platform transmission module, a video recommendation module and an advertisement insertion module;

[0069] The platform data module is composed of a database of the video platform, platform software and a local server, and is used to record and store user information, video element information, interaction data and playback history, set a neural network model in the local server, and store the interaction information of a single user in a data pool of the server as a training library for the neural network model;

[0070] The platform data module includes: a database unit and an artificial intelligence unit;

[0071] The database unit is used to store user information and video playback elements, the video playback elements include: video tags, duration, creator, picture quality and advertisement length, and interactive data include: click-through rate and video playback duration;

[0072] The artificial intelligence unit is set in a local database to analyze video playback data, and the neural network models used include: R-CNN model, DeepLab model, U-Net model and Watershed model.

[0073] The video association module is used to extract the user's historical playback data with the user's username as the primary key, and store the user's historical playback data in a training pool of a local server for training a neural network model for a single user's video playback behavior to obtain a loss model. The loss model takes video elements as input and user interaction data as output, fits the association model between the interaction data and each video element, and generates a joint influence function between the video playback time and each video element;

[0074] The video association module includes: a video classification unit, a loss model unit and a joint impact unit;

[0075] The video classification unit is used to extract the user's historical playback data and classify the historical playback data according to video elements;

[0076] The loss model unit is used to train a neural network model to fit the influence relationship of each video element on user interaction data;

[0077] The joint influence unit is used to express the fitting result in a functional form, and to generate a joint influence function with the interaction data as the dependent variable and each video element as the independent variable.

[0078] The cross-platform transmission module is used to cut the joint influence function according to the preset definition domain interval of each video element, classify the user group according to the cutting result, and for each type of user group, package the joint influence function of all users in the group and upload it to the cloud. The cloud classifies and summarizes the data packets from each platform with the group label as the primary key, and receives the cross-platform data packet call;

[0079] The cross-platform transmission module includes: a group preference unit and a cloud communication unit;

[0080] The group preference unit is used to divide user groups according to the range of video elements and compress the joint influence functions corresponding to the groups into data packets;

[0081] The cloud communication unit is used to classify and aggregate data packets from various platforms according to group labels, and is responsible for cloud uploading and downloading of data packets.

[0082] The video recommendation module is used to parse the cloud data packet on the second platform, perform consistency fitting on the joint influence function in each type of group label data, and generate a typical influence function corresponding to the group label. Each time the user plays a video on the second platform, the video elements are input into each typical function, and a predicted sequence of user interaction data is output. The distance between the actual interaction data and each element of the sequence is used as the cumulative playback coefficient of the group label corresponding to the element;

[0083] The video recommendation module includes: a user identification unit, a group analysis unit and a play accumulation unit;

[0084] The user identification unit is used to fit the joint influence function in each data packet to generate a typical influence function, wherein the typical influence function represents the average preference level of the group label;

[0085] The group analysis unit is used to obtain the user's operation behavior on the second platform and determine the user's group label;

[0086] The playback accumulation unit is used to score the group label to which the user belongs according to the difference between the user's actual interaction data and the interaction data predicted by each typical influence function.

[0087] The advertisement insertion module is used to extract videos from the video library under each video element classification according to the playback coefficient, and use the video playback time as a constraint condition to calculate the video arrangement when the length of each video advertisement is the longest, generate a video stream according to the video arrangement, and make automatic recommendations to users.

[0088] The advertisement insertion module includes: a video extraction unit and a constraint conversion unit;

[0089] The video extraction unit is used to partition and store the video according to the intervals where the video elements are located, and extract videos from each partition according to the user's group label;

[0090] The constraint conversion unit is used to control the arrangement of the extracted videos using the video playback duration as a constraint condition so as to maximize the sum of the video advertisement lengths.

[0091] like Figure 2 As shown, a cross-platform video promotion method based on artificial intelligence includes the following steps:

[0092] Step S1. Collect user interaction data and video playback elements in the first platform, and for each user ID in the platform, generate a data table in the local server with the interaction data as the table header and the video playback elements as the table items;

[0093] Step S1 includes:

[0094] Step S11. Collecting user interaction data and video playback elements on the first platform, wherein the interaction data is the video click-through rate or the video playback duration, and the video playback elements include: video tags, duration, creator, picture quality, and advertisement length;

[0095] Step S12. For each login behavior of each user ID, a data linked list is constructed, and the data linked list has n rows and m columns, where n is the total number of items of interactive data and video playback elements, and m is the total number of times the user plays videos on the platform during the current login. The interactive data is fixed as the header of the data linked list, and the video playback element is the table item of the data linked list. All the data linked lists generated under the user ID are stored in the local server.

[0096] Step S2. Generate a training library from the data tables of all users in a preset period, train a neural network model with the data in the training library, and obtain a loss model. The loss model takes the video playback elements as input and the user's interaction data as output. The loss model is used to fit the influence of the video playback elements on the interaction data of each user to obtain a joint influence function;

[0097] Step S2 includes:

[0098] Step S21. The platform statistical cycle is preset, and the local server generates a new data space every other statistical cycle as the storage location of the newly generated data linked list within the statistical cycle;

[0099] Step S22. Using a neural network model to learn a training library consisting of all data linked lists under a single user ID;

[0100] The video playback elements in the training library are used as input items, and the user interaction data is used as output items. For each data linked list, the functional relationship between the output variable and the input variable is fitted. n-1 ), where x n-1 represents the n-1th input variable, and the neural network models include: R-CNN model, DeepLab model, U-Net model and Watershed model;

[0101] Step S23. According to the selected model category, use elastic network regression to fit the loss function:

[0102]

[0103] Where LOSS is the loss function, SE is the root mean square error of the model, Y is the regularization strength of the model, k is the number of data lists in the training library, wi represents the preset weight of the i-th data list in the training library, r is the proportion of the regularization term, Fi(x1,x2,…,x n-1) represents the functional relationship between the output variable and the input variable in the i-th data linked list, represents the function Fi(x1,x2,…,x n-1 )’s linear regression function;

[0104] Step S24. Calculate the regularization strength Y and the regularization term ratio r when the loss function LOSS takes the minimum value, and generate a positive definite matrix P according to the values ​​of Y and r, so that G(x1, x2, ..., x n-1 )=P·[x1,x2,…,x n-1 ] T , where T is the transpose symbol, G(x1,x2,…,x n-1 ) is the joint influence function.

[0105] Step S3. Classify the videos according to the video playback elements, substitute the video playback elements in each category into the joint influence function, and classify the joint influence function according to the video type where the joint influence function takes the maximum value;

[0106] Step S3 includes:

[0107] Step S31. Arrange and combine the playback elements of the video to obtain h video types, substitute the playback elements corresponding to each video type into the joint influence function, and the obtained function values ​​constitute a prediction set. Classify the joint influence function into the video type corresponding to the maximum value in the prediction set;

[0108] Step S32: Classify the joint influence function corresponding to each user ID, and hide the user ID in the classification result, and only keep the function formula.

[0109] Step S4. Pack the joint influence function according to the classification results, mark the obtained data packet with the video type and upload it to the cloud. The cloud classifies and summarizes the data packets from each platform with the video type as the keyword, and receives the cross-platform data packet retrieval;

[0110] Step S4 includes:

[0111] Step S41. Pack all joint influence functions contained in each video type and upload them to the cloud. The cloud obtains the data packets sent by each platform and classifies and summarizes the data packets using the video type as a keyword.

[0112] Step S42. When the user logs in to the second platform for the first time, the second platform downloads all data packets from the cloud and stores them in isolation in the local server according to the video type.

[0113] Step S5. The second platform parses the data packet, performs consistency fitting on the joint influence function of each video type, obtains a typical function, substitutes the playback elements of the video watched by the user on the second platform into each typical function, accumulates scores for each group label according to the difference between the predicted results and the actual interaction data, extracts videos according to the score ratio, and generates a recommended video stream.

[0114] Step S5 includes:

[0115] Step S51. The second platform generates a typical function of each video type in the local server through a consistency equation:

[0116]

[0117] Among them, B(x1,x2,…,x n-1 ) represents the typical function of the video type, v represents the number of joint influence functions in the video type, Gj(x1,x2,…,x n-1 ) represents the jth joint influence function in the video type;

[0118] Solve the consistency equation to obtain the typical function corresponding to each video type;

[0119] Step S52. Obtain the video playback elements of the video that the user has played in the second platform, substitute the video playback elements into the typical functions of each video type, and record the obtained values ​​as the value set {A1, A2, ..., At, ..., Ah}, where At represents the function value of the typical function in the t-th video type;

[0120] Step S53. Obtain the actual interaction data A0 of the user for the played video, calculate the difference AS between each element in the value set and the actual interaction data, where AS = |At-A0|, and is the cumulative score UR of the video type corresponding to the element, where UR = 1 / AS;

[0121] Step S54: extract videos from the video library corresponding to each video type according to the accumulated scores, generate video streams, and push the video streams to the user.

[0122] Embodiment: There are 2 users and 4 video elements on the first platform. User 1 watched a video with video elements (1, 2, 1, 4) for 20 minutes in one login, and user 2 watched a video with video elements (2, 1, 3, 3) for 50 minutes, and did not watch the other videos. Then, joint influence functions F1 and F2 are generated for user 1 and user 2 respectively, where the output of F1 is 20 when the input is (1, 2, 1, 4), and the output is 0 otherwise, and the output of F2 is 50 when the input is (2, 1, 3, 3), and the output is 0 otherwise. After uploading to the cloud, the user logs in to the second platform and watches a video with video elements (1, 2, 1, 4) for 10 minutes. It is then judged that the viewer belongs to the group to which user 1 belongs, and a video stream is generated according to the group preference.

[0123] It should 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.

[0124] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A cross-platform video promotion method based on artificial intelligence, characterized in that: The method comprises the following steps: Step S1. Collect user interaction data and video playback elements in the first platform, and for each user ID in the platform, generate a data table in the local server with the interaction data as the table header and the video playback elements as the table items; Step S2. Generate a training library from the data tables of all users in a preset period, train a neural network model with the data in the training library, and obtain a loss model. The loss model takes the video playback elements as input and the user's interaction data as output. The loss model is used to fit the influence of the video playback elements on the interaction data of each user to obtain a joint influence function; Step S3. Classify the videos according to the video playback elements, substitute the video playback elements in each category into the joint influence function, and classify the joint influence function according to the video type where the joint influence function takes the maximum value; Step S4. Pack the joint influence function according to the classification results, mark the obtained data packet with the video type and upload it to the cloud. The cloud classifies and summarizes the data packets from each platform with the video type as the keyword, and receives the cross-platform data packet retrieval; Step S5. The second platform parses the data packet, performs consistency fitting on the joint influence function of each video type, obtains a typical function, substitutes the playback elements of the video watched by the user on the second platform into each typical function, accumulates scores for each group label according to the difference between the predicted results and the actual interaction data, extracts videos according to the score ratio, and generates a recommended video stream.

2. The cross-platform video promotion method based on artificial intelligence according to claim 1, characterized in that: Step S1 includes: Step S11. Collecting user interaction data and video playback elements on the first platform, wherein the interaction data is the video click-through rate or the video playback duration, and the video playback elements include: video tags, duration, creator, picture quality, and advertisement length; Step S12. For each login behavior of each user ID, a data linked list is constructed, and the data linked list has n rows and m columns, where n is the total number of items of interactive data and video playback elements, and m is the total number of times the user plays videos on the platform during the current login. The interactive data is fixed as the header of the data linked list, and the video playback element is the table item of the data linked list. All the data linked lists generated under the user ID are stored in the local server.

3. The cross-platform video promotion method based on artificial intelligence according to claim 2, characterized in that: Step S2 includes: Step S21. The platform statistical cycle is preset, and the local server generates a new data space every other statistical cycle as the storage location of the newly generated data linked list within the statistical cycle; Step S22. Using a neural network model to learn a training library consisting of all data linked lists under a single user ID; The video playback elements in the training library are used as input items and the user interaction data are used as output items. For each data linked list, the functional relationship between the output variable and the input variable is fitted. n-1 ), where x n-1 represents the n-1th input variable, and the neural network models include: R-CNN model, DeepLab model, U-Net model and Watershed model; Step S23. According to the selected model category, use elastic network regression to fit the loss function: Where LOSS is the loss function, SE is the root mean square error of the model, Y is the regularization strength of the model, k is the number of data lists in the training library, wi represents the preset weight of the i-th data list in the training library, r is the proportion of the regularization term, Fi(x1,x2,…,x n-1 ) represents the functional relationship between the output variable and the input variable in the i-th data linked list, represents the function Fi(x1,x2,…,x n-1 )’s linear regression function; Step S24. Calculate the regularization strength Y and the regularization term ratio r when the loss function LOSS takes the minimum value, and generate a positive definite matrix P according to the values ​​of Y and r, so that G(x1, x2, ..., x n-1 )=P·[x1,x2,…,x n-1 ] T , where T is the transpose symbol, G(x1,x2,…,x n-1 ) is the joint influence function.

4. The cross-platform video promotion method based on artificial intelligence according to claim 3 is characterized in that: Step S3 includes: Step S31. Arrange and combine the playback elements of the video to obtain h video types, substitute the playback elements corresponding to each video type into the joint influence function, and the obtained function values ​​constitute a prediction set. Classify the joint influence function into the video type corresponding to the maximum value in the prediction set; Step S32. Classify the joint influence function corresponding to each user ID, and hide the user ID in the classification result, and only keep the function formula; Step S4 includes: Step S41. Pack all joint influence functions contained in each video type and upload them to the cloud. The cloud obtains the data packets sent by each platform and classifies and summarizes the data packets using the video type as a keyword. Step S42. When the user logs in to the second platform for the first time, the second platform downloads all data packets from the cloud and stores them in isolation in the local server according to the video type.

5. The cross-platform video promotion method based on artificial intelligence according to claim 4 is characterized in that: Step S5 includes: Step S51. The second platform generates a typical function of each video type in the local server through a consistency equation: Among them, B(x1,x2,…,x n-1 ) represents the typical function of the video type, v represents the number of joint influence functions in the video type, Gj(x1,x2,…,x n-1 ) represents the jth joint influence function in the video type; Solve the consistency equation to obtain the typical function corresponding to each video type; Step S52. Obtain the video playback elements of the video that the user has played in the second platform, substitute the video playback elements into the typical functions of each video type, and record the obtained values ​​as the value set {A1, A2, ..., At, ..., Ah}, where At represents the function value of the typical function in the t-th video type; Step S53. Obtain the actual interaction data A0 of the user for the played video, calculate the difference AS between each element in the value set and the actual interaction data, where AS = |At-A0|, and is the cumulative score UR of the video type corresponding to the element, where UR = 1 / AS; Step S54: extract videos from the video library corresponding to each video type according to the accumulated scores, generate video streams, and push the video streams to the user.

6. A cross-platform video promotion system based on artificial intelligence, characterized in that: The system includes the following modules: platform data module, video association module, cross-platform transmission module, video recommendation module and advertisement insertion module; The platform data module is composed of a database of the video platform, platform software and a local server, and is used to record and store user information, video element information, interaction data and playback history, set a neural network model in the local server, and store the interaction information of a single user in a data pool of the server as a training library for the neural network model; The video association module is used to extract the user's historical playback data with the user's username as the primary key, and store the user's historical playback data in a training pool of a local server for training a neural network model for a single user's video playback behavior to obtain a loss model. The loss model takes video elements as input and user interaction data as output, fits the association model between the interaction data and each video element, and generates a joint influence function between the video playback time and each video element; The cross-platform transmission module is used to cut the joint influence function according to the preset definition domain interval of each video element, classify the user group according to the cutting result, and for each type of user group, package the joint influence function of all users in the group and upload it to the cloud. The cloud classifies and summarizes the data packets from each platform with the group label as the primary key, and receives the cross-platform data packet call; The video recommendation module is used to parse the cloud data packet on the second platform, perform consistency fitting on the joint influence function in each type of group label data, and generate a typical influence function corresponding to the group label. Each time the user plays a video on the second platform, the video elements are input into each typical function, and a predicted sequence of user interaction data is output. The distance between the actual interaction data and each element of the sequence is used as the cumulative playback coefficient of the group label corresponding to the element; The advertisement insertion module is used to extract videos from the video library under each video element classification according to the playback coefficient, and use the video playback time as a constraint condition to calculate the video arrangement when the length of each video advertisement is the longest, generate a video stream according to the video arrangement, and make automatic recommendations to users.

7. The cross-platform video promotion system based on artificial intelligence according to claim 6, characterized in that: The platform data module includes: a database unit and an artificial intelligence unit; The database unit is used to store user information and video playback elements, the video playback elements include: video tags, duration, creator, picture quality and advertisement length, and interactive data include: click-through rate and video playback duration; The artificial intelligence unit is set in a local database to analyze video playback data, and the neural network models used include: R-CNN model, DeepLab model, U-Net model and Watershed model.

8. The cross-platform video promotion system based on artificial intelligence according to claim 7, characterized in that: The video association module includes: a video classification unit, a loss model unit and a joint impact unit; The video classification unit is used to extract the user's historical playback data and classify the historical playback data according to video elements; The loss model unit is used to train a neural network model to fit the influence relationship of each video element on user interaction data; The joint influence unit is used to express the fitting result in a functional form, and to generate a joint influence function with the interaction data as the dependent variable and each video element as the independent variable.

9. The cross-platform video promotion system based on artificial intelligence according to claim 8, characterized in that: The cross-platform transmission module includes: a group preference unit and a cloud communication unit; The group preference unit is used to divide user groups according to the range of video elements and compress the joint influence functions corresponding to the groups into data packets; The cloud communication unit is used to classify and aggregate data packets from various platforms according to group labels, and is responsible for uploading and downloading data packets to the cloud; The video recommendation module includes: a user identification unit, a group analysis unit and a play accumulation unit; The user identification unit is used to fit the joint influence function in each data packet to generate a typical influence function, wherein the typical influence function represents the average preference level of the group label; The group analysis unit is used to obtain the user's operation behavior on the second platform and determine the user's group label; The playback accumulation unit is used to score the group label to which the user belongs according to the difference between the user's actual interaction data and the interaction data predicted by each typical influence function.

10. The cross-platform video promotion system based on artificial intelligence according to claim 9, characterized in that: The advertisement insertion module includes: a video extraction unit and a constraint conversion unit; The video extraction unit is used to partition and store the video according to the intervals where the video elements are located, and extract videos from each partition according to the user's group label; The constraint conversion unit is used to control the arrangement of the extracted videos using the video playback duration as a constraint condition so as to maximize the sum of the video advertisement lengths.

Images