Method, device, equipment and storage medium for recommending videos
By evaluating the degree of user willingness to pay for paid videos, using ordering tendency information and estimate models to recommend videos, the problem of inaccurate video recommendation in the prior art is solved and the user ordering rate is improved.
Patent Information
- Application Number
- CN202011282308.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-11-16
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2040-11-16
AI Technical Summary
The existing video recommendation method is not accurate enough, resulting in low ordering rate for users to recommend videos.
By obtaining the user's viewing behavior data and video details data, the user's willingness to pay for paid videos is evaluated, and the user's ordering intention information is determined using the ordering tendency information and estimate models, and then recommending videos that may be of interest.
Improves the user's order rate for recommended videos and enhances the accuracy of recommended videos.
Smart Images

Figure CN114513700B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of recommended videos, and in particular, relates to a method, device, equipment and storage medium for recommending videos. Background Art
[0002] With the development of technology, users have more and more ways to watch videos, and watching videos through Internet TV is one of the common ways for users.
[0003] The amount of video content is growing explosively. To attract users to watch, most personalized recommendations are used to recommend videos that users may be interested in. In the personalized recommendation method, most of them first calculate the similarity between videos, and then make recommendations based on the historical viewing behavior of the target user and the similarity of the videos. In this way, videos that users are interested in are found.
[0004] However, the current method of recommending videos to users is not accurate enough, resulting in a low subscription rate for recommended videos by users. Summary of the invention
[0005] The embodiments of the present invention provide a method, apparatus, device and storage medium for recommending videos, which can evaluate the user's willingness to pay for paid videos, and then recommend videos that may be of interest to the user based on the user's willingness to pay for paid videos, thereby increasing the user's subscription rate to the recommended videos.
[0006] In a first aspect, an embodiment of the present invention provides a method for recommending videos, the method comprising:
[0007] Obtaining the user's video viewing behavior data and video detail data, wherein the behavior data includes the time data when the user starts playing the video, the time data when the user ends playing the video, and the user's ordering data;
[0008] Determining ordering tendency information based on the behavior data and the feedback model, wherein the ordering tendency information represents the possibility of the user ordering the video;
[0009] The ordering intention information is determined according to the ordering tendency information, the behavior data, the video detail data and the estimation model, wherein the ordering intention information represents the user's ordering intention degree for the video.
[0010] In a possible implementation, the method further includes:
[0011] Preprocessing the behavior data and the video detail data to obtain preprocessed behavior data and video detail data;
[0012] The preprocessing includes at least one of the following:
[0013] Determine the viewing time of the same video by the same user based on the time data of when the user starts playing the video and the time data of when the user ends playing the video;
[0014] Delete duplicate data in video details data;
[0015] Determine the total duration of each video based on the video detail data;
[0016] A user who has completed an ordering behavior among the at least one user is determined according to the user ordering data of the at least one user.
[0017] In a possible implementation, determining order tendency information based on behavior data and a feedback model includes:
[0018] Determining first ordering tendency information according to the time data when the user starts playing the video and the time data when the user ends playing the video, wherein the first ordering tendency information represents the user's preference for the video;
[0019] Determining second ordering tendency information according to the user ordering data, wherein the second ordering tendency information represents the user's basic ordering willingness for the video;
[0020] The ordering tendency information is determined according to the first ordering tendency information and the second ordering tendency information.
[0021] In a possible implementation, before determining the ordering intention information according to the ordering tendency information, the behavior data, the video detail data and the estimation model, the method further includes:
[0022] The prediction model is trained based on ordering propensity information, behavioral data, and video detail data.
[0023] In a possible implementation, the prediction model is a click-through rate (CTR) prediction model;
[0024] The prediction model is trained according to the ordering tendency information, the behavior data and the video detail data, specifically including: the CTR prediction model is trained according to the ordering tendency information, the behavior data and the video detail data.
[0025] In a second aspect, an embodiment of the present invention provides a device for recommending videos, the device comprising:
[0026] An acquisition module, which acquires the behavior data of the user watching the video and the video details data, wherein the behavior data includes the time data when the user starts playing the video, the time data when the user ends playing the video, and the user's ordering data;
[0027] A first determination module is used to determine ordering tendency information according to the behavior data and the feedback model, wherein the ordering tendency information represents the possibility of the user ordering the video;
[0028] The second determination module is used to determine the ordering intention information according to the ordering tendency information, the behavior data, the video detail data and the estimation model, wherein the ordering intention information represents the user's ordering intention degree for the video.
[0029] In a possible implementation, the device further includes:
[0030] A preprocessing module preprocesses the behavior data and the video detail data to obtain the preprocessed behavior data and the video detail data;
[0031] The preprocessing includes at least one of the following:
[0032] Determine the viewing time of the same video by the same user based on the time data of when the user starts playing the video and the time data of when the user ends playing the video;
[0033] Delete duplicate data in video details data;
[0034] Determine the total duration of each video based on the video detail data;
[0035] According to the user ordering data of at least one user, a user with ordering behavior among the at least one user is determined.
[0036] In a possible implementation, the first determining module includes:
[0037] A first determining unit, configured to determine first ordering tendency information according to time data of when the user starts playing the video and time data of when the user ends playing the video, wherein the first ordering tendency information represents the user's preference for the video;
[0038] A second determining unit is used to determine second ordering tendency information according to the user ordering data, wherein the second ordering tendency information represents the user's basic ordering willingness for the video;
[0039] The third determining unit is used to determine the ordering tendency information according to the first ordering tendency information and the second ordering tendency information.
[0040] In a third aspect, an embodiment of the present invention provides a device for recommending videos, the device comprising: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, it implements the method in the first aspect or any possible implementation of the first aspect.
[0041] In a fourth aspect, an embodiment of the present invention provides a computer storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the method in the first aspect or any possible implementation of the first aspect.
[0042] The technical solution provided by the embodiments of the present invention brings at least the following beneficial effects:
[0043] The present invention obtains the behavior data and video details of users watching videos, determines the subscription tendency information according to the behavior data and feedback model of users watching videos, and determines the subscription intention information according to the subscription tendency information, behavior data, video details data of users watching videos and the estimation model. It can evaluate the degree of willingness of users to pay for paid videos, and then recommend videos that may be of interest to users according to the degree of willingness of users to pay for paid videos, thereby increasing the subscription rate of users to the recommended videos. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the following is a brief introduction to the drawings required for use in the embodiment of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0045] Figure 1 is a flow chart of a method for recommending videos provided by an embodiment of the present invention;
[0046] Figure 2 is a schematic diagram of another method for recommending videos provided by an embodiment of the present invention;
[0047] Figure 3 is a flow chart of an algorithm model for building video recommendation provided by an embodiment of the present invention;
[0048] Figure 4 is a schematic diagram of the structure of a device for recommending videos provided by an embodiment of the present invention;
[0049] Figure 5 It is a schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0050] The features and exemplary embodiments of various aspects of the present invention will be described in detail below. In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only configured to explain the present invention and are not configured to limit the present invention. For those skilled in the art, the present invention can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present invention by illustrating examples of the present invention.
[0051] 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. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0052] With the development of technology, users have more and more ways to watch videos, and watching videos through Internet TV is one of the common ways for users.
[0053] The amount of video content is growing explosively. To attract users to watch, most personalized recommendations are used to recommend videos that users may be interested in. In the personalized recommendation method, most of them first calculate the similarity between videos, and then make recommendations based on the historical viewing behavior of the target user and the similarity of the videos. In this way, videos that users are interested in are found.
[0054] However, the current method of recommending videos to users is not accurate enough, resulting in a low subscription rate for recommended videos by users.
[0055] The embodiments of the present invention provide a method, apparatus, device and storage medium for recommending videos, which can evaluate the user's willingness to pay for paid videos, and then recommend videos that may be of interest to the user based on the user's willingness to pay for paid videos, thereby increasing the user's subscription rate to the recommended videos.
[0056] In order to make the description of the present invention easier to understand, the terms in the embodiments of the present invention are explained below.
[0057] The following will be combined Figure 1 A method for recommending videos provided by an embodiment of the present invention is described in detail.
[0058] like Figure 1 As shown, the method may include the following steps:
[0059] S110, obtaining the user's video viewing behavior data and video detail data.
[0060] The behavior data includes the time data when the user starts playing the video, the time data when the user ends playing the video, and the user's ordering data.
[0061] The collection of user behavior data is carried out continuously within a preset operation cycle. In one example, the preset operation cycle may be 2 weeks.
[0062] The time data when the user starts playing the video may be the specific time when the user starts playing the video (eg, Beijing time when the user starts playing the video).
[0063] The time data at which the user finishes playing the video may be the specific time at which the user finishes playing the video.
[0064] User ordering data may include user payment operation data for paid programs. In one example, user ordering data may include: for paid videos, the user enters the ordering page, the corresponding value is 1; the user enters the product package selection page, the corresponding value is 2; the user enters the payment page, the corresponding value is 3; the user fails to order, the corresponding value is 4; the user successfully orders, the corresponding value is 5.
[0065] The video detail data may include one or more of the following data: video address (identity, ID), identity (identity, ID) data of the user watching the video, video name data, total video length data, video type data, plot category data, video director data, video actor data, video distribution region data, video distribution year data, video language data, whether it is a paid video data, etc.
[0066] In one example, an application program may be preinstalled in an Internet TV set-top box, and the application program may be used to collect the original viewing behavior data of the user and send the collected data to the platform server, so that the platform server can analyze the original viewing behavior data of the user.
[0067] In one implementation provided by the present invention, the behavior data and the video detail data are preprocessed to obtain the preprocessed behavior data and the video detail data;
[0068] The preprocessing includes at least one of the following:
[0069] The viewing time data of the same video by the same user is determined based on the time data when the user starts playing the video and the time data when the user ends playing the video.
[0070] In one example, according to the time data of each video playback record of the same user, the duration of each viewing is added up to calculate the total viewing time of the video by the user. According to the playback time data of each user for each video, the duration of each viewing is added up to calculate the viewing time of each user for each video.
[0071] Delete duplicate data in video details.
[0072] In one example, the platform server detects duplicate data in the video detail data, deletes the duplicate data, and aggregates non-duplicate data into a complete database.
[0073] Determine the total duration of each video based on the video detail data.
[0074] In one example, the total duration of each video is determined based on the total duration data of the video in the video detail data of each video.
[0075] A user who has completed an ordering behavior among the at least one user is determined according to the user ordering data of the at least one user.
[0076] The user ordering data may include the user's payment operation data for the paid program, and may include: the user enters the product package selection page, the user enters the payment page, the user order fails, the user order succeeds, and other data. When the user order succeeds, it indicates that the ordering behavior is completed. According to the user ordering data of at least one user, the user who has successfully ordered is determined as the user who has completed the ordering behavior.
[0077] The method provided by the embodiment of the present invention can enable the platform server to obtain the original data of the user's video viewing and pre-process the obtained data, so that the platform server can analyze the user's preference for the video based on the pre-processed data.
[0078] S120: Determine order tendency information according to the behavior data and the feedback model.
[0079] The ordering tendency information represents the possibility of a user ordering a video.
[0080] In one implementation provided by the present invention, determining order tendency information according to the behavior data and the feedback model includes:
[0081] The first ordering tendency information is determined according to the time data when the user starts playing the video and the time data when the user ends playing the video, wherein the first ordering tendency information represents the user's preference for the video.
[0082] The second ordering tendency information is determined according to the user ordering data, wherein the second ordering tendency information represents the user's basic ordering willingness for the video.
[0083] The ordering tendency information is determined according to the first ordering tendency information and the second ordering tendency information.
[0084] In one example, for a user watching a video, the viewing time of the user is determined based on the time data when the user starts playing the video and the time data when the user ends playing the video. The viewing time of the user is divided by the total duration of the program, and the result is named the user viewing score, recorded as Sx. The above operation is performed on the behavior data of N users (N is an integer greater than 0) who watch the program to obtain N user viewing score values Sx. The largest value is recorded as Smax, and the smallest value is recorded as Smin. The calculation method of the first ordering tendency information is shown in formula (1).
[0085]
[0086] Wherein, F1 represents the first ordering tendency information. Wherein, Smax is the maximum value of all users' viewing scores. Smin is the minimum value of all users' viewing scores. Ceil is the upward rounding function.
[0087] In one embodiment provided by the present invention, for a paid video, when a user enters the ordering page, it can be characterized that the user has a basic ordering intention for the paid video. In one example, the user ordering data includes: for a paid video, the user enters the ordering page, and the corresponding second ordering tendency information is 1; the user enters the product package selection page, and the corresponding second ordering tendency information is 2; the user enters the payment page, and the corresponding second ordering tendency information is 3; the user fails to order, and the corresponding second ordering tendency information is 4; the user successfully orders, and the corresponding second ordering tendency information is 5. The larger the value of the second ordering tendency information, the greater the user's basic ordering intention for the video.
[0088] In one embodiment provided by the present invention, the feedback model is shown in formula (2).
[0089] F=Ceil(F1×M+F2×N) (2)
[0090] Wherein, F represents order tendency information. F1 represents first order tendency information. F2 represents second order tendency information. M represents the weight of F1, and N represents the weight of F2. M and N are values preset according to demand. Ceil represents a rounding up function.
[0091] The embodiment of the present invention determines the first ordering tendency information according to the time data of the user starting to play the video and the time data of the user ending to play the video, determines the second ordering tendency information according to the user ordering data, and determines the ordering tendency information according to the first ordering tendency information and the second ordering tendency information. The user's willingness to order the video can be obtained, and then the video that may be of interest to the user can be recommended according to the willingness to order.
[0092] S130, determining order intention information according to the order tendency information, the behavior data, the video detail data and the estimation model.
[0093] In one embodiment of the present invention, before determining the order intention information according to the order tendency information, the behavior data, the video detail data and the estimation model, the method further includes:
[0094] The prediction model is trained based on ordering propensity information, behavioral data, and video detail data.
[0095] First, all the acquired user ordering tendency information and behavior data for each paid video are randomly shuffled. In one example, if the user data is less than 50,000, the data is divided into a training set and a validation set in a ratio of 8:2; if the user data exceeds 50,000, 10,000 of them are extracted as a validation set, and the remaining data is used as a training set. Then set the hyperparameters. The hyperparameters that need to be set in the estimation model of the present invention are one or more of the learning rate, regularization parameter, and number of latent factors. The values of the hyperparameters are changed according to demand during the training process. Reinitialize the training parameters. The training parameters required in the model estimation model of the present invention include one or more of the bias constant, first-order coefficients, and second-order cross-term matrix weights. In one example, the bias constant is initialized to 0, and the first-order coefficients and second-order cross-term matrix weights are randomly initialized using a normal distribution with a mean of 0 and a standard deviation of 1. Finally, the model is iterated by the gradient descent method to obtain a trained estimation model.
[0096] After obtaining the trained estimation model, the test set samples are input into the model to obtain the model's prediction results for the test set.
[0097] In one embodiment provided by the present invention, the model effect is evaluated by a model evaluation index (Area under the curve, AUC). When the model effect does not reach the expected effect, the parameters in the hyperparameters or training parameters are adjusted, and the model training is performed again until the estimated model obtained can achieve the expected effect.
[0098] The method provided in the embodiment of the present invention can obtain an estimation model that meets the expected effect by training the estimation model, so as to determine the ordering intention information.
[0099] In one example, the prediction model is a click-through rate (CTR) prediction model.
[0100] The prediction model is trained according to the ordering tendency information, the behavior data and the video detail data, specifically including: the CTR prediction model is trained according to the ordering tendency information, the behavior data and the video detail data.
[0101] The method provided in the embodiment of the present invention can obtain a CTR prediction model that meets the expected effect by training the CTR prediction model, so as to be used for determining ordering intention information.
[0102] In an embodiment provided by the present invention, order intention information is determined based on order tendency information, behavior data, video detail data and an estimation model, wherein the order intention information represents the user's willingness to order the video.
[0103] The method provided by the embodiment of the present invention can determine the subscription intention information based on the subscription tendency information, behavior data, video details data of the videos watched by the user and the estimation model. It can evaluate the user's willingness to pay for the paid videos, and then recommend videos that may be of interest to the user based on the user's willingness to pay for the paid videos, thereby increasing the user's subscription rate for the recommended videos.
[0104] In one example, the method for recommending videos provided by the present invention is as follows: Figure 2 As shown. The user terminal may be a television and a set-top box 210. The data collection and storage layer may include a set-top box soft probe 220 and a data warehouse 230. The set-top box soft probe 220 is an application pre-installed in the television set-top box. The application may be used to collect the behavior data and video detail data of the user watching the video at the user terminal, and send the collected data to the data warehouse 230 of the platform server. Then, based on the behavior data and video detail data in the data warehouse 230, the feedback model 240 and the CTR prediction model 250, a video recommendation result 260 is generated. The recommendation result 260 is a video that the user may be interested in. The recommendation result 260 is sent to the television and the set-top box 210 to the user terminal through the recommendation system communication link 270.
[0105] The method provided by the embodiment of the present invention can determine the subscription intention information based on the subscription tendency information, behavior data, video details data of the videos watched by the user and the estimation model. It can evaluate the user's willingness to pay for the paid videos, and then recommend videos that may be of interest to the user based on the user's willingness to pay for the paid videos, thereby increasing the user's subscription rate for the recommended videos.
[0106] like Figure 2 The method for recommending videos shown is one of the implementation methods of the embodiment of the present invention, and can also be implemented by other methods, which will not be described in detail here.
[0107] In one embodiment of the present invention, the process of building a video recommendation model is as follows: Figure 3 shown.
[0108] 310, collect data.
[0109] Data can be collected based on the set-top box soft probe. The collected data may include behavior data and video detail data. The set-top box soft probe is an application pre-installed in the TV set-top box. The application can be used to collect the user's viewing behavior raw data and send the collected data to the platform server.
[0110] 320, data preprocessing.
[0111] The collected behavior data and video detail data are preprocessed according to preset preprocessing rules to obtain preprocessed behavior data and video detail data.
[0112] 330, building a feedback model.
[0113] The feedback model can determine the first ordering tendency information according to the time data when the user starts playing the video and the time data when the user ends playing the video, wherein the first ordering tendency information represents the user's preference for the video.
[0114] The feedback model can also determine second ordering tendency information based on the user ordering data, wherein the second ordering tendency information represents the user's basic ordering willingness for the video.
[0115] The feedback model may also determine the ordering tendency information based on the first ordering tendency information and the second ordering tendency information.
[0116] 340, build a CTR prediction model.
[0117] The CTR prediction model in the embodiment of the present invention can generate video recommendation results through ordering tendency information, behavior data and video detail data.
[0118] 350, training CTR estimation model.
[0119] The prediction model is trained based on ordering propensity information, behavioral data, and video detail data.
[0120] In one example, the ordering tendency information and behavior data of all users for each paid video are first randomly shuffled. In one example, if the user data is less than 50,000, the data is divided into a training set and a validation set in a ratio of 8:2; if the user data exceeds 50,000, 10,000 of them are extracted as a validation set, and the remaining data is used as a training set. Then set the hyperparameters. The hyperparameters that need to be set in the estimation model of the present invention are one or more of the learning rate, regularization parameter, and number of latent factors. The values of the hyperparameters are changed according to demand during the training process. Reinitialize the training parameters. The training parameters required in the model estimation model of the present invention include one or more of the bias constant, first-order coefficients, and second-order cross-term matrix weights. In one example, the bias constant is initialized to 0, and the first-order coefficients and second-order cross-term matrix weights are randomly initialized using a normal distribution with a mean of 0 and a standard deviation of 1. Finally, the model is iterated by the gradient descent method to obtain a trained estimation model.
[0121] 360, verify the effect of CTR estimation model.
[0122] After obtaining the trained CTR prediction model, the test set samples are input into the model to obtain the model's prediction results for the test set.
[0123] In one example, the model effect is evaluated by a model evaluation indicator (Area under the curve, AUC).
[0124] 370, to determine whether the effect is up to standard.
[0125] When the model effect does not reach the expected effect, step 380 is executed.
[0126] When the model effect reaches the expected effect, the trained CTR estimation model is obtained.
[0127] 380, adjust parameters.
[0128] Adjust the parameters in the hyperparameters or training parameters and train the model again until the estimated model can achieve the expected results.
[0129] The method provided in the embodiment of the present invention can construct a feedback model and a CTR prediction model, and obtain a CTR prediction model that meets the expected effect for determining ordering intention information.
[0130] like Figure 3 The process of constructing the algorithm model for video recommendation shown is one of the implementation methods of the embodiment of the present invention, and can also be implemented by other methods, which will not be described in detail here.
[0131] Figure 44 is a schematic diagram of a video recommendation device structure provided by an embodiment of the present invention. The video recommendation device 400 may include an acquisition module 410, a first determination module 420, and a second determination module 430. Among them:
[0132] The acquisition module 410 is used to acquire the user's video viewing behavior data and video detail data, wherein the behavior data includes the time data when the user starts playing the video, the time data when the user ends playing the video, and the user's subscription data.
[0133] The collection of user behavior data is carried out continuously within a preset operation cycle. In one example, the preset operation cycle may be 2 weeks.
[0134] The time data when the user starts playing the video may be the specific time when the user starts playing the video (eg, Beijing time when the user starts playing the video).
[0135] The time data at which the user finishes playing the video may be the specific time at which the user finishes playing the video.
[0136] User ordering data may include user payment operation data for paid programs. In one example, user ordering data may include: for paid videos, the user enters the ordering page, the corresponding value is 1; the user enters the product package selection page, the corresponding value is 2; the user enters the payment page, the corresponding value is 3; the user fails to order, the corresponding value is 4; the user successfully orders, the corresponding value is 5.
[0137] The video detail data may include one or more of the following data: video address (identity, ID), identity (identity, ID) data of the user watching the video, video name data, total video length data, video type data, plot category data, video director data, video actor data, video distribution region data, video distribution year data, video language data, whether it is a paid video data, etc.
[0138] The video recommendation device provided in the embodiment of the present invention can obtain the user's video viewing behavior data and video detail data, so as to be used by the device to analyze the user's viewing behavior raw data.
[0139] The first determination module 420 is configured to determine ordering tendency information according to the behavior data and the feedback model, wherein the ordering tendency information represents the possibility of the user ordering the video.
[0140] In an embodiment provided by the present invention, the first determination module 420 may include: a first determination unit, a second determination unit and a third determination unit.
[0141] A first determining unit, configured to determine first ordering tendency information according to time data when the user starts playing the video and time data when the user ends playing the video, wherein the first ordering tendency information represents the user's preference for the video;
[0142] A second determining unit is used to determine second ordering tendency information according to the user ordering data, wherein the second ordering tendency information represents the user's basic ordering willingness for the video;
[0143] The third determining unit is configured to determine the ordering tendency information according to the first ordering tendency information and the second ordering tendency information.
[0144] The video recommendation device provided by the embodiment of the present invention can determine the first ordering tendency information according to the time data of the user starting to play the video and the time data of the user ending to play the video, determine the second ordering tendency information according to the user ordering data, and determine the ordering tendency information according to the first ordering tendency information and the second ordering tendency information. The user's willingness to order the video can be obtained, and then the video that may be of interest to the user can be recommended according to the willingness to order, thereby improving the accuracy of recommending videos to the user.
[0145] The second determination module 430 is used to determine the ordering intention information according to the ordering tendency information, the behavior data, the video detail data and the estimation model, wherein the ordering intention information represents the user's ordering intention degree for the video.
[0146] In one embodiment provided by the present invention, the apparatus for recommending videos further includes a training module 440 for training a prediction model based on ordering tendency information, behavior data, and video detail data.
[0147] The device provided in the embodiment of the present invention can obtain an estimation model that meets the expected effect by training the estimation model, so as to determine the ordering intention information.
[0148] In one example, the prediction model is a click-through rate (CTR) prediction model.
[0149] The prediction model is trained according to the ordering tendency information, the behavior data and the video detail data, specifically including: the CTR prediction model is trained according to the ordering tendency information, the behavior data and the video detail data.
[0150] The method provided in the embodiment of the present invention can obtain a CTR prediction model that meets the expected effect by training the CTR prediction model, so as to be used for determining ordering intention information.
[0151] The device provided by the embodiment of the present invention can determine the subscription intention information based on the subscription tendency information, behavior data, video details data of the videos watched by the user and the estimation model. It can evaluate the user's willingness to pay for the paid videos, and then recommend videos that may be of interest to the user based on the user's willingness to pay for the paid videos, thereby increasing the user's subscription rate for the recommended videos.
[0152] The video recommendation device provided in the embodiment of the present invention may further include a preprocessing module 450, which is used to preprocess the behavior data and the video detail data to obtain preprocessed behavior data and video detail data;
[0153] In one embodiment provided by the present invention, the pre-processing module 450 may include at least one of the following: a third determination module, a deletion module, a fourth determination module, and a fifth determination module. Among them:
[0154] A third determination module is used to determine the viewing time data of the same user for the same video according to the time data when the user starts playing the video and the time data when the user ends playing the video;
[0155] A deletion module, used to delete duplicate data in the video detail data;
[0156] A fourth determination module, configured to determine a total duration of each of the videos according to the video detail data;
[0157] The fifth determining module is used to determine a user who has placed an order among the at least one user according to the user ordering data of the at least one user.
[0158] The device provided by the embodiment of the present invention can pre-process the acquired data to analyze the user's preference for the video.
[0159] The video recommendation device provided by the embodiment of the present invention can determine the subscription tendency information based on the user's video viewing behavior data and feedback model, and determine the subscription intention information based on the subscription tendency information, behavior data, video details data of the user's video viewing, and the estimation model. It can evaluate the user's willingness to pay for paid videos, and then recommend videos that may be of interest to the user based on the user's willingness to pay for paid videos, thereby increasing the user's subscription rate for the recommended videos.
[0160] The device for recommending videos provided by the embodiment of the present invention performs Figure 1 Each step in the method shown can achieve the technical effect of recommending videos that may be of interest to users based on the user's willingness to pay for paid videos, thereby increasing the user's subscription rate to the recommended videos. For the sake of brevity, it will not be described in detail here.
[0161] Figure 5 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention is shown.
[0162] The electronic device may include a processor 501 and a memory 302 storing computer program instructions.
[0163] Specifically, the processor 501 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiment of the present invention.
[0164] The memory 502 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 502 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 502 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 502 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 502 is a non-volatile solid-state memory. In a specific embodiment, the memory 502 includes a read-only memory (ROM). In appropriate cases, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM) or a flash memory or a combination of two or more of these.
[0165] The processor 501 reads and executes the computer program instructions stored in the memory 502 to implement Figure 1 Any one of the methods for recommending videos in the illustrated embodiments.
[0166] In one example, the electronic device may further include a communication interface 503 and a bus 510. Figure 5 As shown, the processor 501, the memory 502, and the communication interface 503 are connected via a bus 510 and communicate with each other.
[0167] The communication interface 503 is mainly used to implement the communication between the modules, devices, units and / or equipment in the embodiment of the present invention.
[0168] Bus 510 includes hardware, software or both, and the parts of electronic equipment are coupled to each other. For example, but not limitation, bus may include accelerated graphics port (AGP) or other graphics bus, enhanced industrial standard architecture (EISA) bus, front side bus (FSB), hypertransport (HT) interconnection, industrial standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 510 may include one or more buses. Although the embodiment of the present invention describes and shows a specific bus, the present invention considers any suitable bus or interconnection.
[0169] The electronic device can execute the method for recommending videos in the embodiment of the present invention, thereby realizing the combination Figure 1 Describes the method of recommending videos.
[0170] In addition, in combination with the method for recommending videos in the above embodiments, the present invention can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when the computer program instructions are executed by a processor, any of the methods for recommending videos in the above embodiments is implemented.
[0171] It should be clear that the present invention is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present invention.
[0172] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.
[0173] It should also be noted that the exemplary embodiments mentioned in the present invention describe some methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiments, or in a different order from the embodiments, or several steps can be performed simultaneously.
[0174] The above is only a specific implementation of the present invention. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the system, module and unit described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be covered within the protection scope of the present invention.
Claims
1. A method for recommending videos, It is characterized in that The method comprises: Obtaining user video viewing behavior data and video detail data, wherein the behavior data includes the time data when the user starts playing the video, the time data when the user ends playing the video, and user subscription data, wherein the user subscription data includes the user's payment operation data for paid programs; Determining ordering tendency information according to the behavior data and the feedback model, wherein the ordering tendency information represents the possibility of the user ordering the video; Determining order intention information according to the order tendency information, the behavior data, the video detail data and the estimation model, wherein the order intention information represents the user's order intention degree for the video; Determining the order tendency information according to the behavior data and the feedback model includes: Determine first ordering tendency information according to the time data when the user starts playing the video and the time data when the user ends playing the video, wherein the first ordering tendency information represents the user's preference for the video; Determining second ordering tendency information according to the user ordering data, wherein the second ordering tendency information represents the user's basic ordering willingness for the video; The ordering tendency information is determined according to the first ordering tendency information and the second ordering tendency information.
2. The method according to claim 1, It is characterized in that The method further comprises: Preprocessing the behavior data and the video detail data to obtain preprocessed behavior data and video detail data; Wherein, the preprocessing includes at least one of the following: Determine the viewing time data of the same video by the same user according to the time data when the user starts playing the video and the time data when the user ends playing the video; Deleting duplicate data in the video details data; Determine the total duration of each video according to the video detail data; A user who has completed an ordering behavior among the at least one user is determined according to the user ordering data of the at least one user.
3. The method according to claim 1, It is characterized in that Before determining the ordering intention information according to the ordering tendency information, the behavior data, the video detail data and the estimation model, the method further includes: The prediction model is trained based on the ordering tendency information, the behavior data and the video detail data.
4. The method according to claim 1 or 3, It is characterized in that The prediction model is a click-through rate (CTR) prediction model; The training of the prediction model according to the ordering tendency information, the behavior data and the video detail data specifically includes: training the CTR prediction model according to the ordering tendency information, the behavior data and the video detail data.
5. A device for video recommendation, It is characterized in that The device comprises: An acquisition module, used to acquire the behavior data of the user watching the video and the video details data, wherein the behavior data includes the time data of the user starting to play the video, the time data of the user ending to play the video, and the user subscription data, wherein the user subscription data includes the payment operation data of the user for the paid program; A first determination module is used to determine ordering tendency information according to the behavior data and the feedback model, wherein the ordering tendency information represents the possibility of the user ordering the video; A second determination module is used to determine ordering intention information according to the ordering tendency information, the behavior data, the video detail data and the estimation model, wherein the ordering intention information represents the user's ordering intention degree for the video; The first determining module includes: A first determining unit, configured to determine first ordering tendency information according to time data of when the user starts playing the video and time data of when the user ends playing the video, wherein the first ordering tendency information represents the user's preference for the video; A second determining unit is used to determine second ordering tendency information according to the user ordering data, wherein the second ordering tendency information represents the user's basic ordering willingness for the video; The third determining unit is configured to determine the ordering tendency information according to the first ordering tendency information and the second ordering tendency information.
6. The device according to claim 5, It is characterized in that The device also includes: A preprocessing module, used for preprocessing the behavior data and the video detail data to obtain preprocessed behavior data and video detail data; Wherein, the preprocessing includes at least one of the following: Determine the viewing time data of the same video by the same user according to the time data when the user starts playing the video and the time data when the user ends playing the video; Deleting duplicate data in the video details data; Determine the total duration of each video according to the video detail data; According to the user ordering data of at least one user, a user with ordering behavior among the at least one user is determined.
7. An electronic device, It is characterized in that The device comprises: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the method for recommending videos according to any one of claims 1 to 4 is implemented.
8. A computer storage medium, It is characterized in that The computer storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the method for recommending videos according to any one of claims 1 to 4 is implemented.
