CN113254843A - Information pushing method and device and storage medium - Google Patents

Information pushing method and device and storage medium Download PDF

Info

Publication number
CN113254843A
CN113254843A CN202110727241.6A CN202110727241A CN113254843A CN 113254843 A CN113254843 A CN 113254843A CN 202110727241 A CN202110727241 A CN 202110727241A CN 113254843 A CN113254843 A CN 113254843A
Authority
CN
China
Prior art keywords
information
recommended
target
score
information 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.)
Granted
Application number
CN202110727241.6A
Other languages
Chinese (zh)
Other versions
CN113254843B (en
Inventor
张应碧
郭瑞
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.)
Tencent Technology Shenzhen 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 CN202110727241.6A priority Critical patent/CN113254843B/en
Publication of CN113254843A publication Critical patent/CN113254843A/en
Application granted granted Critical
Publication of CN113254843B publication Critical patent/CN113254843B/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
    • G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
    • 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 Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

The embodiment of the application discloses an information pushing method, an information pushing device and a storage medium, wherein an information set is obtained; acquiring interactive feedback data of each content information in the information set, and performing positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set; acquiring an information set to be recommended; determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set; and sequencing and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended. Therefore, the information pushing efficiency is improved.

Description

Information pushing method and device and storage medium
Technical Field
The application relates to the technical field of internet, in particular to an information pushing method, an information pushing device and a storage medium.
Background
The appearance and popularization of the internet bring a great deal of Information to users, and the demand of the users on the Information in the Information age is met, but the quantity of the Information on the internet greatly increases along with the rapid development of the network, so that the users are difficult to obtain the Information which is really useful for the users when facing a great amount of Information, and the use efficiency of the Information is reduced on the contrary, which is the problem of so-called Information Overload. In order to solve the problem of information overload, technical personnel provide a recommendation system to recommend information, products and the like interested by a user to a personalized information recommendation system of the user according to the information demand, the interest and the like of the user, so that the use efficiency of the information is improved.
However, in the prior art, the measurement of positive and negative feedback effects does not have an accurate measurement index, and the effect is generally estimated by adopting the subjective experience of technicians or extracting recommendation samples and the like, so that the measurement of positive and negative feedback effects has the problems of strong subjectivity, untimely feedback and the like, the effective monitoring and iterative optimization of the recommendation system are not facilitated, and the information push efficiency is low.
Disclosure of Invention
The embodiment of the application provides an information pushing method, an information pushing device and a storage medium, which can improve the information pushing efficiency.
The embodiment of the application provides an information pushing method, which comprises the following steps:
acquiring an information set, wherein the information set comprises at least one piece of content information;
acquiring interactive feedback data of each content information in the information set, and performing positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set;
acquiring an information set to be recommended, wherein the information set to be recommended comprises at least one piece of content information to be recommended;
determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set;
and sequencing and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended.
Correspondingly, an embodiment of the present application provides an information pushing apparatus, including:
a first acquisition unit configured to acquire an information set, the information set including at least one piece of content information;
the acquisition unit is used for acquiring interactive feedback data of each content information in the information set and carrying out positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set;
the second obtaining unit is used for obtaining an information set to be recommended, and the information set to be recommended comprises at least one piece of content information to be recommended;
the determining unit is used for determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set;
and the pushing unit is used for carrying out sequencing adjustment on each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and pushing the target information set to be recommended.
In one embodiment, the determining unit includes:
a third obtaining subunit, configured to obtain first attribute information of each piece of content information to be recommended in the information set to be recommended, second attribute information of each piece of content information in the positive feedback information set, and third attribute information of each piece of content information in the negative feedback information set;
a first calculating subunit, configured to calculate matching degrees between the first attribute information and the second attribute information, and between the first attribute information and the third attribute information;
and the first determining subunit is configured to determine, according to the matching degree, a first matching score of the first attribute information and the second attribute information, and a second matching score of the first attribute information and the third attribute information.
In one embodiment, the pushing unit includes:
the second determining subunit is used for determining the maximum value of the first matching scores of each piece of content information to be recommended and each piece of content information in the positive feedback information set as a first target matching score;
the third determining subunit is used for determining the maximum value of the second matching scores of each piece of content information to be recommended and each piece of content information in the negative feedback information set as a second target matching score;
and the first sequencing subunit is used for carrying out sequencing adjustment on each piece of content information to be recommended according to the first target matching score and the second target matching score.
In an embodiment, the first sorting subunit is configured to:
weighting the first target matching score and the second target matching score corresponding to each piece of content information to be recommended to obtain a target score of each piece of content information to be recommended;
and sequencing each piece of content information to be recommended according to the target score.
In an embodiment, the information pushing apparatus further includes:
the second calculating unit is used for calculating the positive feedback measurement scores of the target information set to be recommended and the positive feedback information set according to the first matching score, wherein the target information set to be recommended comprises at least one piece of target content information to be recommended;
the third calculating unit is used for calculating negative feedback weighing scores of the target information set to be recommended and the negative feedback information set according to the second matching score;
and the first adjusting unit is used for adjusting the recommendation system based on the positive feedback weighing fraction and the negative feedback weighing fraction.
In one embodiment, the second computing unit includes:
the fourth determining subunit is configured to determine the content information quantity of the target to-be-recommended content information in the target to-be-recommended information set;
a fourth calculating subunit, configured to calculate, according to the first matching score and the content information quantity, a first coverage score of the target information set to be recommended and the positive feedback information set;
a fifth calculating subunit, configured to calculate first relevance scores of the target information set to be recommended and the positive feedback information set according to the first matching score and the number of content information;
and the first processing subunit is used for obtaining the positive feedback measurement scores of the target information set to be recommended and the positive feedback information set based on the first coverage score and the first relevance score.
In one embodiment, the fourth calculating subunit includes:
the first assignment module is used for assigning the first matching score to obtain a first assigned matching score;
the first accumulation module is used for accumulating the first assignment matching scores to obtain a first coverage score of each target content information to be recommended and the positive feedback information set;
the second accumulation module is used for accumulating the first coverage scores corresponding to each target content information to be recommended to obtain a first coverage total score of the target information set to be recommended;
and the first average module is used for carrying out average processing on the first total coverage score according to the content information quantity to obtain a first coverage score of the target information set to be recommended and the positive feedback information set.
In an embodiment, the first assignment module is configured to:
detecting a numerical magnitude of the first matching score;
if the numerical value of the first matching score meets a preset condition, assigning the numerical value of the first matching score as a first preset score;
if the numerical value of the first matching score does not meet the preset condition, assigning the numerical value of the first matching score as a second preset score;
and obtaining a first assignment matching score based on the first preset score and the second preset score.
In an embodiment, the fifth calculating subunit is configured to:
acquiring the maximum value of the first matching score of each piece of target content information to be recommended and each piece of content information in the positive feedback information set to obtain the positive feedback target matching score of each piece of target content information to be recommended and the positive feedback information set;
accumulating the positive feedback target matching scores to obtain a positive feedback target matching total score;
and carrying out average processing on the total matching scores of the positive feedback targets according to the number of the content information to obtain a target information set to be recommended and a first relevance score of the positive feedback information set.
In one embodiment, the third computing unit includes:
the fifth determining subunit is configured to determine the content information quantity of the target to-be-recommended content information in the target to-be-recommended information set;
a sixth calculating subunit, configured to calculate a second coverage score of the target information set to be recommended and the negative feedback information set according to the second matching score and the number of content information;
a seventh calculating subunit, configured to calculate second relevance scores of the target information set to be recommended and the negative feedback information set according to the second matching score and the number of content information;
and the second processing subunit is used for obtaining negative feedback measurement scores of the target information set to be recommended and the negative feedback information set based on the second coverage score and the second relevance score.
In an embodiment, the sixth calculating subunit includes:
the second assignment module is used for assigning the second matching score to obtain a second assigned matching score;
the third accumulation module is used for accumulating the second assignment matching scores to obtain a second coverage score of each target content information to be recommended and the negative feedback information set;
the fourth accumulation module is used for accumulating the second coverage scores corresponding to each target content information to be recommended to obtain a second coverage total score of the target information set to be recommended;
and the second averaging module is used for carrying out averaging processing on the second total coverage score according to the content information quantity to obtain a second coverage score of the target information set to be recommended and the negative feedback information set.
In an embodiment, the second assigning unit is configured to:
detecting a numerical magnitude of the second matching score;
if the numerical value of the second matching score meets a preset condition, assigning the numerical value of the second matching score as a third preset score;
if the numerical value of the second matching score does not meet the preset condition, assigning the numerical value of the second matching score as a fourth preset score;
and obtaining a second assignment matching score based on the third preset score and the fourth preset score.
In an embodiment, the seventh computing subunit is configured to:
obtaining the maximum value of the second matching score of each target content information to be recommended and each content information in the negative feedback information set to obtain the negative feedback target matching score of each target content information to be recommended and the negative feedback information set;
accumulating the negative feedback target matching scores to obtain a negative feedback target matching total score;
and carrying out average processing on the negative feedback target matching total score according to the content information quantity to obtain a second relevance score of the target information set to be recommended and the negative feedback information set.
In addition, a storage medium is further provided, where multiple instructions are stored in the storage medium, and the instructions are suitable for being loaded by a processor to perform any of the steps in the information pushing method provided in the embodiments of the present application.
In addition, the embodiment of the present application further provides a computer device, which includes a processor and a memory, where the memory stores an application program, and the processor is configured to run the application program in the memory to implement the information push method provided in the embodiment of the present application.
Embodiments of the present application also provide a computer program product or a computer program, which includes computer instructions stored in a storage medium. The processor of the computer device reads the computer instructions from the storage medium, and executes the computer instructions, so that the computer device executes the steps in the information pushing method provided by the embodiment of the application.
The embodiment of the application acquires the information set; acquiring interactive feedback data of each content information in the information set, and performing positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set; acquiring an information set to be recommended; determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set; and sequencing and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended. Therefore, positive feedback information sets and negative feedback information sets are obtained by timely carrying out positive and negative feedback classification on the content information in the information sets, a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information sets and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information sets are calculated, and then sequencing adjustment is carried out on each piece of content information to be recommended according to the first matching score and the second matching score, so that the target information sets to be recommended are obtained and pushed more accurately in real time, and the information pushing efficiency is improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic view of an implementation scenario of an information push method provided in an embodiment of the present application;
fig. 2 is a schematic flowchart of an information pushing method according to an embodiment of the present application;
fig. 3 is a schematic flowchart of an information pushing method according to an embodiment of the present application;
fig. 4 is another specific flowchart of an information pushing method according to an embodiment of the present application;
fig. 5 is another schematic flowchart of an information pushing method provided in an embodiment of the present application;
FIG. 6 is a schematic structural diagram of an information pushing apparatus provided in an embodiment of the present application;
fig. 7 is a schematic structural diagram of a computer device provided in an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The embodiment of the application provides an information pushing method, an information pushing device and a storage medium. The information pushing apparatus may be integrated in a computer device, and the computer device may be a server or a terminal.
For a better understanding of the embodiments of the present application, reference is made to the following terms:
the recommendation system comprises: the system is a personalized information recommendation system which recommends information, products and the like which are interested by a user to the user according to the information requirements, interests and the like of the user, namely an algorithm service set which displays contents in a personalized manner on the basis of the characteristics of the user.
Brushing: a batch of ordered content lists provided by a single recommendation service. Is a unit of measure, e.g., 1 brush, 2 brushes, the present brush, the secondary brush, etc. The current brush refers to a batch of content lists which are sequentially presented and provided by the recommendation service at the current time, and the second brush refers to a batch of content lists which are sequentially presented and provided by the recommendation service at the next time.
Referring to fig. 1, taking an information push device integrated in a computer device as an example, fig. 1 is a schematic view of an implementation environment scene of an information push method provided in an embodiment of the present application, where the information push method includes a server a and a terminal B, where the server a may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as cloud service, a cloud database, cloud computing, a cloud function, cloud storage, Network service, cloud communication, middleware service, domain name service, security service, Network acceleration service (CDN), and a big data and artificial intelligence platform, and the information push method or the apparatus disclosed in the present application, where the plurality of servers may form a block chain, and the servers are nodes on the block chain. The server A can obtain an information set; acquiring interactive feedback data of each content information in the information set, and performing positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set; acquiring an information set to be recommended; determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set; and sequencing and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended.
The terminal B may be various computer devices capable of displaying information, such as a smart phone, a tablet computer, a notebook computer, a desktop computer, and a smart television, but is not limited thereto. The terminal B and the server a may be directly or indirectly connected through a wired or wireless communication manner, the terminal B may be installed with a client, which may be a mobile client, a game client, a World Wide Web (Web for short) client, an instant messaging client, and the like, and the server a may receive data uploaded by the terminal B through the client to perform a corresponding information push operation, which is not limited herein.
It should be noted that the schematic diagram of the implementation environment scenario of the information pushing method shown in fig. 1 is only an example, and the implementation environment scenario of the information pushing method described in the embodiment of the present application is for more clearly explaining the technical solution of the embodiment of the present application, and does not constitute a limitation to the technical solution provided by the embodiment of the present application. As can be known to those skilled in the art, with the evolution of information push and the appearance of new service scenarios, the technical solution provided in the present application is also applicable to similar technical problems.
The recommendation system is used for solving the problem of information overload caused by the appearance and popularization of the Internet, and recommends information, products and the like which are interested by a user to the personalized information push system of the user according to the information demand, interest and the like of the user. Compared with a search engine, the recommendation system carries out personalized calculation by researching the interest preference of the user, and the system finds the interest points of the user, so that the user is guided to find the own information requirement. A good recommendation system not only can provide personalized services for users, but also can establish close relations with the users, and the users can generate dependence on the recommendation.
When a user browses contents in a recommendation stream scene through a client, the recommendation system can acquire consumption data of the user in real time and give positive and negative feedback to the contents in the follow-up process so as to meet the instant interest points of the user. Specifically, when the user is more interested in a certain type of content, the recommendation system can continuously recommend similar content in subsequent refreshing; when the user is not interested in certain content, the recommendation system can reduce the occurrence of the content in subsequent refreshing, so that the information consumption experience of the user is improved. However, in the prior art, there is no accurate measure index for the recommendation effect of the positive/negative feedback adjustment of the recommendation system, and generally, the method is to measure the quality of the effect by a technician according to subjective experience judgment of the content pushed by the recommendation system, or to measure the quality of the effect by a technician investigating a sample extracted from the content pushed by the recommendation system, so that the measurement of the positive/negative feedback adjustment effect in the prior art has the problems of strong subjectivity or untimely feedback, etc., which may cause the situation of only simply recommending or shielding a certain kind of content, so that the content obtained by the user is single, which is not beneficial to the interest diversity and interest point mining of the user, and is further not beneficial to effectively monitoring and iterating the recommendation system, resulting in low information pushing efficiency.
In order to solve the above problems, an embodiment of the present application provides an information pushing method, which measures an effect of a recommendation system on a user by determining a measurement index, and further performs effective monitoring and iterative optimization on the recommendation system according to the measurement index, so as to obtain content information with a better recommendation effect more accurately and in real time and push the content information to the user, thereby improving the efficiency of information pushing.
The specific implementation process is described in detail below. It should be noted that the following description of the embodiments is not intended to limit the preferred order of the embodiments.
The embodiment will be described in terms of an information pushing apparatus, which may be specifically integrated in a computer device, where the computer device may be a server or a terminal, and the application is not limited herein.
Referring to fig. 2, fig. 2 is a schematic flow chart illustrating an information pushing method according to an embodiment of the present application. The information pushing method can be executed by computer equipment, wherein the information pushing method comprises the following steps:
in step 101, a set of information is obtained.
The information set may be an entirety formed by at least one piece of content information, where the content information refers to a carrier carrying information and may be consumed by being displayed to a user, so that the user obtains corresponding information. In one embodiment, the content information may include, but is not limited to, video information, audio information, text information, and image information. The information set can be obtained from the internet through a recommendation system by responding to a request triggered by a user, or the information set can be obtained from a storage device.
In an embodiment, please refer to fig. 3, where fig. 3 is a specific flowchart of an information pushing method provided in the embodiment of the present application, which may obtain an information set, where the information set is a batch of content information lists that are sequentially displayed and currently provided by a recommendation system. For example, when a user opens a client and triggers a service request of a recommendation flow scenario, the recommendation system may obtain a total set of recommendation information, which may be obtained by an upstream module of the recommendation system, and may collect historical interest information of the user, where the historical interest information may be some classification information of interest of the user, such as political classification, celebrity classification, sports classification, or movie classification, and may also be some tag information of interest of the user, such as laugh, sadness, dispute, suspicion, or science fiction, and the historical interest information may also be other types of interest information, which is not limited herein.
In an embodiment, historical interest information of a user in a current client can be collected, historical interest information consumed by most users in the client can be collected, and historical interest information of the user in other clients can be collected, so that content information in the recommendation information total set is screened according to the historical interest information to obtain an information set.
In step 102, interactive feedback data of each content information in the information set is collected, and positive and negative feedback classification is performed on the content information according to the distribution of the interactive feedback data, so as to obtain a positive feedback information set and a negative feedback information set.
The interactive feedback data may be data for interaction between the user and the content information, and whether the user is interested in the content information may be evaluated based on the interactive feedback data. Further, the content information may be classified according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set, where the positive feedback information set is a set composed of content information classified into positive feedback categories in the information set, that is, may be an entirety composed of content information that may be interested in the user in the information set, and the negative feedback information set is a set composed of content information classified into negative feedback categories in the information set, that is, may be an entirety composed of content information that may not be interested in the user in the information set.
After the user acquires the information set, each content information has different interactive feedback data according to different user likeness, in order to adjust the content information subsequently pushed by the recommendation system according to the user preference to achieve a better recommendation effect, the positive and negative feedback classification can be performed on the content information according to the interactive feedback data of each content information in the acquired information set and the distribution of the interactive feedback data, specifically, the interactive feedback data of each content information in the information set by the user can be collected, the positive feedback content information set and the negative feedback content information set can be obtained by screening according to the interactive feedback data according to preset conditions, so that the positive and negative feedback sequencing adjustment can be performed on a new recommendation system presentation list according to the positive feedback content information set and the negative feedback content information set, therefore, the recommendation effect is optimized, and the information pushing efficiency is improved. Please refer to fig. 3, interactive feedback data of each content information in the information set may be collected, a preset index condition is determined based on the interactive feedback data, the preset index condition includes a positive feedback preset index condition and a negative feedback preset index condition, content information meeting the positive feedback preset index condition is screened from the information set, a positive feedback information set N is obtained based on the content information meeting the positive feedback preset index condition in the information set, content information meeting the negative feedback preset index condition is screened at the same time, and a negative feedback information set L is obtained based on the content information meeting the negative feedback preset index condition in the information set.
In an embodiment, after the information set is obtained and the content information in the information set is pushed to the client, the interaction feedback data of the user for each content information in the information set may be collected, where the interaction feedback data may include data of whether each content information is clicked and consumed by the user, historical click rate data of each content information, play completion or dwell time data of each content information in the brush, interaction data of each content information in the brush, and the like, and the interaction data may include, but is not limited to, data of like, collection, comment, share, concern, and tile flipping. It should be noted that the interactive feedback data may also include other types of interactive feedback data, and is not limited herein.
Thus, a preset index condition can be determined according to one, part or all of the interactive feedback data, for example, for the positive feedback preset index condition, the preset index condition can include one, part or all of index conditions that a user clicks, the historical click rate reaches X%, the playing completion rate exceeds Y%, or the dwell time data exceeds Z seconds, and interactive data exists, and when the content information in the information set meets the positive feedback preset index condition, the content information meeting the condition is classified into the positive feedback information set; the negative feedback preset index conditions can include one, part or all of index conditions of no click by a user, historical click rate less than A%, playing completion lower than B%, dwell time data less than C seconds, no interaction data and the like, and when the content information in the information set meets the negative feedback preset index conditions, the content information meeting the conditions is classified into the negative feedback information set. Wherein X, Y, A and B are values from 0 to 100, and Z and C are values greater than 0, the setting of the preset index condition may be determined according to actual conditions, and may be one, part or all of the above index conditions, or may include other conditions, which is not limited herein.
In step 103, a set of information to be recommended is obtained.
The information set to be recommended can be an integral formed by at least one piece of content information to be recommended, the content information to be recommended is content information in the information set to be recommended, and the content information to be recommended is a carrier for bearing information and is used for being pushed to a user for consumption so that the user can obtain corresponding information.
In an embodiment, please continue to refer to fig. 3, an information set M to be recommended is obtained by the recommendation system, where the information set M to be recommended is a next content information list sequentially displayed on the basis of the current brush provided by the recommendation system, that is, a second brush, and the information set M to be recommended includes at least one piece of content information to be recommended. The recommendation system can acquire the total set M of information to be recommended and collect historical interest information of the user, wherein the historical interest information can be historical interest information exhibited by the user in the client, historical interest information exhibited by most users in the client and historical interest information exhibited by the user in other clients, so that content information in the total set of information to be recommended is screened according to the historical interest information to obtain the set of information to be recommended. In an embodiment, in order to increase the accuracy of the feedback adjustment, corresponding calculation may be performed by increasing the number of samples, and specifically, a plurality of sets of information to be recommended may be obtained continuously, that is, a plurality of brushes of sets of information to be recommended may be obtained to increase the number of samples.
In step 104, a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set are determined.
The first matching score is an integral formed by the matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set, and the second matching score is an integral formed by the matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set.
In an embodiment, the matching degree between each piece of content information to be recommended and each piece of content information in the positive feedback information set and between each piece of content information to be recommended and each piece of content information in the negative feedback information set can be calculated, so that the matching degree between each piece of content information to be recommended and each piece of content information in the positive feedback information set and between each piece of content information to be recommended and each piece of content information in the negative feedback information set are evaluated according to the matching degree, and the piece of content information to be recommended is adjusted according to the matching degree between the piece of content information to be recommended and the positive and negative feedback information set, so that the information pushing efficiency is improved. Specifically, with continued reference to fig. 3, a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set N and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set L may be determined according to the matching degree.
In an embodiment, the matching degree between each piece of content information to be recommended and each piece of content information in the positive feedback information set and each piece of content information to be recommended and each piece of content information in the negative feedback information set can be calculated according to attribute information carried by each piece of content information, the attribute information can be classification information, label information and the like of each piece of content information, wherein the classification information can include first-level classification information, second-level classification information, third-level classification information and the like, and in addition, the matching degree between each piece of content information to be recommended and each piece of content information in the positive feedback information set and between each piece of content information to be recommended and each piece of content information in the negative feedback information set can be calculated according to information such as the distribution time, the playing time, the number of words, the color or the file size of the content information, and a specific calculation method can be selected according to actual situations, and are not limited herein.
Specifically, the first attribute information of each piece of content information to be recommended in the set of information to be recommended, the second attribute information of each piece of content information in the set of positive feedback information, and the third attribute information of each piece of content information in the set of negative feedback information may be obtained. And respectively calculating the matching degree of the first attribute information of each piece of content information to be recommended in the information set to be recommended, the second attribute information of each piece of content information in the positive feedback information set and the third attribute information of each piece of content information in the negative feedback information set. And then determining a first matching score of the first attribute information and the second attribute information and a second matching score of the first attribute information and the third attribute information according to the matching degree, namely determining a first matching score of each content information to be recommended and each content information in the positive feedback information set and a second matching score of each content information to be recommended and each content information in the negative feedback information set.
In an embodiment, a specific example may be given by taking a case of calculating a matching degree between each piece of content information to be recommended and each piece of content information in the positive feedback information set and between each piece of content information to be recommended and each piece of content information in the negative feedback information set according to the attribute information of the piece of content information, and meanwhile, it may be assumed that the piece of content information m to be recommended in the piece of content information set is to be recommended1、m2And m3Content information n in a set of positive feedback information1And n2Content information l in a negative feedback information set1And l2Wherein the content information m to be recommended1The carried attribute information comprises first-level classification information 'sports' and second-level classification information 'basketball', and the content information m to be recommended2The carried attribute information comprises first-level classification information 'food' and second-level classification information 'milk', and the content information m to be recommended3The carried attribute information comprises first-level classification information 'movie' and second-level classification information 'inference', and the content information n1The carried attribute information comprises first-level classification information 'sports' and second-level classification information 'football', and the content information n2The carried attribute information comprises first-level classification information 'movie' and second-level classification information 'suspicion', and the content information l1The carried attribute information comprises first-level classification information 'food' and second-level classification information 'milk', and the content information l2The carried attribute information includes first-level classification information "movie" and second-level classification information "horror", and the score is 1 when the first-level classification information is the same between two pieces of content information, and the score is 2 when the first-level classification information is the same and the second-level classification information is the same.
Therefore, the content information m to be recommended can be obtained through the calculation of the matching degree1With the content information n in the positive feedback information set1The matching degree of the positive feedback information set is that the first-stage classification information is the same, the second-stage classification information is different, and the matching degree of the positive feedback information set is the same as the content information n in the positive feedback information set1The matching degree of the content information m to be recommended is that the first-level classification information and the second-level classification information are different, and the same principle is that the content information m to be recommended is1The first class classification information and the content information l in the second class classification and negative feedback information set1And l2All are different, so that the content information m to be recommended can be obtained according to the matching degree1With the content information n in the positive feedback information set1Has a first matching score of 1, and the content information n in the positive feedback information set2The first matching score of (1) is 0 score, and the content information m to be recommended1And the content information l in the negative feedback information set1And l2Are all 0 points.
Similarly, the content information m to be recommended can be obtained2With the content information n in the positive feedback information set1And n2The matching degree of the content information m to be recommended is that the first-level classification information and the second-level classification information are different, and the same applies to2The first-level classification information and the second-level classification information are the same as the content information L in the negative feedback information set, and the content information m to be recommended2And the content information l in the negative feedback information set1And l2The matching degree of the content information m is that the first-level classification information and the second-level classification information are different, so that the content information m to be recommended can be obtained according to the matching degree2With the content information n in the positive feedback information set1And n2The first matching scores of (1) are all 0 score, and the content information m to be recommended2And the content information l in the negative feedback information set1And l2The second matching scores of (1) are 2 points and 0 points.
Similarly, the content information m to be recommended can be obtained3With the content information n in the positive feedback information set1The matching degree of the content information m to be recommended is that the first-level classification information and the second-level classification information are different3With content in the positive feedback information setInformation n2The matching degree of the first-level classification information and the second-level classification information is the same, and the content information m to be recommended can be obtained in the same way3The first class information and the second class information and the content information L in the negative feedback information set1All different, content information m to be recommended3And the content information l in the negative feedback information set2The first-level classification information is the same, the second-level classification information is different, and the content information m to be recommended can be obtained according to the matching degree3With the content information n in the positive feedback information set1Is 0 score, and the content information n in the positive feedback information set2Has a first matching score of 2 points, has a second matching score of 0 points with the content information l1 in the negative feedback information set, and has a second matching score of 0 points with the content information l in the negative feedback information set2Is 1.
In step 105, ranking and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended.
The target information set to be recommended is obtained by sequencing and adjusting according to the information set to be recommended and is an integral of content information to be recommended, which needs to be pushed to the client, in the information set to be recommended. For convenience of illustration, the content information to be recommended in the target information set to be recommended may be named as target content information to be recommended, and therefore, the target information set to be recommended includes at least one piece of target content information to be recommended.
In order to push content information that may be of interest to a user and reduce content information that may not be of interest to the user as much as possible, please refer to fig. 3, the order of each piece of content information to be recommended in the set M to be recommended may be adjusted according to a first matching score and a second matching score obtained from the matching degree of the set M to be recommended, the set N of positive feedback information, and the set L of negative feedback information, so as to obtain a target set of information to be recommended that may be of interest to the user and push the target set of information to be recommended. In an embodiment, the content information to be recommended with ranking greater than a preset threshold in the information set to be recommended after the ranking adjustment can be obtained, and a target information set to be recommended is obtained and pushed to the client.
In an embodiment, a maximum value of the first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set may be obtained, so as to obtain a first target matching score of each piece of content information to be recommended and the positive feedback information set, and a maximum value of the second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set is obtained, so as to obtain a second target matching score of each piece of content information to be recommended and the negative feedback information set. And performing sequencing adjustment on each piece of content information to be recommended according to the first target matching score and the second target matching score.
Wherein, it can be assumed that the set of information to be recommended M = (M)1,m2,...,mi) Positive feedback information set N = (N)1,n2,...,np) Negative feedback information set L = (L)1,l2,...,lq) Wherein M represents the target content information to be recommended in the target information set M to be recommended, MiRepresenting ith target content information to be recommended in a target information set T to be recommended, N representing content information in a positive feedback information set N, NpRepresents the p-th content information in the positive feedback information set N, L represents the content information in the negative feedback information set L, LqRepresents the q content information in the negative feedback information set L, and each content information M to be recommended is assumediThe first matching Score with each content information in the positive feedback information set is Score (m)i，np) Each content information M to be recommendediThe first matching Score with each content information in the negative feedback information set is Score (m)i，lq) If the target matching score of the jth content information to be recommended is:
Figure 269955DEST_PATH_IMAGE001
the second target matching score of the jth content information to be recommended is as follows:
Figure 780571DEST_PATH_IMAGE002
for example, please continue to refer to the content in step 104, and for the information m of the content to be recommended in the information set to be recommended1，m2，m3Content information n in a set of positive feedback information1And n2Content information l in a negative feedback information set1And l2And the information m of the content to be recommended can be obtained by calculation1With each content information n in the set of positive feedback information1And n2Is 1 point and 0 point, respectively, and thus, the content information m to be recommended1Matching score with the first target of the positive feedback information set is 1 point; similarly, the information m of the content to be recommended can be obtained by calculation2With each content information n in the set of positive feedback information1And n2Are all 0 points, and therefore, the content information m to be recommended2The matching score of the first target of the positive feedback information set is 0 score; similarly, the information m of the content to be recommended can be obtained by calculation3With each content information n in the set of positive feedback information1And n2Is 0 point and 2 points, respectively, and thus, the content information m to be recommended3The first target matching score with the positive feedback information set is 2.
Similarly, the information m of the content to be recommended can be obtained by calculation1And each content information l in the negative feedback information set1And l2Are all 0 points, and therefore, the content information m to be recommended1The matching score of the second target of the negative feedback information set is 0; the information m of the content to be recommended can be obtained by calculation2And each content information l in the negative feedback information set1And l2Respectively 2 points and 0 points, and therefore, the content information m to be recommended2The matching score of the second target of the negative feedback information set is 2 points; the information m of the content to be recommended can be obtained by calculation3And each of the negative feedback information setsVolume information l1And l2Is 0 point and 1 point, respectively, and thus, the content information m to be recommended3The second target matching score with the set of negative feedback information is 1. Therefore, the content information m to be recommended can be respectively treated according to the first target matching score and the second target matching score1、m2And m3And adjusting the sequence in the information set to be recommended.
In an embodiment, the adjusting the ranking of each piece of content information to be recommended according to the first target matching score and the second target matching score may include:
(1) weighting the first target matching score and the second target matching score corresponding to each piece of content information to be recommended to obtain a target score of each piece of content information to be recommended;
(2) and sequencing each piece of content information to be recommended according to the target score.
Wherein, the first target matching score and the second target matching score corresponding to each piece of content information to be recommended may be weighted according to a preset weight, and the specific weight may be adjusted according to an actual process, for example, the first target matching score corresponding to each piece of content information to be recommended may be multiplied by a first preset weight, the second target matching score corresponding to each piece of content information to be recommended may be multiplied by a second preset weight, the first preset weight can be larger than the second preset weight, and the first target matching score multiplied by the first preset weight and the second target matching score multiplied by the second preset weight are accumulated to obtain the target score of each piece of content information to be recommended, therefore, the positions of the content information to be recommended in the information set to be recommended can be reordered from large to small according to the target score, and the content information to be recommended which is ranked in the front can be preferentially pushed to the client.
Therefore, the position of the content information to be recommended with a higher first target matching score in the information set to be recommended can be moved forward, the position of the content information to be recommended with a higher second target matching score in the information set to be recommended can be moved backward, the content information to be recommended related to the positive feedback information set which is possibly interesting to the user can be pushed to the user as much as possible, the content information to be recommended related to the negative feedback information set which is possibly uninterestive to the user can be pushed to the user as little as possible, and the information pushing effect is improved.
In an embodiment, a difference processing may be performed on a first target matching score and a second target matching score corresponding to each piece of content information to be recommended, that is, the second target matching score is subtracted from the first target matching score of each piece of content information to be recommended, so as to obtain a target score of each piece of content information to be recommended, and thus each piece of content information to be recommended is sorted according to the target score, wherein when at least two pieces of content information to be recommended have the same target score, the size of the first target matching score corresponding to the piece of content information to be recommended with the same target score is detected, and the piece of content information to be recommended with the larger first target matching score is sorted preferentially; when the content information to be recommended with the same target score and the same first target matching score exists, detecting the size of the corresponding second target matching score, and performing priority ordering on the content information to be recommended with the smaller second target matching score; and when at least two pieces of content information to be recommended exist, wherein the target scores of the content information to be recommended are the same, and the corresponding first target matching score and the second target matching score of the content information to be recommended are the same, the sequencing of the at least two pieces of content information to be recommended is randomly arranged.
For example, please continue to refer to the above content, for the information (m) of the content to be recommended in the information set to be recommended1，m2，m3) Content information n in a set of positive feedback information1And n2The content information l1 and l2 in the negative feedback information set can obtain the content information m to be recommended1The matching score with the first target of the positive feedback information set is 1 score, and the matching score with the second target of the negative feedback information set is 0 score; information m of content to be recommended2The matching score with the first target of the positive feedback information set is 0 score, and the matching score with the second target of the negative feedback information set is 2 scores; information m of content to be recommended3Matching score with the first target of the positive feedback information set is 2 points, and matching score with the negative feedback information set is 2 pointsThe second target match score for the interest set is 1. Content information m to be recommended1Performing difference processing on the corresponding first target matching score and the corresponding second target matching score to obtain a target score of 1-0= 1; content information m to be recommended2Performing difference processing on the corresponding first target matching score and the corresponding second target matching score to obtain a target score of 0- (2) = -2 points; content information m to be recommended3And performing difference processing on the corresponding first target matching score and the second target matching score to obtain a target score of 2-1= 1. And then m can be matched according to the size of the target fraction1、m2And m3Sorting from large to small, wherein, m is1And m3Is the same, and thus m can be detected1And m3M is obtained according to the size of the corresponding first target matching fraction3The corresponding first target matching score is larger, therefore, m can be matched1、m2And m3Sorting from big to small to obtain a target information set (m) to be recommended3，m1，m2) And pushed to the client.
In one embodiment, in order to satisfy the interest diversity and interest point mining of users, avoid the situation of only recommending or shielding a certain type of content, the first target matching score and the second target matching score satisfying the condition may be filtered by setting a preset threshold, therefore, the position of each piece of content information to be recommended in the information set to be recommended is adjusted in sequence, and specifically, first target to-be-recommended content information of which the corresponding first target matching score is greater than a first preset threshold value in each to-be-recommended content information can be acquired, and second target to-be-recommended content information with a second target matching score larger than a second preset threshold, the first preset threshold and the first preset threshold may be the same or different, and may be 0 or 1, and the specific value may be determined according to an actual situation, which is not limited herein; determining the ranking number of forward to-be-moved lists of the first target to-be-recommended content information subjected to ranking adjustment in the to-be-recommended information set through a first preset adjusting algorithm, for example, the first target to-be-recommended content information can be rankedThe ranking of the information is moved forward by 1, 2 or 3, etc. to obtain the ranking number to be moved forward, or the ranking number to be moved forward can be determined according to the size of the first target matching score of the first target content information to be recommended, for example, for the first target content information to be recommended with the first target matching score of 2, the ranking number to be moved forward can be 2, for the first target content information to be recommended with the first target matching score of 3, the ranking number to be moved forward can be 3, etc., to obtain the ranking number to be moved forward; determining the negative rank number to be moved of the second target content information to be recommended in the information set to be recommended by using a second preset adjusting algorithm, for example, moving the rank of the second target content information to be recommended backward by 1, 2, or 3, and so on, to obtain the negative rank number to be moved, or determining the rank number to be moved backward according to the size of the second target matching score of the second target content information to be recommended, for example, for the second target content information to be recommended with a second target matching score of 2, the rank number to be moved backward may be 2, for the second target content information to be recommended with a second target matching score of 3, the rank number to be moved backward may be 3, and so on, to obtain the negative rank number to be moved; determining the target number of ranks to be moved of each piece of content information to be recommended in the set of information to be recommended according to the positive number of ranks to be moved and the negative number of ranks to be moved of each piece of content information to be recommended, for example, suppose that the piece of content information m to be recommended is4The ranking number of the target to-be-moved is 5 ranks by positive movement, wherein the positive to-be-moved ranking number is 8, and the negative to-be-moved ranking number is 3; and adjusting the ranking of each piece of content information to be recommended based on the target ranking number to be moved.
After the target number of ranks to be moved is determined, if the number of ranks to be moved is the same, the number of ranks to be moved is adjusted according to specific conditions, in one embodiment, when the information of the content to be recommended after being adjusted according to the target number of ranks to be moved is the same as the number of ranks to be recommended of the information of the content to be recommended without being adjusted in ranks, the information of the content to be recommended after being adjusted according to the target number of ranks to be moved can be preferentially sorted, and the information of the content to be recommended without being adjusted in ranks can be arranged at the next position of the information of the content to be recommended after being adjusted according to the target number of ranks to be moved; when the to-be-recommended content information to be ranked adjusted according to the target to-be-moved ranking number is the same as the to-be-recommended content information to be ranked adjusted according to the target to-be-moved ranking number, the size of a first target matching score corresponding to the to-be-recommended content information to be ranked in the same way can be detected, and the to-be-recommended content information with the larger first target matching score is subjected to priority ranking; when the content information to be recommended with the same rank and the same first target matching score exists, the size of the corresponding second target matching score can be detected, and the content information to be recommended with the smaller second target matching score is prioritized; when there are at least two pieces of content information to be recommended, which have the same rank and the corresponding first target matching score and second target matching score, the ranking between the at least two pieces of content information to be recommended may be randomly arranged. Therefore, the content information to be recommended is flexibly sequenced, the situation that only a certain type of content is recommended or shielded is avoided, the possibility that the content information to be recommended with low matching degree with the positive feedback information set and the negative feedback information set is pushed to the user is increased, the interest diversity and interest point mining of the user are met, and the information pushing effect is improved.
Specifically, it may be assumed that the content information m to be recommended5M which is adjusted according to the number of the target to-be-moved ranks and is not required to be ranked and adjusted6Is the same, m can be prioritized5Sorting is performed, and m can be6Is arranged at m5The next position of (a); suppose content information m to be recommended5M is adjusted according to the number of the target to-be-moved ranks and the number of the target to-be-moved ranks6Is the same, at this point m can be determined5And m6The corresponding first target matching score can prioritize the content information to be recommended with a larger first target matching score, and the content information to be recommended with a smaller first target matching scoreAnd arranging the content information to be recommended at the next position of the content information to be recommended with the larger first target matching score.
In an embodiment, after the target information set to be recommended is obtained, positive feedback measurement scores of the target information set to be recommended and a positive feedback information set may be calculated according to a first matching score, negative feedback measurement scores of the target information set to be recommended and the negative feedback information set may be calculated according to a second matching score, the measurement scores include a relevance score and a coverage score for measuring effects of positive feedback and negative feedback, where the relevance score is used to represent a degree of relevance between content information to be recommended in the target information set to be recommended and content information to be recommended in the information set to be recommended, the coverage score is used to represent a coverage between content information to be recommended in the target information set to be recommended and content information to be recommended in the information set to be recommended, and thus relevance and coverage of the target information set to be recommended may be measured according to the positive feedback measurement scores and the negative feedback measurement scores, and then, the recommendation system is adjusted to obtain a target information set to be recommended with a better effect, and the information pushing efficiency is improved.
As can be seen from the above, in the embodiments of the present application, an information set is obtained; acquiring interactive feedback data of each content information in the information set, and performing positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set; acquiring an information set to be recommended; determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set; and sequencing and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended. Therefore, positive feedback information sets and negative feedback information sets are obtained by timely carrying out positive and negative feedback classification on the content information in the information sets, a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information sets and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information sets are calculated, and sequencing adjustment is carried out on each piece of content information to be recommended according to the first matching score and the second matching score, so that the target information sets to be recommended are obtained and pushed more accurately in real time, the recommendation effect is improved, and the information pushing efficiency is improved.
The method described in the above examples is further illustrated in detail below by way of example.
In this embodiment, the information pushing apparatus will be described by taking an example in which the information pushing apparatus is specifically integrated in a server. Specifically, please refer to fig. 4, and fig. 4 is a schematic flowchart illustrating another specific flow of an information push method according to an embodiment of the present application.
For a better description of the embodiments of the present application, please refer to fig. 4 and 5 together. As shown in fig. 5, fig. 5 is another schematic flow chart of the information pushing method according to the embodiment of the present application. The specific process is as follows:
in step 201, the server obtains an information set, collects interactive feedback data of each content information in the information set, and performs positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set.
The server obtains an information set, wherein the information set is a batch of content information lists which are sequentially displayed and are currently provided by the recommendation system, the information set comprises at least one piece of content information, the content information comprises but is not limited to content information such as video information, audio information, text information, image information and multimedia information, the multimedia information can be content information which integrates at least two of the content information such as the video information, the audio information, the text information, the image information and the multimedia information, and for example, the information set can be a text which integrates the text information and the image information. For convenience of description, the content information in the present embodiment is specifically described by taking video information as an example. The server may obtain the information set according to a request triggered by the user, for example, when the user opens the client and triggers a service request for recommending a streaming scene, the recommendation system may obtain a total set of recommendation information, and may collect historical interest information of the user, where the historical interest information may be some classification information that the user is interested in, for example, a political classification, a celebrity classification, a sports classification, or a movie classification, and may also be some tag information that the user is interested in, for example, a glad, sad, dispute, suspense, or science fiction, and the historical interest information may also be other types of interest information, which is not limited herein.
After the user acquires the information set, each piece of video information has different interactive feedback data according to different degrees of user liking, in order to adjust the video information subsequently pushed by the recommendation system according to the user liking to achieve a better recommendation effect, the server can acquire the interactive feedback data of each piece of video information in the information set, and perform positive and negative feedback classification on the video information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set, so that the sequencing adjustment of positive and negative feedback can be performed on a new recommendation system display list according to the positive feedback video information set and the negative feedback video information set. The interactive feedback data may include data of whether each piece of video information is clicked and consumed by the user in the brush, historical click rate data of each piece of video information, playing completion of each piece of video information in the brush, interactive data of each piece of video information in the brush, and the like, and the interactive data may include, but is not limited to, data of likes, comments, shares, concerns, tiles turning, and the like. It should be noted that the interactive feedback data may also include other types of interactive feedback data, and is not limited herein.
In an embodiment, interactive feedback data of each video information in the information set may be collected, a preset index condition is determined based on the interactive feedback data, the preset index condition includes a positive feedback preset index condition and a negative feedback preset index condition, video information meeting the positive feedback preset index condition is screened from the information set to obtain a positive feedback information set, and video information meeting the negative feedback preset index condition is screened to obtain a negative feedback information set.
For example, the positive feedback preset index condition may be:
index condition 1: in the historical Y-brush, the click rate reaches X% (e.g., 20%);
index condition 2: the playing completion degree of the brush reaches Y% (for example, 20%);
index condition 3: the brush has interactive data (including, but not limited to, likes/concerns/shares/flips/comments, etc.).
The negative feedback preset index condition may be:
index condition 1: in the historical Y-brush, the click rate is below A% (e.g., 20%);
index condition 2: in the history Y brush, the play completion is lower than B% (e.g., 20%);
index condition 3: the present brush has interactive data (e.g., uninteresting buttons).
The history Y-brush refers to a video information list presented in order by the Y-brush provided by the recommendation system before the current brush, that is, a set of Y-brush information provided by the recommendation system before the current brush, and the value of Y may be determined according to an actual situation, which is not limited herein. It should be noted that the above index conditions are only examples provided for easy understanding, and the positive feedback preset index condition and the negative feedback preset index condition may be one, a part, or all of the above index conditions, or may include other index conditions, and may be flexibly selected in an actual situation, which is not limited herein.
In step 202, the server obtains a set of information to be recommended, and obtains first attribute information of each piece of content information to be recommended in the set of information to be recommended, second attribute information of each piece of content information in the set of positive feedback information, and third attribute information of each piece of content information in the set of negative feedback information.
The server obtains a to-be-recommended information set provided by an upstream module of the recommendation system, the to-be-recommended information set is a next video information list which is sequentially displayed and provided by the recommendation system on the basis of the current brush, and the video information which is finally pushed to the client is obtained from the to-be-recommended information set. In order to evaluate the matching degree of each to-be-recommended video information in the to-be-recommended information set and each video information in the positive feedback information set and the matching degree of each to-be-recommended video information in the to-be-recommended information set and each video information in the negative feedback information set, the first attribute information of each to-be-recommended video information in the to-be-recommended information set, the second attribute information of each video information in the positive feedback information set and the third attribute information of each video information in the negative feedback information set can be acquired, and the matching degree is evaluated according to the first attribute information, the second attribute information and the third attribute information.
The attribute information may be classification information, label information, and the like of each piece of video information, where the classification information may include first-level classification information, second-level classification information, third-level classification information, and the like, where the attribute information includes first attribute information, second attribute information, and third attribute information, and the video information includes video information to be recommended. In addition, the matching degree between each piece of video information to be recommended and each piece of video information in the positive feedback information set and between each piece of video information to be recommended and each piece of video information in the negative feedback information set can be calculated according to the information such as the distribution time, the playing time length or the file size of the video information, and the specific calculation method can be selected according to the actual situation and is not limited herein.
In step 203, the server calculates the matching degree of the first attribute information with the second attribute information and the third attribute information, and determines a first matching score of the first attribute information and the second attribute information and a second matching score of the first attribute information and the third attribute information according to the matching degree.
Specifically, the server may obtain first attribute information of each piece of video information to be recommended in the information set to be recommended, second attribute information of each piece of video information in the positive feedback information set, and third attribute information of each piece of video information in the negative feedback information set. And sequentially calculating the matching degree of the first attribute information of each piece of video information to be recommended in the information set to be recommended, the second attribute information of each piece of video information in the positive feedback information set and the third attribute information of each piece of video information in the negative feedback information set. And then determining a first matching score of the first attribute information and the second attribute information and a second matching score of the first attribute information and the third attribute information according to the matching degree, namely determining a first matching score of each piece of video information to be recommended and each piece of video information in the positive feedback information set and a second matching score of each piece of video information to be recommended and each piece of video information in the negative feedback information set.
In step 204, the server determines the maximum value of the first matching scores of each piece of content information to be recommended and each piece of content information in the positive feedback information set as a first target matching score, and determines the maximum value of the second matching scores of each piece of content information to be recommended and each piece of content information in the negative feedback information set as a second target matching score.
Specifically, the server calculates the maximum value of the first matching score of each piece of video information to be recommended and each piece of video information in the positive feedback information set, so as to obtain the first target matching score of each piece of video information to be recommended and the positive feedback information set, and simultaneously calculates the maximum value of the second matching score of each piece of video information to be recommended and each piece of video information in the negative feedback information set, so as to obtain the second target matching score of each piece of video information to be recommended and the negative feedback information set. The first target matching score is the maximum value of the first matching scores, and the second target matching score is the maximum value of the second matching scores.
In step 205, the server performs weighting processing on the first target matching score and the second target matching score corresponding to each piece of content information to be recommended to obtain a target score of each piece of content information to be recommended, and sorts each piece of content information to be recommended according to the target score.
Wherein, the server performs weighting processing on the first target matching score and the second target matching score corresponding to each piece of video information to be recommended according to a preset weight, and the specific weight can be adjusted according to an actual process, for example, the first target matching score corresponding to each video information to be recommended is multiplied by a first preset weight, the second target matching score is multiplied by the first preset weight, and a second preset weight is obtained, the first preset weight can be larger than the second preset weight, and the weighted first target matching score and the weighted second target matching score are accumulated to obtain the target score of each piece of video information to be recommended, therefore, the positions of the video information to be recommended in the information set to be recommended can be reordered according to the target scores from large to small, and the video information to be recommended which is ranked in the front can be preferentially pushed to the client. Therefore, the position of the video information to be recommended with a higher first target matching score in the information set to be recommended can be moved forward, the position of the video information to be recommended with a higher second target matching score in the information set to be recommended can be moved backward, the video information to be recommended related to the positive feedback information set which is possibly interesting to the user can be pushed to the user as much as possible, the video information to be recommended related to the negative feedback information set which is possibly uninterestive to the user can be pushed to the user as little as possible, and the information pushing effect is improved.
In step 206, the server determines the content information quantity of the target to-be-recommended content information in the target to-be-recommended information set, and calculates first coverage scores of the target to-be-recommended information set and the positive feedback information set according to the first matching score and the content information quantity.
In order to measure positive and negative feedback effects, after a target information set to be recommended is obtained, a server calculates positive feedback measurement scores of the target information set to be recommended and a positive feedback information set according to a first matching score, and calculates negative feedback measurement scores of the target information set to be recommended and a negative feedback information set according to a second matching score, wherein the measurement scores comprise a correlation score and a coverage score and are used for measuring positive feedback and negative feedback effects, the correlation score is used for representing the correlation degree between video information to be recommended in the target information set to be recommended and video information to be recommended in the information set to be recommended, the coverage score is used for representing the coverage degree between the video information to be recommended in the target information set to be recommended and the video information to be recommended in the information set to be recommended, and therefore the positive feedback measurement score and the negative feedback measurement score can be used for measuring the coverage degree between the video information to be recommended in the target information set to be recommended The relevance and the coverage are achieved, the recommendation system is adjusted, a target information set to be recommended with a better effect is obtained, and the information pushing efficiency is improved.
For the positive feedback measurement score, calculating first coverage scores of the target information set to be recommended and the positive feedback information set according to a first matching score and the number of video information, and calculating first relevance scores of the target information set to be recommended and the positive feedback information set according to the first matching score and the number of the video information; and obtaining a positive feedback measurement score of the target information set to be recommended and the positive feedback information set based on the first coverage score and the first relevance score.
The step of calculating the first coverage scores of the target information set to be recommended and the positive feedback information set according to the first matching score and the number of the video information may include:
(1) performing assignment processing on the first matching score to obtain a first assigned matching score;
(2) accumulating the first assignment matching scores to obtain a first coverage score of each target video information to be recommended and the positive feedback information set;
(3) accumulating the first coverage scores corresponding to each target video information to be recommended to obtain a first coverage total score of the target information set to be recommended;
(4) and averaging the first total coverage score according to the number of the video information to obtain a first coverage score of the target information set to be recommended and the positive feedback information set.
In order to calculate the first coverage score, the first matching score may be assigned to obtain a first assigned matching score, and in an embodiment, the magnitude of each first matching score may be detected; if the numerical value of the first matching score meets the preset condition, assigning the numerical value of the first matching score as a first preset score; if the numerical value of the first matching score does not meet the preset condition, assigning the numerical value of the first matching score as a second preset score; and obtaining a first assignment matching score based on the first preset score and the second preset score, wherein in order to more accurately measure the recommendation effect of the recommendation system, the fitting degree of the target information set to be recommended and the positive feedback information set can be represented according to the coverage of the target information set to be recommended and the positive feedback information set.
In an embodiment, for better effect measurement, if the preset condition is greater than 0, the first preset score is 0, and the second preset score is 1, the first matching score greater than 0 is assigned as 1, which indicates that the video information to be recommended corresponding to the first matching score matches with one video information in the positive feedback set; and assigning a first matching score not greater than 0 to be 0, which indicates that the video information to be recommended corresponding to the first matching score is not matched with certain video information in the positive feedback set. Specifically, it may be assumed that the target set of information to be recommended T = (T)1,t2,...,tk) Wherein k is the number of video information in the target information set to be recommended, T represents the target video information to be recommended in the target information set to be recommended T, and meanwhile, the positive feedback information set N = (N) is continuously assumed1,n2,...,np) Negative feedback information set L = (L)1,l2,...,lq) Where N denotes video information in a set N of positive feedback information, NpRepresenting the p-th video information in the positive feedback information set N, wherein p is the total amount of the video information in the positive feedback information set N, L represents the video information in the negative feedback information set L, and LqRepresenting the q video information in the negative feedback information set L, wherein q is the total amount of the video information in the negative feedback information set L, and supposing that each target video information t to be recommended isrThe first matching Score with each video information in the set of positive feedback information is Score (t)r，np) Each video information t to be recommendedrThe first matching Score with each video information in the negative feedback information set is Score (t)r，lq). The first coverage score of the r-th target video information to be recommended and the positive feedback information set is
Figure 763570DEST_PATH_IMAGE003
Where, it can be stated that if 0, then the value is 0; otherwise the value is assigned to 1.
And then accumulating the first coverage scores corresponding to each target video information to be recommended to obtain a first total coverage score of the target information set to be recommended as
Figure 752255DEST_PATH_IMAGE004
The first total coverage score is subjected to average processing according to the number of the video information to obtain a first coverage score of the target information set to be recommended and a first coverage score of the positive feedback information set as
Figure 35469DEST_PATH_IMAGE005
In step 207, the server calculates first relevance scores of the target information set to be recommended and the positive feedback information set according to the first matching score and the content information quantity, and obtains positive feedback measurement scores of the target information set to be recommended and the positive feedback information set based on the first coverage score and the first relevance score.
The server calculates first relevance scores of the target information set to be recommended and the positive feedback information set according to the first matching score and the number of the video information, so that the positive feedback measurement scores of the target information set to be recommended and the positive feedback information set can be obtained according to the first coverage score and the first relevance score.
In an embodiment, positive feedback target matching scores of each target video information to be recommended and the positive feedback information set can be obtained by obtaining the maximum value of the first matching score of each target video information to be recommended and each video information in the positive feedback information set, the positive feedback target matching scores are accumulated to obtain a positive feedback target matching total score, and the positive feedback target matching total score is subjected to average processing according to the number of the video information to obtain a first relevance score of the target information set to be recommended and the positive feedback information set.
Specifically, it can be assumed that the e-th target to-be-recommended video information and the positive feedback target matching score of the positive feedback information set are
Figure 451406DEST_PATH_IMAGE006
And accumulating the video information to be recommended of each target and the positive feedback target matching score of the positive feedback information set to obtain a total positive feedback target matching score of
Figure 187281DEST_PATH_IMAGE007
The total matching score of the positive feedback target is averaged according to the number of the video information to obtain a first relevance score of the target information set to be recommended and the positive feedback information set as
Figure 492840DEST_PATH_IMAGE008
In step 208, the server determines the content information quantity of the target to-be-recommended content information in the target to-be-recommended information set, and calculates a second coverage score of the target to-be-recommended information set and the negative feedback information set according to the second matching score and the content information quantity.
For the negative feedback measurement score, a second coverage score of the target information set to be recommended and the negative feedback information set can be calculated according to a second matching score and the number of video information, and a second correlation score of the target information set to be recommended and the negative feedback information set can be calculated according to the second matching score and the number of the video information; and obtaining negative feedback measurement scores of the target information set to be recommended and the negative feedback information set based on the second coverage score and the second relevance score.
The step of calculating the second coverage scores of the target information set to be recommended and the negative feedback information set according to the second matching score and the number of the video information may include:
(1) performing assignment processing on the second matching score to obtain a second assigned matching score;
(2) accumulating the second assignment matching scores to obtain a second coverage score of each target video information to be recommended and the negative feedback information set;
(3) accumulating the second coverage scores corresponding to each target video information to be recommended to obtain a second coverage total score of the target information set to be recommended;
(4) and averaging the second total coverage score according to the number of the video information to obtain a second coverage score of the target information set to be recommended and the negative feedback information set.
In order to calculate a second coverage score, assignment processing may be performed on the second matching score to obtain a second assigned matching score, and in an embodiment, the magnitude of each second matching score may be detected; if the numerical value of the second matching score meets the preset condition, assigning the numerical value of the second matching score as a third preset score; if the numerical value of the second matching score does not meet the preset condition, assigning the numerical value of the second matching score as a fourth preset score; and obtaining a second assignment matching score based on the third preset score and the fourth preset score, wherein in order to more accurately measure the recommendation effect of the recommendation system, the fitting degree of the target information set to be recommended and the negative feedback information set can be represented according to the coverage of the target information set to be recommended and the negative feedback information set.
In an embodiment, in order to obtain a better measurement effect, it may be assumed that the preset condition is greater than 0, the third preset score is 0, and the fourth preset score is 1, then the second matching score greater than 0 is assigned as 1, which indicates that the video information to be recommended corresponding to the second matching score is matched with some video information in the negative feedback set, and the second matching score not greater than 0 is assigned as 0, which indicates that the video information to be recommended corresponding to the second matching score is not matched with some video information in the negative feedback set, specifically, it may be assumed that the target set of information to be recommended T = (T =) is set for the video information to be recommended corresponding to the second matching score1,t2,...,tk) Wherein T represents the target video information to be recommended in the target information set to be recommended T, and TkAnd representing the kth target video information to be recommended in the target information set T to be recommended. The second coverage score of the r-th target to-be-recommended video information and the negative feedback information set is
Figure 630561DEST_PATH_IMAGE009
Wherein,
Figure 217400DEST_PATH_IMAGE010
can be expressed as if
Figure 440571DEST_PATH_IMAGE011
If 0, the value is assigned to 0; otherwise the value is assigned to 1.
And then accumulating the second coverage scores corresponding to each target video information to be recommended to obtain a second total coverage score of the target information set to be recommended as
Figure 771058DEST_PATH_IMAGE012
The second total coverage score is subjected to average processing according to the number of the video information to obtain a second coverage score of the target information set to be recommended and the negative feedback information set as
Figure 763285DEST_PATH_IMAGE013
In step 209, the server calculates second relevance scores of the target information set to be recommended and the negative feedback information set according to the second matching score and the content information quantity, and obtains negative feedback measurement scores of the target information set to be recommended and the negative feedback information set based on the second coverage score and the second relevance score.
The server calculates second relevance scores of the target information set to be recommended and the negative feedback information set according to the second matching score and the number of the video information, so that negative feedback measurement scores of the target information set to be recommended and the negative feedback information set can be obtained according to the second coverage score and the second relevance score.
In an embodiment, negative feedback target matching scores of each target video information to be recommended and the negative feedback information set can be obtained by obtaining the maximum value of the second matching score of each target video information to be recommended and each video information in the negative feedback information set, the negative feedback target matching scores are accumulated to obtain a negative feedback target matching total score, and the negative feedback target matching total score is averaged according to the number of the video information to obtain the second correlation scores of the target video information set to be recommended and the negative feedback information set.
Specifically, the h-th target to-be-recommended video information and the negative feedback target matching score of the negative feedback information set are
Figure 786604DEST_PATH_IMAGE014
And then accumulating the negative feedback target matching scores of the video information to be recommended of each target to obtain a negative feedback target matching total score of
Figure 497071DEST_PATH_IMAGE015
Matching the negative feedback target to the total score according toThe number of the video information is subjected to average processing, and second relevance scores of the target information set to be recommended and the negative feedback information set are obtained
Figure 867135DEST_PATH_IMAGE016
In step 210, the server adjusts the recommendation system based on the positive feedback metric and the negative feedback metric.
The server calculates a positive feedback score and a negative feedback score, and determines the positive and negative feedback effects according to the positive feedback score and the negative feedback score, so as to adjust the recommendation system, for example, a positive feedback target score can be set, the calculated positive feedback score is compared with the positive feedback target score, when the positive feedback score is greater than the positive feedback score, the positive feedback effect can be indicated to meet the preset requirement, when the positive feedback score is less than the positive feedback score, the positive feedback effect can be indicated to not meet the requirement, the recommendation system can be fed back, the subsequently acquired target information set to be recommended is adjusted, so that the positive feedback effect meets the preset requirement, and meanwhile, the negative feedback target score can be set, and comparing the calculated negative feedback measurement score with the negative feedback target measurement score, when the negative feedback measurement score is smaller than the negative feedback target measurement score, indicating that the negative feedback effect meets the preset requirement, and when the positive feedback measurement score is larger than the positive feedback target measurement score, indicating that the positive feedback effect does not meet the requirement, performing information feedback on the recommendation system, and adjusting a subsequently acquired target information set to be recommended so as to enable the negative feedback effect to meet the preset requirement.
For example, the weight of the weighting processing of the first target matching score and the second target matching score corresponding to each piece of video information to be recommended may be adjusted to realize the adjustment of the target score, and further realize the sorting of each piece of video information to be recommended in the information set to be recommended, so as to adjust the target information set to be recommended pushed to the client until the positive feedback weighing score and the negative feedback weighing score meet the preset requirements.
In an embodiment, in order to obtain more accurate positive feedback measurement scores and negative feedback measurement scores and to more accurately adjust a recommendation system, a preset number of historical target information sets to be recommended may be obtained, video information to be recommended in the historical target information sets to be recommended is video information which is obtained and pushed by the recommendation system according to the positive feedback information sets and the negative feedback information sets, so that the positive feedback measurement scores and the negative feedback measurement scores are calculated based on the historical target information sets to be recommended and the target information sets to be recommended, and the recommendation system is adjusted according to the calculated positive feedback measurement scores and the calculated negative feedback measurement scores.
As can be seen from the above, in the embodiment of the application, the server acquires the information set, acquires the interactive feedback data of each content information in the information set, and performs positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set; the server acquires an information set to be recommended, and acquires first attribute information of each piece of content information to be recommended in the information set to be recommended, second attribute information of each piece of content information in a positive feedback information set and third attribute information of each piece of content information in a negative feedback information set; the server calculates the matching degree of the first attribute information with the second attribute information and the third attribute information respectively, and determines a first matching score of the first attribute information and the second attribute information and a second matching score of the first attribute information and the third attribute information according to the matching degree; the server determines the maximum value of the first matching scores of each piece of content information to be recommended and each piece of content information in the positive feedback information set as a first target matching score, and determines the maximum value of the second matching scores of each piece of content information to be recommended and each piece of content information in the negative feedback information set as a second target matching score; the server carries out weighting processing on the first target matching score and the second target matching score corresponding to each piece of content information to be recommended to obtain a target score of each piece of content information to be recommended, and sorts each piece of content information to be recommended according to the target score; the server determines the content information quantity of the target content information to be recommended in the target information set to be recommended, and calculates first coverage scores of the target information set to be recommended and the positive feedback information set according to the first matching score and the content information quantity; the server calculates first relevance scores of the target information set to be recommended and the positive feedback information set according to the first matching score and the content information quantity, and obtains positive feedback measurement scores of the target information set to be recommended and the positive feedback information set based on the first coverage score and the first relevance score; the server determines the content information quantity of the target content information to be recommended in the target information set to be recommended, and calculates second coverage scores of the target information set to be recommended and the negative feedback information set according to the second matching score and the content information quantity; the server calculates second relevance scores of the target information set to be recommended and the negative feedback information set according to the second matching scores and the content information quantity, and obtains negative feedback measurement scores of the target information set to be recommended and the negative feedback information set on the basis of the second coverage scores and the second relevance scores; the server adjusts the recommendation system based on the positive feedback metric and the negative feedback metric. Therefore, positive feedback information sets and negative feedback information sets are obtained by timely carrying out positive and negative feedback classification on the content information in the information sets, a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information sets and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information sets are obtained, sequencing adjustment is carried out on each piece of content information to be recommended according to the first matching score and the second matching score, positive feedback weighing scores of target information sets to be recommended and the positive feedback information sets and negative feedback weighing scores of the target information sets to be recommended and the negative feedback information sets are calculated, positive and negative feedback effects are weighed, the target information sets to be recommended are obtained and pushed more accurately in real time, and the information pushing efficiency is improved, thereby improving the recommendation effect.
In order to better implement the above method, an embodiment of the present invention further provides an information pushing apparatus, which may be integrated in a computer device, where the computer device may be a server or a terminal.
For example, as shown in fig. 6, for a schematic structural diagram of an information pushing apparatus provided in an embodiment of the present application, the information pushing apparatus may include a first obtaining unit 301, a collecting unit 302, a second obtaining unit 303, a determining unit 304, and a pushing unit 305, as follows:
a first obtaining unit 301, configured to obtain an information set, where the information set includes at least one piece of content information;
an acquisition unit 302, configured to acquire interactive feedback data of each content information in the information set, and perform positive-negative feedback classification on the content information according to distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set;
a second obtaining unit 303, configured to obtain an information set to be recommended, where the information set to be recommended includes at least one piece of content information to be recommended;
a determining unit 304, configured to determine a first matching score between each piece of content information to be recommended and each piece of content information in the positive feedback information set, and a second matching score between each piece of content information to be recommended and each piece of content information in the negative feedback information set;
and the pushing unit 305 is configured to perform sorting adjustment on each piece of content information to be recommended according to the first matching score and the second matching score, obtain a target information set to be recommended, and push the target information set to be recommended.
In one embodiment, the determining unit 304 includes:
a third obtaining subunit, configured to obtain first attribute information of each piece of content information to be recommended in the set of information to be recommended, second attribute information of each piece of content information in the set of positive feedback information, and third attribute information of each piece of content information in the set of negative feedback information;
the first calculating subunit is used for calculating the matching degree of the first attribute information with the second attribute information and the third attribute information respectively;
and the first determining subunit is used for determining a first matching score of the first attribute information and the second attribute information and a second matching score of the first attribute information and the third attribute information according to the matching degree.
In an embodiment, the pushing unit 305 includes:
the second determining subunit is used for determining the maximum value of the first matching scores of each piece of content information to be recommended and each piece of content information in the positive feedback information set as a first target matching score;
the third determining subunit is used for determining the maximum value of the second matching scores of each piece of content information to be recommended and each piece of content information in the negative feedback information set as a second target matching score;
and the first sequencing subunit is used for carrying out sequencing adjustment on each piece of content information to be recommended according to the first target matching score and the second target matching score.
In an embodiment, the first sorting subunit is configured to:
weighting the first target matching score and the second target matching score corresponding to each piece of content information to be recommended to obtain a target score of each piece of content information to be recommended;
and sequencing each piece of content information to be recommended according to the target score.
In an embodiment, the information pushing apparatus further includes:
the second calculating unit is used for calculating the positive feedback measurement scores of the target information set to be recommended and the positive feedback information set according to the first matching score, wherein the target information set to be recommended comprises at least one piece of target content information to be recommended;
the third calculating unit is used for calculating negative feedback weighing scores of the target information set to be recommended and the negative feedback information set according to the second matching score;
and the first adjusting unit is used for adjusting the recommendation system based on the positive feedback weighing fraction and the negative feedback weighing fraction.
In one embodiment, the second computing unit includes:
the fourth determining subunit is configured to determine the content information quantity of the target to-be-recommended content information in the target to-be-recommended information set;
a fourth calculating subunit, configured to calculate first coverage scores of the target information set to be recommended and the positive feedback information set according to the first matching score and the content information amount;
a fifth calculating subunit, configured to calculate first relevance scores of the target information set to be recommended and the positive feedback information set according to the first matching score and the content information amount;
and the first processing subunit is used for obtaining the positive feedback measurement scores of the target information set to be recommended and the positive feedback information set based on the first coverage score and the first relevance score.
In one embodiment, the fourth calculating subunit includes:
the first assignment module is used for assigning the first matching score to obtain a first assigned matching score;
the first accumulation module is used for accumulating the first assignment matching scores to obtain a first coverage score of each target content information to be recommended and the positive feedback information set;
the second accumulation module is used for accumulating the first coverage scores corresponding to each target content information to be recommended to obtain a first coverage total score of the target information set to be recommended;
and the first average module is used for carrying out average processing on the first total coverage score according to the content information quantity to obtain a first coverage score of the target information set to be recommended and the positive feedback information set.
In one embodiment, the first assignment module is configured to:
detecting the numerical value of the first matching score;
if the numerical value of the first matching score meets a preset condition, assigning the numerical value of the first matching score as a first preset score;
if the numerical value of the first matching score does not meet the preset condition, assigning the numerical value of the first matching score as a second preset score;
and obtaining a first assignment matching score based on the first preset score and the second preset score.
In an embodiment, the fifth calculating subunit is configured to:
acquiring the maximum value of the first matching score of each piece of target content information to be recommended and each piece of content information in the positive feedback information set to obtain the positive feedback target matching score of each piece of target content information to be recommended and the positive feedback information set;
accumulating the positive feedback target matching scores to obtain a positive feedback target matching total score;
and averaging the total matching scores of the positive feedback target according to the number of the content information to obtain a target information set to be recommended and a first relevance score of the positive feedback information set.
In one embodiment, the third computing unit includes:
the fifth determining subunit is configured to determine the content information quantity of the target to-be-recommended content information in the target to-be-recommended information set;
a sixth calculating subunit, configured to calculate a second coverage score of the target information set to be recommended and the negative feedback information set according to the second matching score and the content information quantity;
a seventh calculating subunit, configured to calculate second relevance scores of the target information set to be recommended and the negative feedback information set according to the second matching score and the content information amount;
and the second processing subunit is used for obtaining negative feedback measurement scores of the target information set to be recommended and the negative feedback information set based on the second coverage score and the second relevance score.
In one embodiment, the sixth calculating subunit includes:
the second assignment module is used for assigning the second matching score to obtain a second assigned matching score;
the third accumulation module is used for accumulating the second assignment matching scores to obtain a second coverage score of each target content information to be recommended and the negative feedback information set;
the fourth accumulation module is used for accumulating the second coverage scores corresponding to each target content information to be recommended to obtain a second coverage total score of the target information set to be recommended;
and the second averaging module is used for carrying out averaging processing on the second total coverage score according to the content information quantity to obtain a second coverage score of the target information set to be recommended and the negative feedback information set.
In an embodiment, the second assigning unit is configured to:
detecting the numerical magnitude of the second matching score;
if the numerical value of the second matching score meets the preset condition, assigning the numerical value of the second matching score as a third preset score;
if the numerical value of the second matching score does not meet the preset condition, assigning the numerical value of the second matching score as a fourth preset score;
and obtaining a second assignment matching score based on the third preset score and the fourth preset score.
In one embodiment, the seventh computing subunit is configured to:
acquiring the maximum value of the second matching score of each target content information to be recommended and each content information in the negative feedback information set to obtain the negative feedback target matching score of each target content information to be recommended and the negative feedback information set;
accumulating the negative feedback target matching scores to obtain a negative feedback target matching total score;
and averaging the negative feedback target matching total scores according to the content information quantity to obtain a target information set to be recommended and a second relevance score of the negative feedback information set.
In a specific implementation, the above units may be implemented as independent entities, or may be combined arbitrarily to be implemented as the same or several entities, and the specific implementation of the above units may refer to the foregoing method embodiments, which are not described herein again.
As can be seen from the above, in the embodiment of the present application, the first obtaining unit 301 obtains the information set; the acquisition unit 302 acquires interactive feedback data of each content information in the information set, and performs positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set; the second obtaining unit 303 obtains a set of information to be recommended; the determining unit 304 determines a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set; the pushing unit 305 performs sorting adjustment on each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and pushes the target information set to be recommended. Therefore, positive feedback information sets and negative feedback information sets are obtained by timely carrying out positive and negative feedback classification on the content information in the information sets, a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information sets and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information sets are calculated, and sequencing adjustment is carried out on each piece of content information to be recommended according to the first matching score and the second matching score, so that the target information sets to be recommended are obtained and pushed more accurately in real time, the recommendation effect is improved, and the information pushing efficiency is improved.
An embodiment of the present application further provides a computer device, as shown in fig. 7, which shows a schematic structural diagram of a computer device according to an embodiment of the present application, where the computer device may be a server or a terminal, and specifically:
the computer device may include components such as a processor 401 of one or more processing cores, memory 402 of one or more computer-readable storage media, a power supply 403, and an input unit 404. Those skilled in the art will appreciate that the computer device configuration illustrated in FIG. 7 does not constitute a limitation of computer devices, and may include more or fewer components than those illustrated, or some components may be combined, or a different arrangement of components. Wherein:
the processor 401 is a control center of the computer device, connects various parts of the entire computer device using various interfaces and lines, and performs various functions of the computer device and processes data by running or executing software programs and/or modules stored in the memory 402 and calling data stored in the memory 402, thereby monitoring the computer device as a whole. Optionally, processor 401 may include one or more processing cores; preferably, the processor 401 may integrate an application processor, which mainly handles operating systems, user interfaces, application programs, etc., and a modem processor, which mainly handles wireless communications. It will be appreciated that the modem processor described above may not be integrated into the processor 401.
The memory 402 may be used for storing software programs and modules, and the processor 401 executes various functional applications and information push by operating the software programs and modules stored in the memory 402. The memory 402 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data created according to use of the computer device, and the like. Further, the memory 402 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device. Accordingly, the memory 402 may also include a memory controller to provide the processor 401 access to the memory 402.
The computer device further comprises a power supply 403 for supplying power to the various components, and preferably, the power supply 403 is logically connected to the processor 401 via a power management system, so that functions of managing charging, discharging, and power consumption are implemented via the power management system. The power supply 403 may also include any component of one or more dc or ac power sources, recharging systems, power failure detection circuitry, power converters or inverters, power status indicators, and the like.
The computer device may also include an input unit 404, the input unit 404 being operable to receive input numeric or character information and to generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.
Although not shown, the computer device may further include a display unit and the like, which are not described in detail herein. Specifically, in this embodiment, the processor 401 in the computer device loads the executable file corresponding to the process of one or more application programs into the memory 402 according to the following instructions, and the processor 401 runs the application programs stored in the memory 402, thereby implementing various functions as follows:
acquiring an information set; acquiring interactive feedback data of each content information in the information set, and performing positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set; acquiring an information set to be recommended; determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set; and sequencing and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended.
The above operations can be implemented in the foregoing embodiments, and are not described in detail herein. It should be noted that the computer device provided in the embodiment of the present application and the information push method in the foregoing embodiment belong to the same concept, and specific implementation processes thereof are detailed in the foregoing method embodiment and are not described herein again.
As can be seen from the above, in the embodiments of the present application, an information set is obtained; acquiring interactive feedback data of each content information in the information set, and performing positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set; acquiring an information set to be recommended; determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set; and sequencing and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended. Therefore, positive feedback information sets and negative feedback information sets are obtained by timely carrying out positive and negative feedback classification on the content information in the information sets, a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information sets and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information sets are calculated, and sequencing adjustment is carried out on each piece of content information to be recommended according to the first matching score and the second matching score, so that the target information sets to be recommended are obtained and pushed more accurately in real time, the recommendation effect is improved, and the information pushing efficiency is improved.
It will be understood by those skilled in the art that all or part of the steps of the methods of the above embodiments may be performed by instructions or by associated hardware controlled by the instructions, which may be stored in a computer readable storage medium and loaded and executed by a processor.
To this end, the present application provides a storage medium, in which a plurality of instructions are stored, where the instructions can be loaded by a processor to execute the steps in any one of the information pushing methods provided in the present application. For example, the instructions may perform the steps of:
acquiring an information set; acquiring interactive feedback data of each content information in the information set, and performing positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set; acquiring an information set to be recommended; determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set; and sequencing and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended.
Wherein the storage medium may include: read Only Memory (ROM), Random Access Memory (RAM), magnetic or optical disks, and the like.
Since the instructions stored in the storage medium can execute the steps in any information push method provided in the embodiments of the present application, beneficial effects that can be achieved by any information push method provided in the embodiments of the present application can be achieved, for details, see the foregoing embodiments, and are not described herein again.
According to an aspect of the application, there is provided, among other things, a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the method provided in the various alternative implementations provided by the embodiments described above.
The above detailed description is provided for an information pushing method, an information pushing device, and a storage medium provided in the embodiments of the present application, and a specific example is applied in the detailed description to explain the principles and embodiments of the present application, and the description of the above embodiments is only used to help understanding the method and the core idea of the present application; meanwhile, for those skilled in the art, according to the idea of the present application, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present application.

Claims (15)

1. An information pushing method, comprising:
acquiring an information set, wherein the information set comprises at least one piece of content information;
acquiring interactive feedback data of each content information in the information set, and performing positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set;
acquiring an information set to be recommended, wherein the information set to be recommended comprises at least one piece of content information to be recommended;
determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set;
and sequencing and adjusting each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and push the target information set to be recommended.
2. The information pushing method according to claim 1, wherein the determining a first matching score of each content information to be recommended with each content information in the positive feedback information set and a second matching score of each content information to be recommended with each content information in the negative feedback information set comprises:
acquiring first attribute information of each piece of content information to be recommended in the information set to be recommended, second attribute information of each piece of content information in the positive feedback information set and third attribute information of each piece of content information in the negative feedback information set;
calculating the matching degree of the first attribute information with the second attribute information and the third attribute information respectively;
and determining a first matching score of the first attribute information and the second attribute information and a second matching score of the first attribute information and the third attribute information according to the matching degree.
3. The information pushing method according to claim 1, wherein the adjusting of the ranking of each piece of content information to be recommended according to the first matching score and the second matching score comprises:
determining the maximum value of the first matching scores of each piece of content information to be recommended and each piece of content information in the positive feedback information set as a first target matching score;
determining the maximum value of the second matching scores of each piece of content information to be recommended and each piece of content information in the negative feedback information set as a second target matching score;
and performing sequencing adjustment on each piece of content information to be recommended according to the first target matching score and the second target matching score.
4. The information pushing method according to claim 3, wherein the adjusting of the ranking of each piece of content information to be recommended according to the first target matching score and the second target matching score comprises:
weighting the first target matching score and the second target matching score corresponding to each piece of content information to be recommended to obtain a target score of each piece of content information to be recommended;
and sequencing each piece of content information to be recommended according to the target score.
5. The information pushing method according to any one of claims 1 to 4, further comprising:
calculating a positive feedback measurement score of the target information set to be recommended and the positive feedback information set according to the first matching score, wherein the target information set to be recommended comprises at least one piece of target content information to be recommended;
calculating negative feedback weighing scores of the target information set to be recommended and the negative feedback information set according to the second matching score;
and adjusting the recommendation system based on the positive feedback measurement score and the negative feedback measurement score.
6. The information pushing method according to claim 5, wherein the calculating the positive feedback score of the target information set to be recommended and the positive feedback information set according to the first matching score includes:
determining the content information quantity of the target to-be-recommended content information in the target to-be-recommended information set;
calculating a first coverage score of the target information set to be recommended and the positive feedback information set according to the first matching score and the content information quantity;
calculating a first relevance score of the target information set to be recommended and the positive feedback information set according to the first matching score and the content information quantity;
and obtaining positive feedback measurement scores of the target information set to be recommended and the positive feedback information set based on the first coverage score and the first relevance score.
7. The information pushing method according to claim 6, wherein the calculating a first coverage score of the target information to be recommended set and the positive feedback information set according to the first matching score and the content information amount comprises:
assigning the first matching score to obtain a first assigned matching score;
accumulating the first assignment matching scores to obtain a first coverage score of each target content information to be recommended and the positive feedback information set;
accumulating the first coverage scores corresponding to each target content information to be recommended to obtain a first total coverage score of the target information set to be recommended;
and carrying out average processing on the first total coverage score according to the content information quantity to obtain a first coverage score of the target information set to be recommended and the positive feedback information set.
8. The information push method according to claim 7, wherein the assigning the first matching score to obtain a first assigned matching score includes:
detecting a numerical magnitude of the first matching score;
if the numerical value of the first matching score meets a preset condition, assigning the numerical value of the first matching score as a first preset score;
if the numerical value of the first matching score does not meet the preset condition, assigning the numerical value of the first matching score as a second preset score;
and obtaining a first assignment matching score based on the first preset score and the second preset score.
9. The information pushing method according to claim 6, wherein the calculating a first relevance score of the target information set to be recommended and the positive feedback information set according to the first matching score and the content information quantity comprises:
acquiring the maximum value of the first matching score of each piece of target content information to be recommended and each piece of content information in the positive feedback information set to obtain the positive feedback target matching score of each piece of target content information to be recommended and the positive feedback information set;
accumulating the positive feedback target matching scores to obtain a positive feedback target matching total score;
and carrying out average processing on the total matching scores of the positive feedback targets according to the number of the content information to obtain a target information set to be recommended and a first relevance score of the positive feedback information set.
10. The information pushing method of claim 5, wherein the calculating negative feedback metrics of the target information set to be recommended and the negative feedback information set according to the second matching scores comprises:
determining the content information quantity of the target to-be-recommended content information in the target to-be-recommended information set;
calculating a second coverage score of the target information set to be recommended and the negative feedback information set according to the second matching score and the content information quantity;
calculating a second correlation score of the target information set to be recommended and the negative feedback information set according to the second matching score and the content information quantity;
and obtaining negative feedback measurement scores of the target information set to be recommended and the negative feedback information set based on the second coverage score and the second relevance score.
11. The information pushing method according to claim 10, wherein the calculating a second coverage score of the target information set to be recommended and the negative feedback information set according to the second matching score and the content information amount comprises:
performing assignment processing on the second matching score to obtain a second assigned matching score;
accumulating the second assignment matching scores to obtain a second coverage score of each target content information to be recommended and the negative feedback information set;
accumulating the second coverage scores corresponding to each target content information to be recommended to obtain a second coverage total score of the target information set to be recommended;
and averaging the second total coverage score according to the number of the content information to obtain a second coverage score of the target information set to be recommended and the negative feedback information set.
12. The information push method according to claim 11, wherein the assigning the second matching score to obtain a second assigned matching score includes:
detecting a numerical magnitude of the second matching score;
if the numerical value of the second matching score meets a preset condition, assigning the numerical value of the second matching score as a third preset score;
if the numerical value of the second matching score does not meet the preset condition, assigning the numerical value of the second matching score as a fourth preset score;
and obtaining a second assignment matching score based on the third preset score and the fourth preset score.
13. The information pushing method according to claim 12, wherein the calculating a second relevance score of the target information set to be recommended and the negative feedback information set according to the second matching score and the content information amount comprises:
obtaining the maximum value of the second matching score of each target content information to be recommended and each content information in the negative feedback information set to obtain the negative feedback target matching score of each target content information to be recommended and the negative feedback information set;
accumulating the negative feedback target matching scores to obtain a negative feedback target matching total score;
and carrying out average processing on the negative feedback target matching total score according to the content information quantity to obtain a second relevance score of the target information set to be recommended and the negative feedback information set.
14. An information pushing apparatus, comprising:
a first acquisition unit configured to acquire an information set, the information set including at least one piece of content information;
the acquisition unit is used for acquiring interactive feedback data of each content information in the information set and carrying out positive and negative feedback classification on the content information according to the distribution of the interactive feedback data to obtain a positive feedback information set and a negative feedback information set;
the second obtaining unit is used for obtaining an information set to be recommended, and the information set to be recommended comprises at least one piece of content information to be recommended;
the determining unit is used for determining a first matching score of each piece of content information to be recommended and each piece of content information in the positive feedback information set and a second matching score of each piece of content information to be recommended and each piece of content information in the negative feedback information set;
and the pushing unit is used for carrying out sequencing adjustment on each piece of content information to be recommended according to the first matching score and the second matching score to obtain a target information set to be recommended and pushing the target information set to be recommended.
15. A storage medium, characterized in that the storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute the steps of the information pushing method according to any one of claims 1 to 13.
CN202110727241.6A 2021-06-29 2021-06-29 Information pushing method and device and storage medium Active CN113254843B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110727241.6A CN113254843B (en) 2021-06-29 2021-06-29 Information pushing method and device and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110727241.6A CN113254843B (en) 2021-06-29 2021-06-29 Information pushing method and device and storage medium

Publications (2)

Publication Number Publication Date
CN113254843A true CN113254843A (en) 2021-08-13
CN113254843B CN113254843B (en) 2021-10-01

Family

ID=77190257

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110727241.6A Active CN113254843B (en) 2021-06-29 2021-06-29 Information pushing method and device and storage medium

Country Status (1)

Country Link
CN (1) CN113254843B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104281622A (en) * 2013-07-11 2015-01-14 华为技术有限公司 Information recommending method and information recommending device in social media
CN109189944A (en) * 2018-09-27 2019-01-11 桂林电子科技大学 Personalized recommending scenery spot method and system based on user's positive and negative feedback portrait coding
CN109241451A (en) * 2018-11-08 2019-01-18 北京点网聚科技有限公司 A kind of content combined recommendation method, apparatus and readable storage medium storing program for executing
US20190050494A1 (en) * 2017-08-08 2019-02-14 Accenture Global Solutions Limited Intelligent humanoid interactive content recommender
CN110097397A (en) * 2019-04-04 2019-08-06 北京字节跳动网络技术有限公司 Information-pushing method, device and electronic equipment based on feedback
CN110555112A (en) * 2019-08-22 2019-12-10 桂林电子科技大学 interest point recommendation method based on user positive and negative preference learning

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104281622A (en) * 2013-07-11 2015-01-14 华为技术有限公司 Information recommending method and information recommending device in social media
US20190050494A1 (en) * 2017-08-08 2019-02-14 Accenture Global Solutions Limited Intelligent humanoid interactive content recommender
CN109189944A (en) * 2018-09-27 2019-01-11 桂林电子科技大学 Personalized recommending scenery spot method and system based on user's positive and negative feedback portrait coding
CN109241451A (en) * 2018-11-08 2019-01-18 北京点网聚科技有限公司 A kind of content combined recommendation method, apparatus and readable storage medium storing program for executing
CN110097397A (en) * 2019-04-04 2019-08-06 北京字节跳动网络技术有限公司 Information-pushing method, device and electronic equipment based on feedback
CN110555112A (en) * 2019-08-22 2019-12-10 桂林电子科技大学 interest point recommendation method based on user positive and negative preference learning

Also Published As

Publication number Publication date
CN113254843B (en) 2021-10-01

Similar Documents

Publication Publication Date Title
CN110297848B (en) Recommendation model training method, terminal and storage medium based on federal learning
CN109902708B (en) Recommendation model training method and related device
CN110503531B (en) Dynamic social scene recommendation method based on time sequence perception
CN107038213B (en) Video recommendation method and device
CN110543598B (en) Information recommendation method and device and terminal
KR102112973B1 (en) Estimating and displaying social interest in time-based media
TW202007178A (en) Method, device, apparatus, and storage medium of generating features of user
US20200221181A1 (en) Content Recommendation System and Method-Based Implicit Ratings
CN111159564A (en) Information recommendation method and device, storage medium and computer equipment
CN110430471A (en) It is a kind of based on the television recommendations method and system instantaneously calculated
CN110490683B (en) Offline collaborative multi-model hybrid recommendation method and system
CN112241327A (en) Shared information processing method and device, storage medium and electronic equipment
KR101859620B1 (en) Method and system for recommending content based on trust in online social network
CN113643070A (en) Intelligent information pushing method and system based on big data
CN111597446A (en) Content pushing method and device based on artificial intelligence, server and storage medium
CN110032678A (en) Service resources method for pushing and device, storage medium and electronic device
WO2023087933A1 (en) Content recommendation method and apparatus, device, storage medium, and program product
CN117216362A (en) Content recommendation method, device, apparatus, medium and program product
CN113656681A (en) Object evaluation method, device, equipment and storage medium
CN114245185B (en) Video recommendation method, model training method, device, electronic equipment and medium
CN111177564B (en) Product recommendation method and device
CN113254843B (en) Information pushing method and device and storage medium
CN112905885A (en) Method, apparatus, device, medium, and program product for recommending resources to a user
CN115619503A (en) Article recommendation method and device, storage medium and computer equipment
CN114418701A (en) Method and device for generating recommendation list, electronic equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
REG Reference to a national code

Ref country code: HK

Ref legal event code: DE

Ref document number: 40052750

Country of ref document: HK