CN117473107A

CN117473107A - Training method, device, equipment and storage medium for media content click model

Info

Publication number: CN117473107A
Application number: CN202311501588.4A
Authority: CN
Inventors: 杨镜子
Original assignee: Beijing Zitiao Network Technology Co Ltd
Current assignee: Beijing Zitiao Network Technology Co Ltd
Priority date: 2023-11-10
Filing date: 2023-11-10
Publication date: 2024-01-30

Abstract

The embodiment of the disclosure provides a media content click model training method, device, equipment and storage medium. The method comprises the following steps: acquiring a media content sample and a sorting position corresponding to the media content sample and a search result; inputting the target feature set corresponding to the media content sample into a media content click model to be trained, and determining a non-deviation click rate corresponding to the media content sample; and determining position deviation information corresponding to the media content samples according to the sorting positions, determining a target loss value according to the position deviation information and the unbiased click rate, and training the media content click model to be trained according to the target loss value. The embodiment of the disclosure can improve the click rate deviation caused by the sorting positions, thereby relieving the problem that the click rate prediction result deviates from the real search requirement due to the sorting positions.

Description

Training method, device, equipment and storage medium for media content click model

Technical Field

The embodiment of the disclosure relates to a search technology, in particular to a media content click model training method, device, equipment and storage medium.

Background

Currently, the ranking position of the candidate media content in the search result is related to the interaction information corresponding to the candidate media content.

For example, if the candidate media content is clicked, focused, or browsed more times, its ranking position may be the more advanced. The more the candidate media content ranked at the front is clicked, focused or browsed, the higher the click rate of the candidate documents is predicted by the click model, so that the ranking position of the candidate media content in the search result is continuously moved forward.

However, it is possible that the later ranked candidate media content in the search results is the resource that better meets the user's search needs, and that these resources are rarely clicked, focused, or browsed due to the later ranked positions. The click rate of the candidate media contents is predicted to be lower through the click model, so that the sorting position of the candidate media contents in the search result is more backward, and the problem that the click rate prediction result deviates from the real search requirement exists. The user needs to spend more time to browse the candidate media content meeting the search requirement, and the user experience is poor.

Disclosure of Invention

The embodiment of the disclosure provides a training method, device, equipment and storage medium for a media content click model, which can relieve the problem that a click rate prediction result deviates from a real search requirement due to ordering positions.

In a first aspect, an embodiment of the present disclosure provides a media content click model training method, including:

acquiring a media content sample and a sorting position corresponding to the media content sample and a search result;

inputting the target feature set corresponding to the media content sample into a media content click model to be trained, and determining a non-deviation click rate corresponding to the media content sample, wherein the target feature set is determined based on the influence degree of the feature existence determined by a feature selection module on a model prediction result, and the training ending condition of the feature selection module is determined based on a discarded feature proportion;

and determining position deviation information corresponding to the media content samples according to the sorting positions, determining a target loss value according to the position deviation information and the unbiased click rate, and training the media content click model to be trained according to the target loss value.

In a second aspect, an embodiment of the present disclosure further provides a media content click model training apparatus, including:

the sample acquisition module is used for acquiring a media content sample and an ordering position corresponding to the media content sample and a search result;

the click rate determining module is used for inputting the target feature set corresponding to the media content sample into a media content click model to be trained, and determining a bias-free click rate corresponding to the media content sample, wherein the target feature set is determined based on the influence degree of the feature existence determined by the feature selecting module on a model prediction result, and the training ending condition of the feature selecting module is determined based on a discarded feature proportion;

and the model training module is used for determining position deviation information corresponding to the media content samples according to the sorting positions, determining a target loss value according to the position deviation information and the unbiased click rate, and training the media content click model to be trained according to the target loss value.

In a third aspect, embodiments of the present disclosure further provide an electronic device, including:

one or more processors;

storage means for storing one or more programs,

the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the media content click model training method as described in any embodiment of the present disclosure.

In a fourth aspect, the disclosed embodiments also provide a storage medium containing computer-executable instructions, which when executed by a computer processor, are for performing a media content click model training method as described in any of the embodiments of the disclosure.

The embodiment of the disclosure provides a media content click model training method, device, equipment and storage medium, wherein the influence degree of characteristics required by model training on a model prediction result is determined through a characteristic selection module, whether the characteristics are discarded or not is determined according to the influence degree, a target characteristic set is obtained, and the target characteristic set corresponding to a media content sample is input into a media content click model to be trained, so that the influence of characteristic redundancy on the model training speed and the influence on the model online prediction efficiency are avoided. Then, determining a target loss value according to position deviation information and unbiased click rate corresponding to the media content sample, and performing model training according to the target loss value can improve click rate deviation caused by sorting positions, so that the problem that a click rate prediction result deviates from a real search requirement due to sorting positions is solved.

Drawings

The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. The same or similar reference numbers will be used throughout the drawings to refer to the same or like elements. It should be understood that the figures are schematic and that elements and components are not necessarily drawn to scale.

FIG. 1 is a flowchart of a training method for a media content click model according to an embodiment of the present disclosure;

FIG. 2 is a schematic diagram of a model training framework provided by the present disclosure;

fig. 3 is a flowchart of a click rate determining method according to an embodiment of the present disclosure;

FIG. 4 is a schematic diagram of a training device for a media content click model according to an embodiment of the present disclosure;

fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure.

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure have been shown in the accompanying drawings, it is to be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for illustration purposes only and are not intended to limit the scope of the present disclosure.

It should be understood that the various steps recited in the method embodiments of the present disclosure may be performed in a different order and/or performed in parallel. Furthermore, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.

The term "including" and variations thereof as used herein are intended to be open-ended, i.e., including, but not limited to. The term "based on" is 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. Related definitions of other terms will be given in the description below.

It should be noted that the terms "first," "second," and the like in this disclosure are merely used to distinguish between different devices, modules, or units and are not used to define an order or interdependence of functions performed by the devices, modules, or units.

It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.

The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

It will be appreciated that prior to using the technical solutions disclosed in the embodiments of the present disclosure, the user should be informed and authorized of the type, usage range, usage scenario, etc. of the personal information related to the present disclosure in an appropriate manner according to the relevant legal regulations.

For example, in response to receiving an active request from a user, a prompt is sent to the user to explicitly prompt the user that the operation it is requesting to perform will require personal information to be obtained and used with the user. Thus, the user can autonomously select whether to provide personal information to software or hardware such as an electronic device, an application program, a server or a storage medium for executing the operation of the technical scheme of the present disclosure according to the prompt information.

As an alternative but non-limiting implementation, in response to receiving an active request from a user, the manner in which the prompt information is sent to the user may be, for example, a popup, in which the prompt information may be presented in a text manner. In addition, a selection control for the user to select to provide personal information to the electronic device in a 'consent' or 'disagreement' manner can be carried in the popup window.

It will be appreciated that the above-described notification and user authorization process is merely illustrative and not limiting of the implementations of the present disclosure, and that other ways of satisfying relevant legal regulations may be applied to the implementations of the present disclosure.

It will be appreciated that the data (including but not limited to the data itself, the acquisition or use of the data) involved in the present technical solution should comply with the corresponding legal regulations and the requirements of the relevant regulations.

Fig. 1 is a schematic flow chart of a media content click model training method provided by an embodiment of the present disclosure, where the embodiment of the present disclosure is applicable to a case of optimizing a click model, the method may be performed by a model training apparatus, and the apparatus may be implemented in a form of software and/or hardware, optionally, implemented by an electronic device, and the electronic device may be a server or the like.

As shown in fig. 1, the method includes:

s110, acquiring a media content sample and a sorting position of the media content sample corresponding to a search result.

Wherein the media content sample may be historical media content that has been historically searched. Alternatively, the media content sample may be a document, a picture, audio or video, etc. corresponding to the historical search request, and the embodiment contrast of the present disclosure is not particularly limited.

The historical media content may have different ranking positions in the search results. After displaying the search results in response to the historical search request, a ranking position of each historical media content corresponding to the historical search request in the search results may be recorded.

And when the click model is trained, the historical media content corresponding to the historical search request in the set time period can be obtained as a training sample. In addition, the ranking positions of the media content samples corresponding to the search results are also obtained.

Illustratively, historical media content corresponding to the historical search request is taken as a media content sample. And taking the ranking position of the historical media content in the search result as the ranking position of the media content sample corresponding to the search result.

S120, inputting the target feature set corresponding to the media content sample into a media content click model to be trained, and determining a non-deviation click rate corresponding to the media content sample.

The target feature set is determined based on the influence degree of the feature existence determined by the feature selection module on the model prediction result, and the training ending condition of the feature selection module is determined based on the discarded feature proportion. The proportion of the discarded features characterizes the proportion of the discarded features in the total features, the proportion of the discarded features can be a preset value, and the proportion of the discarded features can be dynamically adjusted according to different application scenes.

During long-term iterations of clicking on the model, the input features of the model training may gradually increase. Along with the change of interaction information or the migration of interest information, some old features cannot adapt to a new scene, the old features which are not adaptive to the new scene need to be screened out from the original feature set, and the old features are discarded on the basis that the prediction effect of the click model is not affected, so that feature redundancy is avoided, and the online prediction speed and the offline training speed of the click model are improved.

Illustratively, determining the target feature set using the feature selection module includes: acquiring an original feature set corresponding to the media content click model to be trained; inputting the original feature set into the feature selection module, and determining a discarding factor of each original feature in the original feature set according to the feature selection module; and determining a target feature set according to the discarding factor of each original feature in the original feature set. Wherein the discard factor may characterize how much the feature presence affects the predicted outcome of the media content click model.

Optionally, determining the target feature set according to the discarding factor of each original feature in the original feature set includes: determining a feature vector corresponding to the original feature in the original feature set; and determining a target feature set according to the feature vector and the discarding factor corresponding to the original feature.

For example, determining a feature vector corresponding to an original feature through an embedding layer (embedding), then determining whether the original feature needs to be discarded by combining a discarding factor and the feature vector corresponding to the original feature determined by the feature selection module, if yes, discarding the original feature, otherwise, reserving the original feature, and forming a target feature set according to the reserved original feature.

In some embodiments, the feature selection module may be trained such that the feature selection module learns the significance of the existence of features of the training sample to the model predictions and continuously optimizes the significance of iterations during the model training process. The feature selection module determines a learnable discard factor for each feature, multiplying the discard factor by the feature vector as an embedded vector for the feature. If the value of the discarding factor tends to 0, the embedded vector of the feature tends to 0, so that the effect of discarding the feature is realized. If the discarding factor value tends to 1, the embedded vector of the feature also tends to 1, so that the effect of preserving the feature is realized.

Alternatively, the importance of the target feature may be used to characterize the extent to which the presence of the target feature affects the model's prediction results. A target feature set determined using a pre-trained feature selection module, comprising: acquiring an original feature set corresponding to the media content click model to be trained; inputting the original feature set into the pre-trained feature selection module, and determining the importance degree of each original feature in the original feature set according to the feature selection module; and determining a target feature set according to the importance degree of each original feature in the original feature set.

For example, the original features in the original feature set corresponding to the click model are input into a feature selection module, the importance degree corresponding to the original features is determined according to the feature selection module, and the original features with the importance degree close to 0 are deleted from the original feature set, so that the target feature set is obtained. Wherein, the importance approaching 0 may be smaller than the set threshold value.

Optionally, during the training of the feature selection module, a duty ratio of the retained features in the original feature set is determined. Judging whether training of the feature selection module is finished or not according to the proportion of the duty ratio and the discarding feature. Wherein determining whether to end training of the feature selection module according to the ratio of the duty ratio to the discard feature may include: and judging whether training of the feature selection module is finished or not according to the ratio of the duty ratio to the discarding feature ratio. Or judging whether to end training of the feature selection module according to the difference value of the duty ratio and the discarding feature ratio. Or judging whether to finish training the feature selection module according to the numerical value of the ratio of the duty ratio to the discarding feature.

For example, in the training process of the feature selection module, the number of the reserved features is counted, the duty ratio of the reserved features in the original feature set is determined, the difference value between the duty ratio of the reserved features in the original feature set and the proportion of the discarded features is calculated, and whether the feature selection module converges is judged according to the difference value. And if the feature selection module converges, ending training. If the feature selection module is not converged, continuing to iterate and optimize the feature selection module to obtain the feature selection module with the training completed.

In the embodiment of the disclosure, the unbiased click rate may be a click rate that eliminates a deviation between a predicted value and a true value of the click rate caused by posterior features. The posterior features may include ordering positions, interactive information such as clicking, browsing, etc.

Fig. 2 is a schematic diagram of a model training framework provided by the present disclosure. As shown in fig. 2, the model training framework includes a position deviation determination model 210 and a media content click model 220. The target feature set corresponding to the media content sample is input into the media content click model 220, and the unbiased click rate corresponding to the media content sample is predicted through the media content click model 220.

It should be noted that, the media content click model to be trained uses the target feature set of the media content sample as an input feature, outputs a non-deviation click rate corresponding to the media content sample, and uses the real click rate of the media content sample to supervise the training of the media content click model to be trained.

S130, determining position deviation information corresponding to the media content samples according to the sorting positions, determining a target loss value according to the position deviation information and the unbiased click rate, and training the media content click model to be trained according to the target loss value.

The position deviation information may be an effect of a position of the media content sample on the search result page on the click rate. In general, the higher the top ranking position on the search results page, the easier it is to obtain a higher click rate. Thus, the position deviation information may characterize click rate differences due to different ranking positions of the media content samples.

As shown in fig. 2, the ranking positions of the media content samples corresponding to the search results are input into the position deviation determining model 210, and position deviation information output by the position deviation determining model 210 is acquired.

In some embodiments, determining the target loss value based on the positional deviation information and the unbiased click rate includes: determining the biased click rate corresponding to the media content sample according to the position bias information and the unbiased click rate corresponding to the media content sample; acquiring a real click rate corresponding to the media content sample; and determining the target loss value according to the biased click rate and the real click rate corresponding to the media content sample.

The biased click rate may be a click rate having a bias between a predicted value and a true value of the click rate due to posterior features. The actual click rate may be a true value of the click rate corresponding to the media content sample. For example, the click rate of the media content samples on the search results page is taken as the true click rate. The target loss value may characterize a deviation between a biased click rate and a true click rate of the media content sample.

Optionally, determining the position deviation information corresponding to the media content sample and the addition operation result of the unbiased click rate; and determining the biased click rate corresponding to the media content sample according to the addition operation result.

For example, the position deviation information corresponding to the media content sample is directly added to the non-deviation click rate to obtain the deviation click rate corresponding to the media content sample. Then, determining a target loss value according to the biased click rate, the real click rate and the preset loss function of the media content sample. Determining whether the click model meets the training ending condition according to the target loss value, if the training ending condition is not met, adjusting parameters of the media content click model, calculating a new biased click rate for the unbiased click rate output by the click model after parameter adjustment, determining a new target loss value according to the new biased click rate and the real click rate, and performing iterative training on the click model by adopting the mode until the training ending condition is met.

According to the technical scheme, the influence degree of the characteristics required by model training on the model prediction result is determined through the characteristic selection module, whether the characteristics are discarded or not is determined according to the influence degree, a target characteristic set is obtained, the target characteristic set corresponding to a media content sample is input into the media content click model to be trained, the influence of characteristic redundancy on the model training speed is avoided, and the online prediction efficiency of the model is influenced. Then, determining a target loss value according to position deviation information and unbiased click rate corresponding to the media content sample, and performing model training according to the target loss value can improve click rate deviation caused by sorting positions, so that the problem that a click rate prediction result deviates from a real search requirement due to sorting positions is solved.

Fig. 3 is a flowchart of a click rate determining method according to an embodiment of the present disclosure, where the present disclosure provides an application scenario of a media content click model based on the foregoing embodiment.

As shown in fig. 3, the method includes:

s310, obtaining candidate documents corresponding to the search request, and obtaining target feature sets corresponding to the candidate documents.

Wherein the target feature set is determined based on the importance of the target feature determined using the pre-trained feature selection module. For example, the target feature set includes content matching degree and interaction attribute corresponding to the document. The content matching degree can be characterized by content quality, and is the matching degree of the document content and the search request.

The candidate documents corresponding to the search request and target features such as content quality, average click rate, average display position, display times and the like corresponding to the candidate documents are obtained.

S320, inputting the target feature set corresponding to the candidate document into a media content click model, and determining the predicted click rate of the candidate document according to the media content click model.

The target feature set corresponding to the candidate document is input into a media content click model, and the predicted click rate of the candidate document is predicted based on the content quality, the average click rate, the average display position and the display times corresponding to the candidate document through the media content click model.

As the media content click model is trained based on the model training method provided by the embodiment of the disclosure, the deviation of click rate prediction caused by the sorting positions is improved, and the candidate documents with better content quality have higher predicted click rate than the candidate documents with original display positions in front, so that the search results more meet the search requirement.

S330, determining the sorting positions of the candidate documents in the search results according to the predicted click rate.

According to the technical scheme, the target feature set corresponding to the candidate document is input into the media content click model, the predicted click rate output by the media content click model is obtained, then the sorting position of the candidate document in the search result is determined according to the predicted click rate, the deviation of click rate prediction caused by the sorting position is improved, the candidate document with better content quality is higher in predicted click rate than the candidate document with the original display position, and the search result meets the search requirement better.

Fig. 4 is a schematic structural diagram of a training device for a media content click model according to an embodiment of the present disclosure, where the training device may be implemented in software and/or hardware, and optionally, implemented by an electronic device, and the electronic device may be a server or the like.

As shown in fig. 4, the apparatus includes: sample acquisition module 410, click rate determination module 420, and model training module 430.

A sample acquiring module 410, configured to acquire a media content sample and an ordering position corresponding to the media content sample and a search result;

the click rate determining module 420 is configured to input the target feature set corresponding to the media content sample into a media content click model to be trained, and determine a bias-free click rate corresponding to the media content sample, where the target feature set is determined based on a degree of influence of the feature presence determined by the feature selecting module on the model prediction result, and a training ending condition of the feature selecting module is determined based on a discarded feature proportion;

the model training module 430 is configured to determine position deviation information corresponding to the media content samples according to the sorting positions, determine a target loss value according to the position deviation information and a non-deviation click rate, and train the media content click model to be trained according to the target loss value.

According to the technical scheme provided by the embodiment of the disclosure, the influence degree of the characteristics required by model training on the model prediction result is determined through the characteristic selection module, whether the characteristics are discarded or not is determined according to the influence degree, a target characteristic set is obtained, and the characteristics redundancy is avoided to influence the model training speed and the model online prediction efficiency by inputting the target characteristic set corresponding to the media content sample into the media content click model to be trained. Then, determining a target loss value according to position deviation information and unbiased click rate corresponding to the media content sample, and performing model training according to the target loss value can improve click rate deviation caused by sorting positions, so that the problem that a click rate prediction result deviates from a real search requirement due to sorting positions is solved.

Optionally, the sample acquisition module 410 is specifically configured to:

taking the historical media content corresponding to the historical search request as a media content sample;

and taking the ranking position of the historical media content in the search result as the ranking position of the media content sample corresponding to the search result.

Optionally, the apparatus further includes a feature set determining module, including:

the original feature set acquisition unit is used for acquiring an original feature set corresponding to the media content click model to be trained;

a discarding factor determining unit, configured to input the original feature set into the feature selection module, and determine a discarding factor of each original feature in the original feature set according to the feature selection module;

and the target feature set determining unit is used for determining a target feature set according to the discarding factor of each original feature in the original feature set.

Further, the target feature set determining unit is specifically configured to:

determining a feature vector corresponding to the original feature in the original feature set;

and determining a target feature set according to the feature vector and the discarding factor corresponding to the original feature.

Optionally, the model training module 430 includes:

the click rate determining unit is used for determining the biased click rate corresponding to the media content sample according to the position bias information and the unbiased click rate corresponding to the media content sample;

the click rate acquisition unit is used for acquiring the real click rate corresponding to the media content sample;

and the loss value determining unit is used for determining the target loss value according to the biased click rate and the real click rate corresponding to the media content sample.

Further, the click rate determination unit is specifically configured to:

determining the position deviation information corresponding to the media content sample and the addition operation result of the unbiased click rate;

and determining the biased click rate corresponding to the media content sample according to the addition operation result.

Optionally, the apparatus further comprises a training end determination module for: determining the duty ratio of the retained features in the original feature set; judging whether training of the feature selection module is finished or not according to the proportion of the duty ratio and the discarding feature.

The media content click model training device provided by the embodiment of the disclosure can execute the media content click model training method provided by any embodiment of the disclosure, and has the corresponding functional modules and beneficial effects of the execution method.

It should be noted that each unit and module included in the above apparatus are only divided according to the functional logic, but not limited to the above division, so long as the corresponding functions can be implemented; in addition, the specific names of the functional units are also only for convenience of distinguishing from each other, and are not used to limit the protection scope of the embodiments of the present disclosure.

Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure. Referring now to fig. 5, a schematic diagram of an electronic device (e.g., a terminal device or server in fig. 5) 500 suitable for use in implementing embodiments of the present disclosure is shown. The terminal devices in the embodiments of the present disclosure may include, but are 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), in-vehicle terminals (e.g., in-vehicle navigation terminals), and the like, and stationary terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 5 is merely an example and should not be construed to limit the functionality and scope of use of the disclosed embodiments.

As shown in fig. 5, the electronic device 500 may include a processing means (e.g., a central processing unit, a graphics processor, etc.) 501, which may perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 502 or a program loaded from a storage means 508 into a Random Access Memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic apparatus 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An edit/output (I/O) interface 505 is also connected to bus 504.

In general, the following devices may be connected to the I/O interface 505: input devices 506 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; an output device 507 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 508 including, for example, magnetic tape, hard disk, etc.; and communication means 509. The communication means 509 may allow the electronic device 500 to communicate with other devices wirelessly or by wire to exchange data. While fig. 5 shows an electronic device 500 having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead.

In particular, according to embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a non-transitory computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication means 509, or from the storage means 508, or from the ROM 502. The above-described functions defined in the methods of the embodiments of the present disclosure are performed when the computer program is executed by the processing device 501.

The electronic device provided by the embodiment of the present disclosure and the training method of the media content click model provided by the foregoing embodiment belong to the same inventive concept, and technical details not described in detail in the present embodiment may be referred to the foregoing embodiment, and the present embodiment has the same beneficial effects as the foregoing embodiment.

The disclosed embodiments provide a computer storage medium having stored thereon a computer program which, when executed by a processor, implements the media content click model training method provided by the above embodiments.

It should be noted that the computer readable medium described in the present disclosure may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this disclosure, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present disclosure, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.

In some implementations, the clients, servers 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"), the internet (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed networks.

The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device.

The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to:

Computer program code for carrying out operations of the present disclosure may be written in one or more programming languages, including, but not limited to, an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units involved in the embodiments of the present disclosure may be implemented by means of software, or may be implemented by means of hardware. Wherein the names of the units do not constitute a limitation of the units themselves in some cases.

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: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a Complex Programmable Logic Device (CPLD), and the like.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The foregoing description is only of the preferred embodiments of the present disclosure and description of the principles of the technology being employed. It will be appreciated by persons skilled in the art that the scope of the disclosure referred to in this disclosure is not limited to the specific combinations of features described above, but also covers other embodiments which may be formed by any combination of features described above or equivalents thereof without departing from the spirit of the disclosure. Such as those described above, are mutually substituted with the technical features having similar functions disclosed in the present disclosure (but not limited thereto).

Moreover, although operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order. In certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the present disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.

Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are example forms of implementing the claims.

Claims

1. A method for training a media content click model, comprising:

2. The method of claim 1, wherein the obtaining the media content samples and the ranking positions of the media content samples corresponding to the search results comprises:

3. The method of claim 1, wherein determining the target feature set using the feature selection module comprises:

acquiring an original feature set corresponding to the media content click model to be trained;

inputting the original feature set into the feature selection module, and determining a discarding factor of each original feature in the original feature set according to the feature selection module;

and determining a target feature set according to the discarding factor of each original feature in the original feature set.

4. A method according to claim 3, wherein said determining a target feature set from the discard factor of each original feature in said set of original features comprises:

5. The method of claim 1, wherein the determining a target loss value based on the positional deviation information and a no deviation click rate comprises:

determining the biased click rate corresponding to the media content sample according to the position bias information and the unbiased click rate corresponding to the media content sample;

acquiring a real click rate corresponding to the media content sample;

and determining the target loss value according to the biased click rate and the real click rate corresponding to the media content sample.

6. The method of claim 5, wherein determining the biased click rate for the media content sample based on the position bias information and the non-biased click rate for the media content sample comprises:

7. The method as recited in claim 1, further comprising:

determining the duty ratio of the retained features in the original feature set;

judging whether training of the feature selection module is finished or not according to the proportion of the duty ratio and the discarding feature.

8. A media content click model training device, comprising:

9. An electronic device, the electronic device comprising:

one or more processors;

storage means for storing one or more programs,

the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the media content click model training method of any of claims 1-7.

10. A storage medium containing computer executable instructions which, when executed by a computer processor, are for performing the media content click model training method of any of claims 1-7.