CN111079000B - Content processing method, device, equipment and storage medium - Google Patents

Content processing method, device, equipment and storage medium Download PDF

Info

Publication number
CN111079000B
CN111079000B CN201911163681.2A CN201911163681A CN111079000B CN 111079000 B CN111079000 B CN 111079000B CN 201911163681 A CN201911163681 A CN 201911163681A CN 111079000 B CN111079000 B CN 111079000B
Authority
CN
China
Prior art keywords
pool
content
determining
interaction
index set
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201911163681.2A
Other languages
Chinese (zh)
Other versions
CN111079000A (en
Inventor
黄仕航
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Yayue Technology Co ltd
Original Assignee
Tencent Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tencent Technology Shenzhen Co Ltd filed Critical Tencent Technology Shenzhen Co Ltd
Priority to CN201911163681.2A priority Critical patent/CN111079000B/en
Publication of CN111079000A publication Critical patent/CN111079000A/en
Application granted granted Critical
Publication of CN111079000B publication Critical patent/CN111079000B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

The embodiment of the invention provides a content processing method, a device, equipment and a storage medium; the method comprises the following steps: determining a historical interaction index set between user behaviors and contents in a current content pool within a historical preset time; performing pool withdrawal on the content in the current content pool to obtain a candidate content pool; predicting the interaction indexes of the contents in the candidate content pool in the future time length according to the historical interaction index set to obtain a predicted interaction index set; if the predicted interaction index set meets a preset condition, determining the candidate content pool as a target content pool which can be recommended to a user; therefore, the target content pool which can be finally presented to the user can be determined by predicting the interaction index of the candidate content pool in the future time length, and the optimization efficiency of the content pool is improved.

Description

Content processing method, device, equipment and storage medium
Technical Field
The present invention relates to the field of data processing, and in particular, to a content processing method, apparatus, device, and storage medium.
Background
In the operation process of news information products, the problem that the content pool needs to be adjusted may be faced in stages, in the related technology, an assumption is made on the optimized content pool, the adjustment direction of the content is assumed to be beneficial to the products or have small influence, and the strategy of returning the content pool is continuously adjusted after online data is compared through an experiment in a heuristic mode; this adds a lot of uncertainty in the exploration process and increases the experimental period.
Disclosure of Invention
Embodiments of the present invention provide a content processing method, an apparatus, a device, and a storage medium, which can determine a target content pool that can be finally presented to a user by predicting an interaction index of a candidate content pool in a future time, so as to improve optimization efficiency of the content pool.
The technical scheme of the embodiment of the invention is realized as follows:
in a first aspect, an embodiment of the present invention provides a content processing method, including: determining a historical interaction index set between user behaviors and contents in a current content pool within a historical preset time; performing pool withdrawal on the content in the current content pool to obtain a candidate content pool; predicting the interaction indexes of the contents in the candidate content pool in the future time length according to the historical interaction index set to obtain a predicted interaction index set; and if the predicted interaction index set meets a preset condition, determining the candidate content pool as a target content pool which can be recommended to the user.
In a second aspect, an embodiment of the present invention provides a content processing apparatus, including: the first determination module is used for determining a historical interaction index set between the user behavior and the content in the current content pool within a historical preset time length; the first pool returning module is used for returning the content in the current content pool to obtain a candidate content pool; the first prediction module is used for predicting interaction indexes of the content in the candidate content pool in the future time length according to the historical interaction index set to obtain a predicted interaction index set; and the second determination module is used for determining the candidate content pool as a target content pool which can be recommended to the user if the predicted interaction index set meets a preset condition.
In a third aspect, an embodiment of the present invention provides a device for content processing, including: a memory for storing executable instructions; and the processor is used for realizing the content processing method when executing the executable instructions stored in the memory.
In a fourth aspect, an embodiment of the present invention provides a storage medium, which stores executable instructions for causing a processor to execute the method for processing content according to the embodiment of the present invention.
The embodiment of the invention has the following beneficial effects: the method comprises the steps that a candidate content pool is obtained by performing pool quitting on the content in a current content pool; then, based on the historical interaction index set, predicting the interaction indexes of the contents in the candidate content pool in the future time length; finally, under the condition that the predicted interaction index set meets preset conditions, taking the candidate content pool as a target content pool which can be recommended to a user, and applying the candidate content pool to an online experiment; therefore, in the process of obtaining the target content pool, the target content pool which can be finally presented to the user is determined by predicting the interaction index of the candidate content pool in advance, so that the target content pool which can be finally presented to the user can be determined without carrying out multiple actual online experiments, and the optimization efficiency of the content pool is improved.
Drawings
FIG. 1 is an alternative architectural diagram of a content processing system provided by an embodiment of the present invention;
FIG. 2A is a schematic diagram of an alternative architecture of a content processing system according to an embodiment of the present invention;
FIG. 2B is a schematic diagram of a content processing system according to an embodiment of the present invention;
fig. 3 is a schematic flow chart of an implementation of a content processing method according to an embodiment of the present invention;
fig. 4A is a schematic flow chart of another implementation of the content processing method according to the embodiment of the present invention;
fig. 4B is a schematic flow chart of another implementation of the content processing method according to the embodiment of the present invention;
FIG. 5 is a block diagram of a content processing system provided by an embodiment of the present invention;
FIG. 6 is another block diagram of a content processing system provided by an embodiment of the present invention;
fig. 7A is a schematic flow chart of another implementation of the content processing method according to the embodiment of the present invention;
FIG. 7B is a block diagram of the data preparation phase of the content processing process according to an embodiment of the present invention;
FIG. 7C is a block diagram of a prediction stage during content processing according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail with reference to the accompanying drawings, the described embodiments should not be construed as limiting the present invention, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.
In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.
In the following description, references to the terms "first \ second \ third" are only to distinguish similar objects and do not denote a particular order, but rather the terms "first \ second \ third" are used to interchange specific orders or sequences, where appropriate, to enable embodiments of the invention described herein to be practiced in other than the order shown or described herein.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used herein is for the purpose of describing embodiments of the invention only and is not intended to be limiting of the invention.
Before further detailed description of the embodiments of the present invention, terms and expressions mentioned in the embodiments of the present invention are explained, and the terms and expressions mentioned in the embodiments of the present invention are applied to the following explanations.
1) A content pool: is a concept in the field of content marketing, namely a system for loading and making content organically related. The content pool has three elements: one is content, two is content containers, and three is the interaction between content. For example, a content pond can be compared with a fish pond, the pond is a container, fish, water, aquatic plants, microorganisms and the like in the pond are all content in the pond, and the interaction of the content creates a valuable ecosystem of the fish pond. The content pool can allow a marketer to construct the content of the website like writing a book, and the purpose of providing the content is to provide value for network users so as to establish the professional degree and authority degree of the industry. Once the content is considered professional and authoritative, the relevant traffic will naturally go to the canal, which is the meaning of building a network content pool.
2) Adjusting property: brand tonality is a market impression based on the appearance of a brand, and is equivalent to human personality in terms of a brand personalization model. Colloquial is what consumers see or feel about brands. The brand identity is not obvious and is often hidden in specific brand expression, but the influence degree of the brand identity on the brand success or failure is far beyond the common imagination. If the brand identity violates the industry attribute, the brand cannot be removed, which is a potential rule for free market and is not diverted to the individual's mind. For example, the starbucks coffee is comfortable, leisure, free, etc. in terms of brand identity, which is conveyed by various factors such as decoration, service, products, etc. in the store.
3) Blockchain (Blockchain): an encrypted, chained transactional memory structure formed of blocks (blocks).
4) Block chain Network (Blockchain Network): the new block is incorporated into the set of a series of nodes of the block chain in a consensus manner.
In the related art, the rise of content information and the heat of fire have brought forward the concept of content as king, and new high-quality content and low-quality outdated content need to be continuously introduced and rejected in the operation process of the existing content information. In the process of clearing and rejecting the content, the tag attributes of the expiration time or the low quality of the article and the like are marked in a manual or machine mode, and then the content is cleared and rejected according to a certain strategy.
When the adjustability of the content pool is found to be inconsistent with the expectation in the product operation process, the adjustability of the content needs to be adjusted strategically or greatly, and at this time, the filtering and pool-returning processing needs to be performed on the existing article according to a certain policy. At this time, it is assumed that the overall adjustment of the content pool may cause a large change in the recommended content of the user, the content consumption index of the user may be affected to some extent, and positive is desirable, while negative needs to be avoided or reduced as much as possible. When the recommendable content is adjusted, the existing scheme is to make a strategy direction with feasible hypothesis, and verify whether the hypothesis is established or not by performing barrel-dividing experiments on users and comparing and observing the changes of an experimental barrel and a comparison barrel on a line. If yes, further expanding the influence of the assumed strategy to the users to the total number of users, and continuously iterating and increasing the adjustment force of the strategy direction to perform an experiment until the state of the content pool reaches an expectation; if the assumption is not true, a new assumption change strategy needs to be made again for experimental trial until a feasible pool quitting rule is found for carrying out pool quitting processing on the current content pool; therefore, the pool withdrawal experiment is directly carried out, so that a lot of uncertainty is increased in the exploration process, and the experiment period is increased.
In view of the foregoing technical problems, embodiments of the present invention provide a content processing method, apparatus, device, and storage medium, where in a scenario where content recommendation needs to be performed for a user, first, a historical interaction index set within a historical preset duration is determined for content in a current content pool; secondly, according to a prediction pool quitting rule, carrying out pool quitting on the content in the current content pool to obtain a candidate content pool; thirdly, predicting future interaction indexes of the content in the current content pool within future time length based on the historical interaction index set; predicting future interaction indexes of the contents in the candidate content pool within the future time length to obtain a future interaction index set; finally, estimating the influence of the future interaction index set on the candidate content pool, if the estimation result meets the condition, indicating that the content in the current content pool is correctly retreated, and recommending the retreated content pool to the user; therefore, the interaction indexes in the future time length are predicted for the contents in the candidate content pool, and whether the predicted interaction indexes meet the preset conditions or not is judged, the candidate content pool after the pool retreating is used as the target content pool which can be recommended to the user, so that the online experiment is not needed for multiple times, the final contents which can be subjected to the online experiment can be determined by predicting the interaction indexes in the future time length of the candidate content pool, and the optimization efficiency of the content pool is improved.
An exemplary application of the content processing device provided by the embodiment of the present invention is described below, and the content processing device provided by the embodiment of the present invention can be implemented as various types of user devices such as a notebook computer, a tablet computer, a desktop computer, a set-top box, a mobile device (e.g., a mobile phone, a portable music player, a personal digital assistant, a dedicated messaging device, and a portable game device), and can also be implemented as a server. In the following, an exemplary application will be explained when the device is implemented as a device or a server.
Referring to fig. 1, fig. 1 is an optional architecture diagram of a content processing system according to an embodiment of the present invention, and in order to implement supporting an exemplary application, first, when a content recommendation request sent by a terminal 10 is received, a historical interaction index set between a user behavior and content in a current content pool within a historical preset time is obtained for the content in the content pool 11; secondly, according to a preset pool quitting rule, performing pool quitting on the content in the content pool 11 to obtain a candidate content pool 12; thirdly, based on the historical interaction index set, predicting interaction indexes of the candidate content pool 12 in the future time length to obtain a predicted interaction index set 13; determining the candidate content pool 13 as a target content pool 14 which can be recommended to the user when the predicted interaction index set 13 meets the condition; therefore, in the process of obtaining the target content pool, the target content pool which can be finally presented to the user is determined by predicting the interaction index of the candidate content pool 12 in advance, so that the content which can be finally subjected to online experiment can be determined without carrying out multiple actual online experiments, and the optimization efficiency of the content pool is improved.
Referring to fig. 2A, fig. 2A is another alternative architecture diagram of a content processing system according to an embodiment of the present invention, which includes a blockchain network 20 (exemplarily showing a server 200 as a native node), a monitoring system 30 (exemplarily showing a device 300 belonging to the monitoring system 30 and a graphical interface 301 thereof), which are described below.
The type of blockchain network 20 is flexible and may be, for example, any of a public chain, a private chain, or a federation chain. Taking a public link as an example, electronic devices such as user equipment and servers of any service entity can access the blockchain network 20 without authorization; taking a federation chain as an example, an electronic device (e.g., a device/server) under the jurisdiction of a service entity after obtaining authorization may access the blockchain network 20, and at this time, become a special type of node in the blockchain network 20, i.e., a client node.
Note that the client node may provide only functionality to support the initiation of transactions by the business entity (e.g., for uplink storage of data or querying of data on the chain), and may be implemented by default or selectively (e.g., depending on the specific business requirements of the business entity) for the functions of the native nodes of the blockchain network 20, such as the ranking function, consensus service, ledger function, etc., described below. Therefore, the data and the service processing logic of the service subject can be migrated to the blockchain network 20 to the maximum extent, and the credibility and traceability of the data and service processing process are realized through the blockchain network 20.
Blockchain network 20 receives a transaction submitted by a client node (e.g., device 300 shown in fig. 2A as belonging to monitoring system 30) from a business entity (e.g., monitoring system 30 shown in fig. 2A), executes the transaction to update or query the ledger, and displays various intermediate or final results of executing the transaction on a user interface of the device (e.g., graphical interface 301 of device 300).
An exemplary application of the blockchain network is described below by taking monitoring system access to the blockchain network to implement uplink for content processing as an example.
The device 300 of the monitoring system 30 accesses the blockchain network 20 to become a client node of the blockchain network 20. The device 300 acquires a historical interaction index set between user behaviors and contents in the current content pool within a historical preset time through a sensor, and performs pool quitting on the contents in the current content pool according to a prediction pool quitting rule to obtain a candidate content pool; thirdly, predicting future interaction indexes of the content in the current content pool within future time length based on the historical interaction index set; predicting future interaction indexes of the contents in the candidate content pool within the future time length to obtain a future interaction index set; finally, estimating the influence of the future interaction index set on the candidate content pool to obtain a target content pool which can be recommended to the user; and, the target content pool is delivered to the server 200 in the blockchain network 20 or stored in the device 300; in the case where the upload logic has been deployed to the device 300 or the user has performed an operation, the device 300 generates a transaction corresponding to the update operation/query operation according to the to-be-processed item/synchronization time query request, specifies an intelligent contract to be called to implement the update operation/query operation and parameters transferred to the intelligent contract in the transaction, and the transaction also carries a digital signature signed by the monitoring system 30 (for example, a digest of the transaction is encrypted by using a private key in a digital certificate of the monitoring system 30), and broadcasts the transaction to the blockchain network 20. The digital certificate can be obtained by registering the monitoring system 30 with the certificate authority 31.
A native node in the blockchain network 20, for example, the server 200 verifies a digital signature carried by the transaction when receiving the transaction, and after the verification of the digital signature is successful, it is determined whether the monitoring system 30 has a transaction right according to the identity of the monitoring system 30 carried in the transaction, and any verification judgment of the digital signature and the right verification will result in a transaction failure. After successful verification, the native node signs its own digital signature (e.g., by encrypting a digest of the transaction using the native node's private key) and continues to broadcast in the blockchain network 20.
After the node with the sorting function in the blockchain network 20 receives the transaction successfully verified, the transaction is filled into a new block and broadcasted to the node providing the consensus service in the blockchain network 20.
The nodes in the blockchain network 20 that provide the consensus service perform a consensus process on the new block to reach agreement, the nodes that provide the ledger function append the new block to the end of the blockchain, and perform the transaction in the new block: updating the key value pair corresponding to the target content pool in the state database for the transaction submitting the processing result of the data to be processed; and for the transaction of inquiring the synchronization time, inquiring the key value pair corresponding to the synchronization time from the state database, and returning an inquiry result. The resulting synchronized time may be displayed in a graphical interface 301 of the device 300.
The native node in the blockchain network 20 may read the target content pool from the blockchain and present the target content pool on a monitoring page of the native node, and the native node may also perform pool returning on poor-quality content in the current content pool by using the current content pool stored in the blockchain, for example, perform pool returning on content in the current content pool according to a preset pool returning rule to obtain a candidate content pool; then, based on the historical interaction index set, predicting the interaction indexes of the contents in the candidate content pool in the future time length; finally, based on the predicted future interaction index set, determining the target content which needs to be finally returned to the pool so as to obtain a target content pool which can be recommended to the user; therefore, the content of the online experiment which can be finally carried out can be determined by predicting the interaction index of the candidate content pool in the future time length, the times of the online experiment required by the pool quitting rule and the influence on an online user are reduced, the period of the attempt of the pool quitting rule is shortened, and the optimization efficiency of the content pool is improved.
In practical applications, different functions may be set for different native nodes of the blockchain network 20, for example, the server 200 is set to have a content processing function and an accounting function, for example, the server analyzes a recommendation request uploaded by a device side, so as to perform a pool-dropping on part of content from a current content pool according to a preset pool-dropping rule to obtain a candidate content pool; and finally, determining a target content pool which can be recommended to the user by predicting the content in the candidate content pool and judging whether the predicted interaction index set meets a preset condition. For this situation, in the transaction process, the server 200 receives the recommendation request sent by the device 300, and determines a target content pool that can be recommended to the user from the current content pool based on the recommendation request by using the server 200; therefore, by predicting the interaction index of the candidate content pool in the future time length, the content of the online experiment which can be finally carried out is determined, the times of the online experiment required by the pool quitting rule and the influence on the online user are reduced, the period of the attempt of the pool quitting rule is shortened, and the optimization efficiency of the content pool is improved.
Referring to fig. 2B, fig. 2B is a schematic structural diagram of a content processing system according to an embodiment of the present invention, and the apparatus 400 shown in fig. 2B includes: at least one processor 410, memory 450, at least one network interface 420, and a user interface 430. The various components in device 400 are coupled together by a bus system 440. It is understood that the bus system 440 is used to enable communications among the components. The bus system 440 includes a power bus, a control bus, and a status signal bus in addition to a data bus. For clarity of illustration, however, the various buses are labeled as bus system 440 in FIG. 2B.
The processor 410 may be an integrated circuit chip having signal processing capabilities such as a general purpose processor, a digital signal processor, or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc., wherein the general purpose processor may be a microprocessor or any conventional processor, etc.
The user interface 430 includes one or more output devices 431, including one or more speakers and/or one or more visual displays, that enable the presentation of media content. The user interface 430 also includes one or more input devices 432, including user interface components that facilitate user input, in some examples, a keyboard, a mouse, a microphone, a touch screen display, a camera, other input buttons and controls.
The memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid state memory, hard disk drives, optical disk drives, and the like. Memory 450 optionally includes one or more storage devices physically located remote from processor 410.
The memory 450 includes either volatile memory or nonvolatile memory, and may include both volatile and nonvolatile memory. The nonvolatile Memory may be a Read Only Memory (ROM), and the volatile Memory may be a Random Access Memory (RAM). The memory 450 described in embodiments of the invention is intended to comprise any suitable type of memory.
In some embodiments, memory 450 is capable of storing data, examples of which include programs, modules, and data structures, or a subset or superset thereof, to support various operations, as exemplified below.
An operating system 451, including system programs for handling various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and handling hardware-based tasks;
a network communication module 452 for communicating to other computing devices via one or more (wired or wireless) network interfaces 420, exemplary network interfaces 420 including: bluetooth, wireless compatibility authentication, and Universal Serial Bus (USB), etc.;
a presentation module 453 for enabling presentation of information (e.g., user interfaces for operating peripherals and displaying content and information) via one or more output devices 431 (e.g., display screens, speakers, etc.) associated with user interface 430;
an input processing module 454 for detecting one or more user inputs or interactions from one of the one or more input devices 432 and translating the detected inputs or interactions.
In some embodiments, the apparatus provided by the embodiments of the present invention may be implemented in software, and fig. 2B illustrates a content processing server 455 stored in the memory 450, which may be software in the form of programs and plug-ins, and the like, and includes the following software modules: a first determination module 4551, a first pool removing module 4552, a first prediction module 4553 and a second determination module 4554; these modules are logical and thus may be combined or further split according to the functionality implemented. The functions of the respective modules will be explained below.
In other embodiments, the apparatus provided in the embodiments of the present invention may be implemented in hardware, and for example, the apparatus provided in the embodiments of the present invention may be a processor in the form of a hardware decoding processor, which is programmed to execute the content processing method provided in the embodiments of the present invention, for example, the processor in the form of the hardware decoding processor may be one or more Application Specific Integrated Circuits (ASICs), DSPs, Programmable Logic Devices (PLDs), Complex Programmable Logic Devices (CPLDs), Field-Programmable Gate arrays (FPGAs), or other electronic components.
The content processing method provided by the embodiment of the present invention will be described in conjunction with exemplary applications and implementations of the apparatus provided by the embodiment of the present invention.
Referring to fig. 3, fig. 3 is a schematic flow chart of an implementation of the content processing method according to the embodiment of the present invention, and is described with reference to the steps shown in fig. 3.
Step S301, determining a historical interaction index set between the user behavior and the content in the current content pool within a historical preset time length.
In some possible implementation manners, the interaction indicators may be aggregated according to the category of the content, so as to obtain a historical interaction indicator set. The historical preset duration may be the latest period of time, for example, the last 7 days; the set of historical interaction metrics includes at least: the amount of exposure (i.e., the number of times an advertisement (e.g., a picture file or a program) is exposed on a website in a certain time, for example, the number of times an advertisement is exposed refers to the number of times an advertisement (e.g., a picture file or a program) is successfully downloaded to a guest browser), the number of clicks (e.g., the number of times an advertisement is clicked by a user), the browsing duration (i.e., the length of time a content is browsed in a certain time, for example, the length of time an advertisement is browsed in a day), the number of comments (e.g., the number of times a user comments for a certain article) and the number of likes (e.g., the number of times a user likes for a certain article), etc. In a specific example, the exposure, the number of clicks, the browsing duration, the number of comments, the number of likes, and the like for the content in the current content pool within the last 7 days are determined.
Step S302, the content in the current content pool is returned to obtain a candidate content pool.
In some embodiments, performing pool quitting on the content in the current content pool according to a preset pool quitting rule to obtain a candidate content pool; the candidate content pool may also be understood as a candidate content pool obtained by dividing the content in the current content pool into a retired pool content and an unretired pool content, and retiring the retired pool content from the current content pool. The preset pool quitting rule may be set by referring to a historical interaction index set, or may be set by a user according to a requirement of the user, for example, the content of which the number of clicks is smaller than a certain threshold is used as the content of the pool quitting requirement, or the content of which the exposure amount is smaller than a certain threshold is used as the content of the pool quitting requirement, and the like.
Step S303, according to the historical interaction index set, predicting the interaction indexes of the contents in the candidate content pool in the future time length to obtain a predicted interaction index set.
In some embodiments, first, according to the historical interaction index set, predicting interaction indexes of the content in the current content pool in a future time length to obtain a first prediction index set; then, a second prediction index set corresponding to the content in the candidate content pool is determined from the first prediction index set, and the second prediction index set is determined as the predicted interaction index set. In brief, firstly, predicting interaction indexes of the content in the current content pool in the future time length by adopting a time series prediction mode based on a historical interaction index set; then, determining a predicted interaction index set corresponding to the content in the candidate content pool from the interaction indexes; therefore, the interaction index of the content in the candidate content pool in the future time length is determined by predicting the interaction index of the content in the current content pool in the future time length, and whether the candidate content pool can be used as the target content pool is determined by judging whether the interaction index meets the preset condition or not, so that the online experiment frequency of the content pool is reduced, and the user experience is improved.
Step S304, if the predicted interaction index set meets a preset condition, determining the candidate content pool as a target content pool which can be recommended to the user.
In some embodiments, the set of predicted interaction metrics is determined to satisfy a preset condition if the set of predicted interaction metrics indicates that the impact of content in the candidate content pool is positive. In a specific example, indexes such as exposure, click number and browsing duration in the predicted interaction index set are all relatively large, which indicates that the user is very required for the contents, that is, the influence of the contents in the candidate content pool is positive, and the candidate content pool is a target content pool which can be recommended to the user; therefore, the interaction indexes in the future time length are predicted for the contents in the candidate content pool, and whether the predicted interaction indexes meet the preset conditions or not is judged, the candidate content pool after the pool retreating is used as the target content pool which can be recommended to the user, so that the online experiment is not needed for multiple times, the final contents which can be subjected to the online experiment can be determined by predicting the interaction indexes in the future time length of the candidate content pool, and the optimization efficiency of the content pool is improved.
In some embodiments, the rule for retiring the content of the current content pool may be implemented by:
firstly, according to the historical interaction index set, determining the historical change trend of the historical interaction index.
For example, the change trend within the current content pool within the 7 days is determined according to the historical interaction index set of the last 7 days.
And then, determining the preset pool quitting rule based on the historical change trend.
For example, if the historical trend indicates that the click rate, exposure or browsing duration of the content of category a is gradually decreasing in the 7 days, the pool quitting rule may be set as: the content with the click quantity smaller than a certain threshold value is the content of the quitting pool; and/or the content of which the exposure is less than a certain threshold value is the content of the returned pool; and/or the content with the browsing time length less than a certain threshold value is the content of the quit pool, and the like.
And finally, the pool of the pool to be returned in the current content pool, which is matched with the preset pool returning rule, is returned to obtain the candidate content pool.
For example, if the preset pool returning rule is that the content with the click volume smaller than a certain threshold is the pool returning content, finding out the content to be returned with the click volume smaller than the certain threshold from the current content pool, marking the pool returning label on the content to be returned, and returning the content containing the pool returning label from the current content pool to obtain a candidate content pool; therefore, by adopting an assumed mode, a pool quitting label is marked on some contents, the contents are quitted from the current content pool to obtain a candidate content pool, and online experiments are not required to be directly carried out according to a pool quitting rule.
In some embodiments, in order to predict the accuracy of the interaction index of the content in the candidate content pool in the future time period, the step S304 may be implemented by the following steps, and fig. 4A is a schematic flow chart of another implementation of the content processing method provided by the embodiment of the present invention, which is described below with reference to fig. 3:
step S401, determining a future change trend of the interaction index in the candidate content pool within the future time length according to the predicted interaction index set and the first prediction index set.
In some embodiments, the set of predicted interaction metrics includes: click rate, exposure, browsing duration and the like; the first set of predictors includes: click rate, exposure, browsing duration, etc. For example, the future variation trend of the interaction index in the candidate content pool within the future 7 days is determined based on the predicted click rate, exposure and browsing duration of the content in the candidate content pool for the future 7 days and the predicted click rate, exposure and browsing duration of the content in the current content pool for the future 7 days.
Step S402, if the future variation trend meets the preset condition, determining the candidate content pool as a target content pool which can be recommended to the user.
In one specific example, if the future trend indicates a forward trend (e.g., a gradually increasing click rate, exposure or browsing duration), it is determined that a preset condition is satisfied, indicating that the content in the candidate content pool is required by the user, the candidate content pool is taken as a target content pool that can be recommended to the user.
In some embodiments, in order to improve the accuracy of predicting the future variation trend, the step S401 may be implemented by the following steps, and fig. 4B is a flowchart of another implementation of the content processing method provided by an embodiment of the present invention, which is described below with reference to fig. 3:
step S421, determining a ratio between the click rate and the corresponding exposure of each category of content in the predicted interaction index set to obtain a first ratio set.
For example, the content in the candidate content pool has 30 categories, and the ratio ctr1(pv/exp) of the click rate (pv) and the exposure (exp) of the content in each of the 30 categories is determined, respectively, to obtain a first ratio set.
Step S422, determining a ratio between the browsing duration of each category of content in the predicted interaction index set and the corresponding click rate, to obtain a second ratio set.
For example, the content in the candidate content pool has 30 categories, and the ratio time _ p _ pv of the browsing duration (time) to the click rate (pv) of the content in each of the 30 categories is determined1(time/pv), a second set of ratios is obtained.
Step S423, determining the future variation trend according to the first set of prediction indexes, the first set of ratios, and the second set of ratios.
In some embodiments, the step S423 may be implemented by:
the method comprises the following steps of firstly, multiplying the exposure of the content in the first prediction index set by the corresponding first ratio in the first ratio set to obtain the target click rate corresponding to each first ratio so as to obtain a target click rate set.
For example, the exposure (exp) to the first ratio (ctr) for each category of content1) Multiplying to obtain the target click quantity pv corresponding to each ratio1(exp*ctr1). In this way, the exposure of the content in the first prediction index set is used as the predicted exposure of the whole content pool after the content pool is predicted to be removed, and since the exposure is not greatly influenced by the content pool in a short period, it is assumed here that the exposure of the removed content is distributed to other content in the corresponding category, and the accuracy of the predicted change trend is not influenced.
And step two, multiplying each target click quantity by a second ratio in the corresponding second ratio set to obtain a target browsing time length corresponding to each target click quantity so as to obtain a target browsing time length set.
For example, target click volume (pv) in target click volume set1) And a second ratio (time _ p _ pv) in the second ratio set1) Multiplying to predict target browsing time1(ii) a Thus, by predicting the interaction index of the content in the candidate content poolThe change trend of the content pool after the content pool is withdrawn in the future time is predicted according to the interaction indexes of the content in the current content pool, so that the influence of the withdrawal of the pool on each index of content consumption can be quantitatively evaluated.
And thirdly, determining the future change trend according to the target click quantity set and the target browsing duration set.
In some possible implementation manners, first, averaging the target click volumes in the target click volume set to obtain a first average value; for example, using the formula avg ((pv)1-pv)/pv) averaging the target click rates to obtain a first average value; then, averaging the target browsing duration in the target browsing duration set to obtain a second average value; for example, using the formula avg ((time)1-time)/time), averaging the target point browsing duration to obtain a second average value; and finally, determining the first average value and the second average value as the future variation trend.
In some embodiments, after obtaining the first average value and the second average value, the process of determining whether the candidate content pool can be used as the target content pool is as follows:
and if the first average value and the second average value in the variation trend are both larger than 0, determining that the variation trend meets a preset condition, and determining that the candidate content pool is a target content pool which can be recommended to the user. Here, if the first average value and the second average value in the variation trend are both greater than 0, it is described that the pool quitting rule has a positive influence on the correlation index, and the larger the absolute value is, the greater the influence is, and it is further described that the content in the candidate content pool is required by the user, so that the candidate content pool is taken as a target content pool that can be recommended to the user.
In other implementations, if the first average or the second average in the trend is less than or equal to 0, determining a difference between the average less than 0 and 0; firstly, updating the preset pool quitting rule according to the difference value; and then, the pool of the pool to be quitted which is matched with the updated preset pool quitting rule in the current content pool is quitted to obtain an updated candidate content pool. Here, if the first average value or the second average value in the variation trend is less than or equal to 0, it indicates that the pool quitting rule has a negative influence on the relevant index, and the larger the absolute value is, the larger the influence is, further indicates that the content in the candidate content pool is not required by the user; therefore, readjusting the preset pool quitting rule based on the difference value, and then performing pool quitting on the content in the current content pool based on the updated preset pool quitting rule to obtain an updated candidate content pool; finally, the change trend of the interaction indexes of the content in the updated candidate content pool in the future time length is estimated, whether the estimated result meets the preset condition or not is judged again (namely whether the first average value and the second average value in the estimated future change trend are larger than 0 or not) until the future change trend meets the preset condition, the updating of the pool quitting rule is stopped, and the target content pool which can be recommended to the user is obtained; therefore, the number of times of on-line experiments of the pool quitting rule and the influence on-line users are reduced, the period of the attempt of the pool quitting rule is shortened, and the optimization efficiency of the content pool is improved.
An exemplary application of the embodiment of the present invention in an actual application scenario will be described below, taking the example of recommending a content pool for a user in news information as an example.
Fig. 5 is a block diagram of a content processing system according to an embodiment of the present invention, and as shown in fig. 5, the content processing system includes the following modules: a data evaluation module 501, a pool-quitting policy module 502 and an online experiment module 503, wherein:
the data evaluation module 501 is configured to evaluate data of the current content pool in an online experiment process, and feed back an evaluation result to the pool returning policy module 502.
For example, the click rate, browsing duration, exposure and the like of the content in the online experiment process of the current content pool are evaluated.
The pool returning strategy module 502 is configured to set a pool returning rule according to the evaluation result, and feed the pool returning rule back to the online experiment module 503.
The online experiment module 503 is configured to perform pool dropping on the content in the current content pool based on a pool dropping rule, apply the content pool after the pool dropping to an actual application, and feed back data generated in the actual application to the data evaluation module 501.
In some embodiments, after receiving the data fed back by the online experiment module 503, the data evaluation module 501 feeds back the data to the pool quitting policy module 502 again to change the set pool quitting rule, and performs the online experiment again, that is, performs the experiment attempt again by making a new pool quitting rule again until a feasible pool quitting rule is found to perform the pool quitting processing on the current content pool; in this way, a hypothetical pool quitting rule is firstly made for the optimized content pool, the adjustment direction of the content is assumed to be beneficial to the product or have small influence on the product, and the online data is compared and then continuously adjusted through a heuristic gradual experiment; this increases many uncertainties in the exploration process, increases the experimental period, and restricts the overall rhythm of the adjustment.
Based on this, the embodiment of the invention provides a content processing method, firstly, the change trend of each index of a content pool in the next several days is estimated; secondly, assuming that the exposure of the content is kept unchanged, the indexes (such as click rate, single-chapter reading duration and the like) of articles which do not leave the pool are used for replacing the overall indexes to calculate the change trends of the indexes such as click rate, browsing duration and the like; thirdly, the influence of the content pool after the pool quitting on the user is estimated through comparison of the change trends of the indexes before and after the pool quitting; therefore, by predicting the trend and influence of content interaction indexes of the pre-retreated tank content after the tank retreating treatment for several days in the future, the tank retreating rule can be adjusted under the condition of not performing an online experiment, and the content which accords with the expected direction is selected for performing the tank retreating experiment; therefore, the time period and uncertainty of experiment exploration are reduced, the rhythm of the experiment is accelerated, and the efficiency and controllability are improved.
Fig. 6 is another block diagram of a content processing system according to an embodiment of the present invention, as shown in fig. 6, the content processing system includes the following modules: a data evaluation module 601, a pool-quitting strategy module 602, an influence estimation module 603 and an online experiment module 604, wherein:
the data evaluation module 601 is configured to obtain a historical interaction index set between a user behavior and content in the current content pool within a historical preset time, and feed back the historical interaction index set to the pool returning policy module 602.
A pool quitting policy module 602, configured to formulate a preset pool quitting rule based on the historical interaction index set, and adopt the preset pool quitting rule to perform pool quitting on the content in the current content pool, so as to obtain a candidate content pool; then, the candidate content pool is fed back to the influence estimation module 603. In some embodiments, the pool exit policy module 602 is further configured to formulate a preset pool exit rule according to the input instruction.
The influence estimation module 603 is configured to estimate an influence of the candidate content pool, and feed back an estimation result to the pool returning policy module 602.
In some embodiments, the influence estimation module 603 estimates the influence of the indicator in the candidate content pool, for example, after the current content pool is returned for an article, the content pool after the pool is returned is estimated, and if the estimation result is not good, the estimation result is fed back to the pool returning policy module 602, so that the pool returning policy module 602 readjusts the pool returning rule, and if the effect is positive, the candidate content pool is fed back to the online experiment module 604.
The online experiment module 604 is configured to recommend the candidate content pool as a target content pool to the user, and feed back the interaction index in the target content pool to the data evaluation module 601.
In the embodiment of the invention, an influence estimation process is added in a pool quitting strategy implementation framework, before an online experiment is not carried out, the influence of a pool quitting rule on an interaction index is iterated and previewed in advance and analyzed, the pool quitting rule is adjusted continuously, the online experiment is implemented until the estimated influence is in an acceptable range or a forward result, the experiment is guided to develop in a controllable and forward direction, and the purpose of optimizing a content pool and improving the quality of the content pool is achieved while the influence of the rule experiment process on user experience is reduced.
Fig. 7A is a schematic flow chart of another implementation of the content processing method according to the embodiment of the present invention, and as shown in fig. 7A, the method may be implemented by the following steps:
step S701, acquiring a historical interaction index set of the content in the current content pool within a historical preset time.
In some embodiments, in order to adapt the content processing influence estimation method to the refined operation, the categories of the content in the content pool are divided, and a historical interaction index set is generated according to the categories of the content. In a specific example, according to historical day-level data of each index (exposure, click rate, reading time and the like) of article granularity, different categories of exposure (exp), click number (pv), browsing time (time) and the like are aggregated according to categories, and historical interaction index sets such as ctr (pv/exp), time _ p _ pv (time/pv) and the like are set.
Step S702, a preset pool quitting rule is formulated.
For example, an interaction index of an article in the last day is taken, a quit pool label is marked on the article which corresponds to the article needing to quit the pool according to a quit pool rule, and a quit pool article set and a non-quit pool article set are obtained respectively.
Step S703, obtaining interaction index sets of the retired pool content and the unreleased pool content respectively.
In some embodiments, the pool content returning and the pool content not returning are aggregated according to the category and the pool returning label to obtain ctr and time _ p _ pv indexes of the pool returning label classified respectively, namely, the pool returning content and the pool content not returning conversion interaction index, and in the estimation of each index of the content pool (namely, candidate content) after the pool returning, the conversion interaction index using the content not returning is used as the conversion interaction index of the whole content pool to estimate the whole index condition. As shown in fig. 7B, the indexes such as the content pool article category information in the module 721, the recent exposure, the click rate, and the browsing duration of the content pool article are aggregated according to categories, so as to obtain the interaction indexes (i.e., exp, pv, time, ctr, and time _ p _ pv) of each category in the content pool in the module 722. Then, labeling the content needing to be returned in the content pool of the latest day according to a preset pool returning rule to obtain indexes such as category information of the content pool, a pool returning label and the exposure, click rate, browsing duration and the like of the content in the content pool in the latest day in the module 723; then, according to the category of the content and whether to print a quit pool label for aggregation, conversion type interaction indexes (ctr and time _ p _ pv) of the quit pool content and the unremoved pool content in the module 724 are obtained.
The above steps S701 to S703 provide the steps required in the data preparation phase, and obtain the interaction index sets of the retired pool content and the unreleased pool content, so as to be used in the subsequent estimation process of the index of the entire content pool, that is, in the phase of applying to the interaction index prediction.
In step S704, a first prediction index set is predicted.
In some embodiments, based on a historical interaction index set, a time series prediction model is adopted to predict interaction indexes of the content in the current content pool in a future time length, so as to obtain a first prediction index set; therefore, the change trend of the content consumption index in the next days can be estimated according to the historical data. As shown in fig. 7C, for the historical interaction index sets (exp, pv, and time) in the module 731, a seasonal Autoregressive Moving Average model (ARIMA) prediction algorithm is used to realize prediction of each index, so as to obtain the predicted indexes (i.e., the first set of predicted indexes) such as exp, pv, and time in the module 732. Thus, the interactive indexes of each historical day form a time series index, and a time series index prediction algorithm can be adopted to predict the condition of each interactive index (mainly comprising exp, ctr and time) in the next few days. The method considers the habit of the user, the time of using the content information by the user is influenced by the work period to a certain extent, and the short-term index time sequence has periodicity with a period of 7 days, so that the prediction of each index is realized by adopting a seasonal ARIMA time sequence prediction algorithm. In this embodiment, the SARIMA model is used to predict the change trend of the interaction index in the next several days, and the model is used as one of time series predictions, and other time series prediction methods may be used instead, for example, a Holt linear trend method, a Holt-winter seasonal prediction model, or a simpler simple average model or a moving average model.
Step S705, a pool withdrawal index prediction algorithm is formulated.
In some embodiments, as shown in fig. 7C, based on the first set of prediction indexes obtained in module 732 (i.e. the predicted interaction indexes exp, pv and time of the non-retired content pool) and the conversion-class interaction indexes (ctr and time _ p _ pv) of the retired content and the non-retired content in module 733, the interaction indexes in the retired content pool are predicted according to the established retired pool prediction algorithm to obtain the predicted interaction indexes (exp and time _ p _ pv) in the candidate content pool in module 7341、pv1And time1) Respectively using exp1、pv1And time1And (4) showing. Since exp is not greatly influenced by the content pool in a short time and the category relationship between the recommendation and the content itself is large, it is assumed here that the exposure of the retired pool content will be distributed to other contents in the corresponding category, ctr and time _ p _ pv of unretired pool content are used as the overall conversion category interaction indexes of each category of the content pool in the next few days, ctr is used respectively1And time _ p _ pv1Expressed, the pool exit prediction algorithm is shown in equation (1):
Figure BDA0002286841510000191
step S706, determining a set of predicted interaction indexes of unremoved pool content.
In some embodiments, a formulated pool returning index prediction algorithm is adopted, and interaction indexes of the contents which do not return from the pool are predicted based on the interaction index sets of the contents which return from the pool and the contents which do not return from the pool, which are obtained in the step S703, so that a predicted interaction index set is obtained; therefore, the change trend of each index of content consumption in the next few days after the withdrawal of the pool is estimated is realized.
The above steps S704 to S706 provide the steps required in the interactive index prediction stage, and obtain the predicted interactive index set and the first prediction index set of the unremoved pool content, so as to be used in the subsequent process of predicting the influence on the entire content pool, that is, in the stage of influence prediction.
In step S707, a preset influence estimation rule is established.
In some embodiments, a preset influence estimation rule is adopted, and influence estimation is performed on the candidate content pool based on the first prediction index set and the predicted interaction index set; in a specific example, the influence estimation rule is mainly performed for pv and time indexes, a change rate is used to measure the influence estimation result, e is used to represent the estimation result, and avg represents an average value of a change trend of several days in the future, as shown in formula (2):
(epv，etime)＝(avg((pv1-pv)/pv),avg((time1-time)/time)) (2)；
step S708, influence estimation is performed on the candidate content pool.
In some embodiments, the influence estimation is performed on the candidate content pool by using the preset influence estimation rule obtained in step S707 to obtain an estimation result. For example, in formula (2), when e is positive, it indicates that the pool receding rule has a positive influence on the relevant index, and the larger the absolute value is, the larger the influence is; when e is negative, the pool quitting rule is indicated to have negative influence on the relevant indexes, and the larger the absolute value is, the larger the influence is. When the estimated influence of the pool returning rule is negative and large, the pool returning rule needs to be adjusted, and when the estimated influence is positive, an online experiment can be carried out, and real data is used for further verification; therefore, the influence of the pool returning on each index of content consumption is quantitatively evaluated, so that a relatively suitable content processing rule is selected to provide a quantitative index reference, the times of on-line experiments of the pool returning bone and the influence on-line users are reduced, the trial period of the pool returning rule is shortened, and the optimization efficiency of the content pool is improved.
The step S707 and the step S708 provide steps required for the influence estimation stage, and influence estimation is performed on the candidate content pool by using a preset influence estimation rule to obtain an estimation result; and finally, determining whether the candidate content pool can be used as a target content pool or not based on the estimation result, so that the target content pool which can be finally presented to the user can be determined by estimating the influence in the candidate content pool without a plurality of online experiments, and the times and the period of the online experiments are reduced.
Continuing with the exemplary structure of the content processing server 455 provided by the embodiments of the present invention as implemented as software modules, in some embodiments, as shown in fig. 2B, the software modules stored in the content processing server 455 of the memory 450 may include: a first determining module 4551, configured to determine a historical interaction index set between a user behavior and content in a current content pool within a historical preset time; a first pool returning module 4552, configured to return the content in the current content pool to obtain a candidate content pool; a first prediction module 4553, configured to predict, according to the historical interaction index set, interaction indexes of the content in the candidate content pool within a future time duration, so as to obtain a predicted interaction index set; a second determining module 4554, configured to determine the candidate content pool as a target content pool that can be recommended to the user if the predicted interaction index set satisfies a preset condition.
In some embodiments, the first determining module 4551 is further configured to: determining the historical change trend of the historical interaction indexes according to the historical interaction index set; determining a preset pool quitting rule based on the historical change trend; correspondingly, the first pool returning module 4552 is further configured to return the pool of the to-be-returned pool, which is matched with the preset pool returning rule, in the current content pool to obtain the candidate content pool.
In some embodiments, the first prediction module 4553 is further configured to: predicting the interaction indexes of the content in the current content pool in the future time length according to the historical interaction index set to obtain a first prediction index set; and determining a second prediction index set corresponding to the content in the candidate content pool from the first prediction index set, and determining the second prediction index set as the predicted interaction index set.
In some embodiments, the second determining module 4554 is further configured to: determining a future change trend of the interaction indexes in the candidate content pool within the future time length according to the predicted interaction index set and the first prediction index set; and if the future change trend meets a preset condition, determining the candidate content pool as a target content pool which can be recommended to the user.
In some embodiments, the second determining module 4554 is further configured to: determining the ratio of the click rate of the content of each category in the predicted interaction index set to the corresponding exposure amount to obtain a first ratio set; determining the ratio of the browsing duration of each category of content in the predicted interaction index set to the corresponding click rate to obtain a second ratio set; determining the future trend of change according to the first set of prediction indexes, the first set of ratios and the second set of ratios.
In some embodiments, the second determining module 4554 is further configured to: multiplying the exposure of the content in the first prediction index set by the corresponding first ratio in the first ratio set to obtain a target click rate corresponding to each first ratio so as to obtain a target click rate set; multiplying each target click quantity by a second ratio in the corresponding second ratio set to obtain a target browsing time length corresponding to each target click quantity so as to obtain a target browsing time length set; and determining the future change trend according to the target click quantity set and the target browsing duration set.
In some embodiments, the second determining module 4554 is further configured to: averaging the target click quantities in the target click quantity set to obtain a first average value; averaging the target browsing duration in the target browsing duration set to obtain a second average value; and determining the first average value and the second average value as the future variation trend. Embodiments of the present invention provide a storage medium storing executable instructions, which when executed by a processor, will cause the processor to perform the method provided by embodiments of the present invention.
In some embodiments, the storage medium may be a memory such as a flash memory, a magnetic surface memory, an optical disk, or an optical disk memory; or may be various devices including one or any combination of the above memories. In some embodiments, executable instructions may be written in any form of programming language (including compiled or interpreted languages), in the form of programs, software modules, scripts or code, and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. By way of example, executable instructions may, but need not, correspond to files in a file system, and may be stored in a portion of a file that holds other programs or data, such as in one or more scripts in a hypertext Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub-programs, or portions of code). By way of example, executable instructions may be deployed to be executed on one in-vehicle computing device or on multiple computing devices located at one site or distributed across multiple sites and interconnected by a communication network. In summary, in the process of recommending the content that the user feels like for the user, in the embodiment of the present invention, first, for the content in the current content pool, a historical interaction index set within a historical preset duration is determined; secondly, according to a prediction pool quitting rule, carrying out pool quitting on the content in the current content pool to obtain a candidate content pool; thirdly, predicting future interaction indexes of the content in the current content pool within future time length based on the historical interaction index set; predicting future interaction indexes of the contents in the candidate content pool within the future time length to obtain a future interaction index set; finally, estimating the influence of the future interaction index set on the candidate content pool, if the estimation result meets the condition, indicating that the content in the current content pool is correctly retreated, and recommending the retreated content pool to the user; therefore, the interaction indexes of the content in the content pool in the future time length are predicted, the content pool after the content pool is withdrawn is predicted based on the interaction indexes in the future, and the candidate content pool after the content pool is withdrawn is used as a target content pool which can be recommended to a user under the condition that the prediction result meets the condition, so that the content which can be finally subjected to online experiment can be determined by predicting the interaction indexes of the candidate content pool in the future time length without multiple online experiments, and the optimization efficiency of the content pool is improved.
The above description is only an example of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, and improvement made within the spirit and scope of the present invention are included in the protection scope of the present invention.

Claims (10)

1. A method for content processing, the method comprising:
determining a historical interaction index set between user behaviors and contents in a current content pool within a historical preset time;
determining unremoved pool contents in the current content pool and pool contents to be returned containing a pool returning label according to a preset pool returning rule;
withdrawing the pool content from the current content pool to obtain a candidate content pool;
predicting interaction indexes of unremoved pool contents in the current content pool in future time according to the historical interaction index set to obtain a first prediction index set;
acquiring a predicted interaction index set from the first prediction index set;
and if the predicted interaction index set meets a preset condition, determining the candidate content pool as a target content pool which can be recommended to the user.
2. The method according to claim 1, wherein before the determining of the unremoved pool content and the to-be-retired pool content in the current content pool according to a preset retirement rule, the method further comprises:
determining the historical change trend of the historical interaction indexes according to the historical interaction index set;
determining the preset pool quitting rule based on the historical change trend;
correspondingly, the pool of the pool to be quitted in the current content pool matched with the preset pool quitting rule is quitted, and the candidate content pool is obtained.
3. The method according to claim 1 or 2, wherein the obtaining of the predicted interaction metric set from the first prediction metric set comprises:
and determining a second prediction index set corresponding to the content in the candidate content pool from the first prediction index set, and determining the second prediction index set as the predicted interaction index set.
4. The method according to claim 3, wherein the determining that the candidate content pool is a target content pool that can be recommended to the user if the set of predicted interaction metrics satisfies a preset condition comprises:
determining a future change trend of the interaction indexes in the candidate content pool within the future time length according to the predicted interaction index set and the first prediction index set;
and if the future change trend meets a preset condition, determining the candidate content pool as a target content pool which can be recommended to the user.
5. The method of claim 4, wherein determining a future trend of interaction indicators in the candidate content pool over the future time period according to the predicted set of interaction indicators and the first set of prediction indicators comprises:
determining the ratio of the click rate of the content of each category in the predicted interaction index set to the corresponding exposure amount to obtain a first ratio set;
determining the ratio of the browsing duration of each category of content in the predicted interaction index set to the corresponding click rate to obtain a second ratio set;
determining the future trend of change according to the first set of prediction indexes, the first set of ratios and the second set of ratios.
6. The method of claim 5, wherein determining the future trend from the first set of predictors, the first set of ratios, and the second set of ratios comprises:
multiplying the exposure of the content in the first prediction index set by the corresponding first ratio in the first ratio set to obtain a target click rate corresponding to each first ratio so as to obtain a target click rate set;
multiplying each target click quantity by a second ratio in the corresponding second ratio set to obtain a target browsing time length corresponding to each target click quantity so as to obtain a target browsing time length set;
and determining the future change trend according to the target click quantity set and the target browsing duration set.
7. The method of claim 6, wherein determining the future trend according to the set of target click volumes and the set of target browsing durations comprises:
averaging the target click quantities in the target click quantity set to obtain a first average value;
averaging the target browsing duration in the target browsing duration set to obtain a second average value;
and determining the first average value and the second average value as the future variation trend.
8. A content processing apparatus, characterized in that the apparatus comprises:
the first determination module is used for determining a historical interaction index set between the user behavior and the content in the current content pool within a historical preset time length;
the first pool returning module is used for determining the unremoved pool contents in the current content pool and the to-be-returned pool contents containing the pool returning labels according to a preset pool returning rule; withdrawing the pool content from the current content pool to obtain a candidate content pool;
the first prediction module is used for predicting interaction indexes of unremoved pool contents in the current content pool in future time according to the historical interaction index set to obtain a first prediction index set;
acquiring a predicted interaction index set from the first prediction index set;
and the second determination module is used for determining the candidate content pool as a target content pool which can be recommended to the user if the predicted interaction index set meets a preset condition.
9. An apparatus for content processing, comprising:
a memory for storing executable instructions;
a processor for implementing the content processing method of any one of claims 1 to 7 when executing executable instructions stored in the memory.
10. A computer-readable storage medium having stored thereon executable instructions for causing a processor to perform the content processing method of any one of claims 1 to 7 when executed.
CN201911163681.2A 2019-11-25 2019-11-25 Content processing method, device, equipment and storage medium Active CN111079000B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911163681.2A CN111079000B (en) 2019-11-25 2019-11-25 Content processing method, device, equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911163681.2A CN111079000B (en) 2019-11-25 2019-11-25 Content processing method, device, equipment and storage medium

Publications (2)

Publication Number Publication Date
CN111079000A CN111079000A (en) 2020-04-28
CN111079000B true CN111079000B (en) 2021-10-19

Family

ID=70311509

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911163681.2A Active CN111079000B (en) 2019-11-25 2019-11-25 Content processing method, device, equipment and storage medium

Country Status (1)

Country Link
CN (1) CN111079000B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114064445A (en) * 2020-08-04 2022-02-18 腾讯科技（深圳）有限公司 Test method, device, equipment and computer readable storage medium
CN113656681B (en) * 2021-07-08 2023-08-11 北京奇艺世纪科技有限公司 Object evaluation method, device, equipment and storage medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106649884A (en) * 2017-01-11 2017-05-10 河南科技大学 Multimedia content recommendation method based on user situational analysis
CN107918622A (en) * 2016-10-10 2018-04-17 阿里巴巴集团控股有限公司 Commending contents, methods of exhibiting, client, server and system
CN110309418A (en) * 2018-04-26 2019-10-08 腾讯科技（北京）有限公司 Recommendation determines method, apparatus, storage medium and computer equipment
US10444957B1 (en) * 2011-10-11 2019-10-15 Google Llc Determining intent of a recommendation on a URL of a web page or advertisement
CN110457589A (en) * 2019-08-19 2019-11-15 上海新共赢信息科技有限公司 A kind of vehicle recommended method, device, equipment and storage medium
CN110473000A (en) * 2019-07-17 2019-11-19 深圳市元征科技股份有限公司 A kind of information recommendation method, server and storage medium

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105872629B (en) * 2016-03-18 2019-08-27 合一网络技术(北京)有限公司 Content recommendation method, apparatus and system
CN109033228A (en) * 2018-06-29 2018-12-18 北京百度网讯科技有限公司 A kind of determination method, apparatus, server and the storage medium of recommendation list
CN110415063B (en) * 2018-07-31 2022-07-05 北京京东尚科信息技术有限公司 Commodity recommendation method and device, electronic equipment and readable medium
CN109272360B (en) * 2018-09-28 2021-09-10 有米科技股份有限公司 Intelligent advertisement recommendation method, system and device
CN110263242B (en) * 2019-01-04 2023-10-20 腾讯科技（深圳）有限公司 Content recommendation method, content recommendation device, computer readable storage medium and computer equipment
CN110163686A (en) * 2019-05-27 2019-08-23 成都魔方城科技有限公司 Desired consumption portrait method and system based on consumer behaviour

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10444957B1 (en) * 2011-10-11 2019-10-15 Google Llc Determining intent of a recommendation on a URL of a web page or advertisement
CN107918622A (en) * 2016-10-10 2018-04-17 阿里巴巴集团控股有限公司 Commending contents, methods of exhibiting, client, server and system
CN106649884A (en) * 2017-01-11 2017-05-10 河南科技大学 Multimedia content recommendation method based on user situational analysis
CN110309418A (en) * 2018-04-26 2019-10-08 腾讯科技（北京）有限公司 Recommendation determines method, apparatus, storage medium and computer equipment
CN110473000A (en) * 2019-07-17 2019-11-19 深圳市元征科技股份有限公司 A kind of information recommendation method, server and storage medium
CN110457589A (en) * 2019-08-19 2019-11-15 上海新共赢信息科技有限公司 A kind of vehicle recommended method, device, equipment and storage medium

Also Published As

Publication number Publication date
CN111079000A (en) 2020-04-28

Similar Documents

Publication Publication Date Title
Bhagat et al. Buy it again: Modeling repeat purchase recommendations
CN104281961B (en) For the advertisement in on-line system and the quality score system of content
CN107918622B (en) Content recommendation and display method, client, server and system
US8224711B2 (en) Method and apparatus for creating contextualized auction feeds
WO2015135353A1 (en) Search method, system, and device
US20180285331A1 (en) Method, server, browser, and system for recommending text information
EP4181026A1 (en) Recommendation model training method and apparatus, recommendation method and apparatus, and computer-readable medium
TW201905736A (en) Information push method and system
JP2021103542A (en) Information providing device, information providing method, and program
CN106997262A (en) Client end interface loading control method, device, client and mobile terminal
CN111079000B (en) Content processing method, device, equipment and storage medium
US10678831B2 (en) Page journey determination from fingerprint information in web event journals
CN111767466A (en) Recommendation information recommendation method and device based on artificial intelligence and electronic equipment
US11734516B2 (en) Systems and methods to generate messages using machine learning on digital assets
CN109783727A (en) Retrieve recommended method, device, computer readable storage medium and electronic equipment
CN113536098A (en) Information recommendation method, device, equipment and computer readable storage medium
CN109063104A (en) Method for refreshing, device, storage medium and the terminal device of recommendation information
US20190065587A1 (en) Page journey determination from web event journals
CN104615452A (en) Information providing method and device
CN111159570A (en) Information recommendation method and server
CN106878772A (en) A kind of program commending method and device
KR20150020390A (en) Content recommendation method, and apparatas and system for providing the method
JP2014071749A (en) Content server device, program and communication system
JP2012256268A (en) Advertisement distribution device and advertisement distribution program
CN109242690A (en) Finance product recommended method, device, computer equipment and readable storage medium storing program for executing

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
REG Reference to a national code

Ref country code: HK

Ref legal event code: DE

Ref document number: 40022486

Country of ref document: HK

SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right

Effective date of registration: 20221123

Address after: 1402, Floor 14, Block A, Haina Baichuan Headquarters Building, No. 6, Baoxing Road, Haibin Community, Xin'an Street, Bao'an District, Shenzhen, Guangdong 518133

Patentee after: Shenzhen Yayue Technology Co.,Ltd.

Address before: 518000 Tencent Building, No. 1 High-tech Zone, Nanshan District, Shenzhen City, Guangdong Province, 35 Floors

Patentee before: TENCENT TECHNOLOGY (SHENZHEN) Co.,Ltd.

TR01 Transfer of patent right