Multimedia resource watching duration prediction method and device, medium, equipment and product

Abstract

The invention relates to a multimedia resource watching duration prediction method, a multimedia resource watching duration prediction device, a medium, equipment and a product, and relates to the technical field of multimedia resource processing. The method comprises the following steps: acquiring corresponding watching probabilities of multimedia resources in a plurality of continuous duration intervals, and adjusting the watching probabilities to obtain adjusted watching probabilities, so that the watching probability of a larger duration interval is smaller than the watching probability corresponding to a last smaller duration interval, and the watching probability of the multimedia resources in the last smaller duration interval is smaller than the watching probability corresponding to the last smaller duration interval. And then, based on each duration interval and the adjusted watching probability corresponding to the duration interval, determining the predicted watching duration of the object for the multimedia resource, so that the limitation of implicit hypothesis data distribution of a duration prediction model can be reduced, and the scale and extreme value of the duration interval are more friendly; and the predicted watching duration of the multimedia resources obtained by the duration prediction model can be more accurate without manual intervention.

Description

Technical Field

[0001] The present disclosure relates to the technical field of multimedia resource processing, and in particular, to a method, device, medium, equipment and product for predicting viewing duration of multimedia resources. Background Art

[0002] In the information flow push system, it is often necessary to estimate the viewing time of different users for different content, and achieve the effect of accurate push by accurately modeling the user's viewing time.

[0003] In the related art, in the modeling of the duration model, regression methods such as mean square error and mean absolute percentage error are generally used for the duration target. However, these regression methods have implicit conditional assumptions, that is, the deviation between the output predicted by the duration model and the actual duration (label) conforms to the normal distribution. The greater the difference between the deviation distribution and the normal distribution, the worse the prediction ability of the duration model. Moreover, these regression methods are sensitive to the scale of the duration target and are not very friendly to the maximum and minimum values ​​in the duration target, resulting in low accuracy in the prediction of viewing duration. Summary of the invention

[0004] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0005] In a first aspect, the present disclosure provides a method for predicting viewing time of multimedia resources, comprising: Inputting the attribute information of the multimedia resource and object to be predicted into the trained duration prediction model to obtain the corresponding viewing probability of the multimedia resource in multiple continuous duration intervals; Adjusting each of the viewing probabilities to obtain an adjusted viewing probability, wherein the adjusted viewing probability makes the viewing probability of a larger duration interval smaller than the viewing probability corresponding to a previous smaller duration interval; Based on each of the duration intervals and the adjusted viewing probabilities corresponding to the duration intervals, a predicted viewing duration of the object for the multimedia resource to be predicted is determined.

[0006] In a second aspect, the present disclosure provides a multimedia resource viewing time prediction device, comprising: A first obtaining module is configured to input the attribute information of the multimedia resource and object to be predicted into the trained duration prediction model to obtain the corresponding viewing probability of the multimedia resource in multiple continuous duration intervals; A second obtaining module is configured to adjust each viewing probability to obtain an adjusted viewing probability, wherein the adjusted viewing probability makes the viewing probability of a larger duration interval smaller than the viewing probability corresponding to the previous smaller duration interval; The determination module is configured to determine the predicted viewing duration of the object for the multimedia resource to be predicted based on each of the duration intervals and the adjusted viewing probabilities corresponding to the duration intervals.

[0007] In a third aspect, the present disclosure provides a computer-readable medium having a computer program stored thereon, which, when executed by a processing device, implements the steps of the method described in the first aspect.

[0008] In a fourth aspect, the present disclosure provides an electronic device, including: a storage device having a computer program stored thereon; A processing device is used to execute the computer program in the storage device to implement the steps of the method described in the first aspect.

[0009] In a fifth aspect, the present disclosure provides a computer program product, including a computer program, which implements the steps of the method described in the first aspect when executed by a processor.

[0010] Based on the above technical scheme, by inputting the attribute information of the multimedia resources and objects to be predicted into the trained duration prediction model, the viewing probability corresponding to the multimedia resources in multiple continuous duration intervals is obtained, and each viewing probability is adjusted to obtain the adjusted viewing probability, so that the viewing probability of a larger duration interval is less than the viewing probability corresponding to the previous smaller duration interval, and then based on each duration interval and the adjusted viewing probability corresponding to the duration interval, the predicted viewing time of the object for the multimedia resource is determined, which can reduce the limitations of the implicit assumptions of the data distribution of the duration prediction model, and is more friendly to the scale and extreme values ​​of the duration interval, and the predicted viewing time of the multimedia resources obtained by the duration prediction model can be more accurate without human intervention.

[0011] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings: Figure 1 is a flowchart of a method for predicting viewing duration of multimedia resources according to some embodiments.

[0013] Figure 2 It is a schematic diagram showing the principle of predicting the viewing duration of multimedia resources according to some embodiments.

[0014] Figure 3 It is a schematic diagram of the structure of a multimedia resource viewing duration prediction device according to some embodiments.

[0015] Figure 4 is a schematic diagram of the structure of an electronic device according to some embodiments. DETAILED DESCRIPTION

[0016] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0017] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0018] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0019] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0020] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0021] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0022] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0023] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0024] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0025] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0026] At the same time, it is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.

[0027] Figure 1 FIG. 1 is a flowchart of a method for predicting the viewing time of a multimedia resource according to some embodiments. Figure 1 As shown, the embodiment of the present disclosure provides a method for predicting the viewing time of a multimedia resource, which can be executed by an electronic device, and specifically can be executed by a device for predicting the viewing time of a multimedia resource, which can be implemented by software and / or hardware and configured in an electronic device. Figure 1 As shown, the method may include the following steps.

[0028] In step 110, the attribute information of the multimedia resources and objects to be predicted is input into the trained duration prediction model to obtain the corresponding viewing probabilities of the multimedia resources in multiple continuous duration intervals.

[0029] Here, the multimedia resources may be video, audio, news information, etc. The multimedia resources to be predicted may be one or more multimedia resources in the database. The attribute information of the object may refer to the attribute information of the user, wherein the attribute information of the object may include the information of the user itself and the interaction information of the user for the multimedia resources, and the interaction information is used to reflect the preference degree of the user for different types of multimedia resources.

[0030] A target duration can be divided into several continuous, non-overlapping time intervals to obtain multiple continuous duration intervals. For example, if the target duration is 30 minutes, the target duration can be divided into 1800 continuous duration intervals, each of which has a duration of 1 second, that is, multiple continuous duration intervals can be (0, 1] seconds, (1, 2] seconds, (2, 3] seconds, (3, 4] seconds, ..., (1799, 1800] seconds.

[0031] The length of the duration interval determines the granularity of the prediction. Shorter duration intervals (such as 1 second) can provide more fine-grained predictions, while longer duration intervals (such as 1 minute) can provide coarser-grained predictions. Of course, the specific length of the duration interval can be determined according to the needs of the actual application.

[0032] It is worth noting that the target duration may be independent of the duration of the multimedia resource. The target duration may be the maximum duration that the subject may spend when viewing a multimedia resource, which is set according to the requirements. For example, the target duration may be set to half an hour, which means that it is assumed that the maximum duration that the subject may spend when viewing a multimedia resource is half an hour. For another example, for a 5-minute video, the maximum duration that the subject may spend watching the video may be 30 minutes.

[0033] The duration prediction model is used to predict the viewing probability of multimedia resources in multiple consecutive duration intervals. The viewing probability corresponding to each duration interval represents the possibility of the object watching the multimedia resource in the duration interval.

[0034] For each duration interval, the duration prediction model predicts whether the subject will stop watching the multimedia resource within the duration interval. For example, if the predicted duration interval is one second, the duration prediction model predicts whether the subject will stop watching the multimedia resource in the next second. Accordingly, the viewing probability of the multimedia resource under multiple consecutive duration intervals can be understood as the probability that the subject stops watching the multimedia resource within the duration interval.

[0035] For example, for a multimedia resource sample, if the actual viewing time of the multimedia resource sample is m (m≤N, N is the target time), the viewing probability of the multimedia resource sample in the first m time intervals is 1, and the viewing probability in the remaining time intervals is 0.

[0036] For example, the duration prediction model can be a pre-trained deep model. The number of units in the last fully connected network layer of the deep model can be set to N, and the fully connected network will output N viewing probabilities ( ). Where N is the number of consecutive duration intervals. Each viewing probability corresponds to a duration interval.

[0037] In step 120, each viewing probability is adjusted to obtain an adjusted viewing probability, wherein the adjusted viewing probability makes the viewing probability of the larger duration interval smaller than the viewing probability corresponding to the previous smaller duration interval.

[0038] Here, after obtaining the viewing probabilities corresponding to the respective duration intervals, the respective viewing probabilities are adjusted so that the viewing probability corresponding to the larger duration interval is smaller than the viewing probability corresponding to the previous smaller duration interval.

[0039] Among them, the larger duration interval is the duration interval on the larger side of the duration value among multiple continuous duration intervals, and the smaller duration interval is the duration interval on the smaller side of the duration value among multiple continuous duration intervals. For example, multiple continuous duration intervals are (0, 1] seconds, (1, 2] seconds, (2, 3] seconds, (3, 4] seconds, ..., (1799, 1800] seconds, then relative to the duration interval of (0, 1] seconds, (1, 2] seconds, (2, 3] seconds, (3, 4] seconds, ..., (1799, 1800] seconds are all larger duration intervals. In other words, the adjusted viewing probability can make the viewing probability corresponding to the latter duration interval less than the viewing probability corresponding to the previous duration interval. For example, the adjusted viewing probability of the duration interval of (2, 3] seconds is 1, the adjusted viewing probability of the duration interval of (3, 4] seconds can be 0.9, and the adjusted viewing probability of the duration interval of (4, 5] seconds can be 0.8.

[0040] It should be understood that when the subject watches multimedia resources, the viewing behavior of the multimedia resources has a certain time dependency, and the viewing probability of the next time interval depends on the behavior of the previous time interval. In other words, the subject's viewing of the multimedia resource at the n+1th second is based on the subject having watched the multimedia resource at the nth second. Only after the subject has watched the multimedia resource at the nth second will the subject have a certain probability of watching the multimedia resource at the n+1th second.

[0041] Therefore, by adjusting the viewing probability corresponding to each continuous duration interval of the multimedia resource output by the duration prediction model, so that the adjusted viewing probability can make the viewing probability corresponding to the larger duration interval smaller than the viewing probability corresponding to the previous smaller duration interval, the duration prediction model can ensure that the viewing probability of the previous second is greater than or equal to the viewing probability of the next second when performing duration modeling, that is, satisfying , where n is time and P is viewing probability. That is, by adjusting the viewing probability, the viewing probability of multiple consecutive time intervals can be order-preserving constraints.

[0042] Based on this, in the process of predicting the viewing time of multimedia resources, the time dependence of viewing behavior is reflected, thereby weakening the limitations of the implicit assumptions of data distribution in the duration prediction model, making the duration prediction model more friendly to duration intervals and extreme values.

[0043] It is worth noting that the duration prediction model actually divides a target duration into multiple continuous duration intervals for classification prediction. The general classification method does not measure the difference between the predicted probability and the actual probability of the duration interval. When a long duration sample is predicted as a short duration or a medium duration, if the predicted probabilities are consistent, the values ​​of the loss function are equivalent, and the duration model cannot learn the difference between the two. However, in the process of multimedia resource recommendation, the degree of deviation from the true value will have a greater impact on the recommendation effect. Even through classification methods such as ordinal regression, it is necessary to determine the duration interval and classification label in advance through the true duration distribution, and some need to determine the probability cut point through the true duration distribution, which is greatly affected by data distribution and manual influence. In the embodiment of the present disclosure, by adjusting the viewing probability so that the adjusted viewing probability makes the viewing probability of the larger duration interval less than the viewing probability corresponding to the previous smaller duration interval, the limitations of the assumed data distribution of the duration prediction model can be weakened, which is more friendly to the duration target scale and extreme value, and no manual intervention is required.

[0044] In step 130 , based on each duration interval and the adjusted viewing probability corresponding to the duration interval, the predicted viewing duration of the object for the multimedia resource to be predicted is determined.

[0045] Here, after obtaining the adjusted viewing probability corresponding to each duration interval, the predicted viewing duration of the object for the multimedia resource to be predicted is calculated based on the duration corresponding to each duration interval and the adjusted viewing probability.

[0046] The predicted viewing time of the object for the multimedia resource refers to the possible viewing time of the user for viewing the multimedia resource. The predicted viewing time can indicate the user's preference for the multimedia resource. The greater the predicted viewing time, the greater the user's preference for the multimedia resource.

[0047] It is worth noting that the preset viewing time can be used to recommend multimedia resources to the object. For example, multimedia resources whose predicted viewing time is greater than a preset time threshold can be recommended to the object.

[0048] Therefore, by inputting the attribute information of the multimedia resources and objects to be predicted into the trained duration prediction model, the viewing probability corresponding to the multimedia resources in multiple continuous duration intervals is obtained, and each viewing probability is adjusted to obtain the adjusted viewing probability, so that the viewing probability of a larger duration interval is less than the viewing probability corresponding to the previous smaller duration interval, and then based on each duration interval and the adjusted viewing probability corresponding to the duration interval, the predicted viewing time of the object for the multimedia resource is determined, which can reduce the limitations of the implicit assumptions of the data distribution of the duration prediction model, and is more friendly to the scale and extreme values ​​of the duration interval, and the predicted viewing time of the multimedia resources obtained by the duration prediction model can be more accurate without human intervention.

[0049] In some feasible implementations, in step 120, the viewing probability corresponding to each duration interval may be reduced to obtain an adjusted viewing probability.

[0050] Here, for each duration interval, the viewing probability may be adjusted to be smaller, so that the adjusted viewing probability makes the viewing probability of a larger duration interval smaller than the viewing probability corresponding to the previous smaller duration interval.

[0051] It should be understood that the magnitude of reduction of each viewing probability may be different, but the ultimate goal is to make the viewing probability of a larger duration interval smaller than the viewing probability corresponding to the previous smaller duration interval.

[0052] In some embodiments, for each duration interval except the first duration interval in multiple consecutive duration intervals, the viewing probability corresponding to the duration interval is converted into a negative viewing probability through a preset function, and then based on the negative viewing probability and the viewing probability corresponding to the previous duration interval of the duration interval, the adjusted viewing probability of the duration interval is obtained.

[0053] Among them, for the first duration interval among multiple consecutive duration intervals, the viewing probability of the first duration interval is not adjusted. For example, multiple consecutive duration intervals are (0, 1] seconds, (1, 2] seconds, (2, 3] seconds, (3, 4] seconds, ..., (1799, 1800] seconds, then the viewing probability of the duration interval of (0, 1] seconds is not adjusted, that is, among all the adjusted viewing probabilities, the viewing probability of the first duration interval will be the largest.

[0054] For each duration interval except the first duration interval among the multiple consecutive duration intervals, the value range of the viewing probability corresponding to the duration interval is transformed into a negative number through a preset function to obtain a negative viewing probability. For example, if the multiple consecutive duration intervals are (0, 1] seconds, (1, 2] seconds, (2, 3] seconds, (3, 4] seconds, ..., (1799, 1800] seconds, the value range of the viewing probability corresponding to the duration intervals of (1, 2] seconds, (2, 3] seconds, (3, 4] seconds, ..., (1799, 1800] seconds is transformed into a negative number through a preset function to obtain the corresponding negative viewing probabilities.

[0055] Exemplarily, the activation function used by the duration prediction model is a monotone continuous type activation function, and the preset function is a monotone continuous type function.

[0056] It should be understood that the monotone continuous type activation function may be a sigmoid activation function or a tanh activation function, etc. The preset function may be a variant of the ELU activation function, or the preset function may be a variant of the exponential function. The ELU activation function is a variant of the Relu activation function and is a monotone increasing type function.

[0057] The ELU activation function is expressed as: (1) The ELU activation function has the advantages of the Relu activation function, but does not have the Dead Relu problem (the phenomenon that the output of some neurons is always 0). By reducing the influence of bias offset, the normal gradient is close to the unit natural gradient, and it will saturate to a negative value under smaller inputs, thus having a certain degree of robustness to noise.

[0058] In the embodiment of the present disclosure, the preset function may be a variant of the ELU activation function. For example, the preset function may be: .

[0059] Since the activation function used in the duration prediction model is a monotonic continuous type of activation function (such as the sigmoid activation function), Can be converted to ,in, is the viewing probability of the nth duration interval, is the viewing probability of the n+1th duration interval.

[0060] The default function is For example, the adjusted viewing probability can be expressed as: (2) in, is the negative viewing probability corresponding to the nth duration interval, It is the viewing probability corresponding to the nth duration interval output by the duration prediction model.

[0061] Next, the adjusted viewing probability may be obtained based on the sum of the negative viewing probability and the corresponding viewing probability of the previous duration interval.

[0062] For example, the adjusted viewing probability can be obtained by a first calculation formula, which is: (3) in, is the adjusted viewing probability corresponding to the n+1th duration interval, is the adjusted viewing probability corresponding to the nth duration interval, is the negative viewing probability corresponding to the n+1th duration interval.

[0063] It is worth noting that due to is always negative, thus ensuring , so that the viewing probability of the multimedia resource in the previous second is greater than or equal to the viewing probability in the next second.

[0064] In some embodiments, the negative viewing probability can be normalized according to the total number of multiple consecutive duration intervals to obtain a target probability, and then the adjusted viewing probability of the duration interval can be obtained according to the sum of the viewing probability corresponding to the previous duration interval of the duration interval and the target probability. Each duration interval represents a classification target, and the total number of multiple consecutive duration intervals represents the number of classification targets. The total number of multiple consecutive duration intervals can be used to normalize the negative viewing probability to avoid gradient explosion during back propagation of the activation function.

[0065] Exemplarily, the adjusted viewing probability may be obtained by a second calculation formula, which is:

[0066] Where N is the total number. It should be understood that by This can avoid gradient explosion during back propagation of the sigmoid activation function.

[0067] Thus, through the above implementation, the viewing probability of multiple consecutive time intervals can be The order-preserving constraint ensures that the duration prediction model can obtain accurate preset viewing duration.

[0068] In some feasible implementations, in step 130, for each duration interval, based on the product of the duration of the duration interval and the corresponding adjusted viewing probability, a target viewing duration corresponding to the duration interval is obtained, and then the predicted viewing time of the object for the multimedia resource to be predicted is determined based on the sum of the target viewing times.

[0069] Here, the duration of the duration interval refers to the length of time occupied by the duration interval. For example, the duration of the duration interval of (0, 1] seconds corresponds to 1 second. Of course, if the duration interval is (0, 60] seconds, the duration of the duration interval of (0, 60] seconds corresponds to 60 seconds.

[0070] For each duration interval, the target viewing duration corresponding to the duration interval can be calculated based on the duration corresponding to the duration interval and the corresponding adjusted viewing probability. The target viewing duration actually indicates how long the user is likely to watch the multimedia resource in the duration interval.

[0071] For example, assuming that the adjusted viewing probability corresponding to the duration interval of (2, 3] seconds is 1, the target viewing time of the duration interval of (2, 3] seconds is 1 second. For another example, assuming that the adjusted viewing probability corresponding to the duration interval of (1799, 1800] seconds is 0, the target viewing time of the duration interval of (1799, 1800] seconds is 0 seconds.

[0072] It should be noted that if the first duration interval is not adjusted, the adjusted viewing probability corresponding to the first duration interval is the viewing probability of the first duration interval output by the duration prediction model, that is, the original viewing probability of the first duration interval.

[0073] After the target viewing durations corresponding to the respective duration intervals are obtained, the sum of the target viewing durations is taken as the predicted viewing duration for the multimedia resource to be predicted.

[0074] Therefore, through the above implementation, the duration prediction model can obtain accurate predicted viewing duration, thereby providing support for content recommendation.

[0075] Figure 2 FIG. 1 is a schematic diagram showing the principle of multimedia resource viewing time prediction according to some embodiments. Figure 2As shown in the figure, the number of units in the last layer of the fully connected network of the deep model is set to n (n corresponds to the number of time intervals), and the fully connected network will output n viewing probabilities (including logit(1), logit(2), logit(3), ..., logit(n)). Then, the viewing probabilities of the n viewing probabilities except the viewing probability of the first time interval (logit(1)) are adjusted so that the viewing probability of the larger time interval is less than the viewing probability corresponding to the previous smaller time interval. Then, the range of the viewing probability is limited to [0, 1] through the sigmoid activation function, and finally, based on the duration of the time interval and the corresponding viewing probability, the predicted viewing time of the object for the multimedia resource is calculated.

[0076] It is worth noting that in the embodiments of the present disclosure, the duration prediction model is only used as an example of a deep model, which does not mean that the duration prediction model can only be a deep model. Other types of machine learning models may also be used.

[0077] It should be noted that in the embodiments of the present disclosure, the training process of the duration prediction model is not described in detail, and multimedia resources and attribute information carrying real duration labels can be used to train the machine learning model to obtain a trained duration prediction model. In the training stage, the cross entropy loss function can be used to train the duration prediction model.

[0078] Exemplarily, the cross entropy loss function may be:

[0079] in, is the real duration label of the nth duration interval.

[0080] Figure 3 FIG. 1 is a schematic diagram of a structure of a multimedia resource viewing duration prediction device according to some embodiments. Figure 3 As shown, the embodiment of the present disclosure provides a multimedia resource viewing time prediction device 300, and the multimedia resource viewing time prediction device 300 includes: The first obtaining module 301 is configured to input the attribute information of the multimedia resource and object to be predicted into the trained duration prediction model to obtain the corresponding viewing probability of the multimedia resource in multiple continuous duration intervals; The second obtaining module 302 is configured to adjust each viewing probability to obtain an adjusted viewing probability, wherein the adjusted viewing probability makes the viewing probability of a larger duration interval smaller than the viewing probability corresponding to the previous smaller duration interval; The determination module 303 is configured to determine the predicted viewing duration of the object for the multimedia resource to be predicted based on each of the duration intervals and the adjusted viewing probabilities corresponding to the duration intervals.

[0081] Optionally, the second obtaining module 302 is specifically configured to: For the viewing probability corresponding to each of the duration intervals, the viewing probability is adjusted to be smaller to obtain the adjusted viewing probability.

[0082] Optionally, the second obtaining module 302 is specifically configured to: For each duration interval except the first duration interval among the plurality of continuous duration intervals, transforming the viewing probability corresponding to the duration interval into a negative viewing probability by using a preset function; Based on the negative viewing probability and the viewing probability corresponding to the previous duration interval of the duration interval, the adjusted viewing probability of the duration interval is obtained.

[0083] Optionally, the second obtaining module 302 is specifically configured to: Normalizing the viewing probability of the negative number according to the total number of the multiple continuous duration intervals to obtain a target probability; The adjusted viewing probability of the duration interval is obtained according to the sum of the viewing probability corresponding to the previous duration interval of the duration interval and the target probability.

[0084] Optionally, the activation function used by the duration prediction model is a monotone continuous type activation function, and the preset function is a monotone continuous type function.

[0085] Optionally, the determining module 303 is specifically configured to: For each of the duration intervals, based on the product of the duration of the duration interval and the corresponding adjusted viewing probability, obtaining a target viewing duration corresponding to the duration interval; The predicted viewing time of the object for the multimedia resource to be predicted is determined according to the sum of the target viewing times.

[0086] The logic of the method executed by each functional module in the multimedia resource viewing duration prediction device 300 can refer to the method related to the above embodiment, and will not be repeated here.

[0087] Reference below Figure 4, which shows a schematic diagram of the structure of an electronic device (such as a terminal device or a server) 400 suitable for implementing the embodiment of the present disclosure. The terminal device in the embodiment of the present disclosure may include but is not limited to mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0088] like Figure 4 As shown, the electronic device 400 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage device 408 into a random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the electronic device 400 are also stored. The processing device 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0089] Typically, the following devices may be connected to the I / O interface 405: an input device 406 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 407 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 408 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 409. The communication device 409 may allow the electronic device 400 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 4 The electronic device 400 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0090] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 409, or installed from the storage device 408, or installed from the ROM 402. When the computer program is executed by the processing device 401, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0091] It should be noted that the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. Computer readable signal media may also be any computer readable medium other than computer readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0092] In some embodiments, the electronic devices may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0093] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0094] The above-mentioned computer-readable medium carries one or more programs. When the above-mentioned one or more programs are executed by the electronic device, the electronic device: inputs the attribute information of the multimedia resources and objects to be predicted into the trained duration prediction model to obtain the viewing probability corresponding to the multimedia resources in multiple continuous duration intervals; adjusts each of the viewing probabilities to obtain an adjusted viewing probability, wherein the adjusted viewing probability makes the viewing probability of a larger duration interval smaller than the viewing probability corresponding to the previous smaller duration interval; based on each of the duration intervals and the adjusted viewing probabilities corresponding to the duration intervals, determines the predicted viewing time of the object for the multimedia resource to be predicted.

[0095] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0096] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0097] The modules involved in the embodiments described in the present disclosure may be implemented by software or hardware, wherein the name of a module does not, in some cases, limit the module itself.

[0098] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0099] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0100] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0101] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0102] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims. Regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Claims

1. A method for predicting the viewing time of multimedia resources, characterized in that: include: Inputting the attribute information of the multimedia resource and object to be predicted into the trained duration prediction model to obtain the corresponding viewing probability of the multimedia resource in multiple continuous duration intervals; Adjusting each of the viewing probabilities to obtain an adjusted viewing probability, wherein the adjusted viewing probability makes the viewing probability of a larger duration interval smaller than the viewing probability corresponding to a previous smaller duration interval; Based on each of the duration intervals and the adjusted viewing probabilities corresponding to the duration intervals, a predicted viewing duration of the object for the multimedia resource to be predicted is determined.

2. The method according to claim 1, characterized in that: The adjusting each viewing probability to obtain an adjusted viewing probability includes: For the viewing probability corresponding to each of the duration intervals, the viewing probability is adjusted to be smaller to obtain the adjusted viewing probability.

3. The method according to claim 2, characterized in that The step of reducing the viewing probability corresponding to each of the time intervals to obtain an adjusted viewing probability includes: For each duration interval except the first duration interval among the plurality of continuous duration intervals, transforming the viewing probability corresponding to the duration interval into a negative viewing probability by using a preset function; Based on the negative viewing probability and the viewing probability corresponding to the previous duration interval of the duration interval, the adjusted viewing probability of the duration interval is obtained.

4. The method according to claim 3, characterized in that The obtaining the adjusted viewing probability of the duration interval based on the viewing probability of the negative number and the viewing probability corresponding to the previous duration interval of the duration interval includes: Normalizing the viewing probability of the negative number according to the total number of the multiple continuous duration intervals to obtain a target probability; The adjusted viewing probability of the duration interval is obtained according to the sum of the viewing probability corresponding to the previous duration interval of the duration interval and the target probability.

5. The method according to claim 3, characterized in that: The activation function used by the duration prediction model is a monotone continuous type activation function, and the preset function is a monotone continuous type function.

6. The method according to any one of claims 1 to 5, characterized in that The determining, based on each of the duration intervals and the adjusted viewing probabilities corresponding to the duration intervals, the predicted viewing duration of the object for the multimedia resource to be predicted includes: For each of the duration intervals, based on the product of the duration of the duration interval and the corresponding adjusted viewing probability, obtaining a target viewing duration corresponding to the duration interval; The predicted viewing time of the object for the multimedia resource to be predicted is determined according to the sum of the target viewing times.

7. A multimedia resource viewing time prediction device, characterized in that: include: A first obtaining module is configured to input the attribute information of the multimedia resource and object to be predicted into the trained duration prediction model to obtain the corresponding viewing probability of the multimedia resource in multiple continuous duration intervals; A second obtaining module is configured to adjust each viewing probability to obtain an adjusted viewing probability, wherein the adjusted viewing probability makes the viewing probability of a larger duration interval smaller than the viewing probability corresponding to the previous smaller duration interval; The determination module is configured to determine the predicted viewing duration of the object for the multimedia resource to be predicted based on each of the duration intervals and the adjusted viewing probabilities corresponding to the duration intervals.

8. A computer readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processing device, the steps of the method according to any one of claims 1 to 6 are implemented.

9. An electronic device, characterized in that: include: a storage device having a computer program stored thereon; A processing device, configured to execute the computer program in the storage device to implement the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images