CN113626709B - Content recommendation method and server based on heat - Google Patents
Content recommendation method and server based on heat Download PDFInfo
- Publication number
- CN113626709B CN113626709B CN202110921287.1A CN202110921287A CN113626709B CN 113626709 B CN113626709 B CN 113626709B CN 202110921287 A CN202110921287 A CN 202110921287A CN 113626709 B CN113626709 B CN 113626709B
- Authority
- CN
- China
- Prior art keywords
- heat
- recommended content
- content
- recommended
- update period
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/43—Querying
- G06F16/435—Filtering based on additional data, e.g. user or group profiles
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The application provides a content recommendation method based on heat, which comprises the following steps: acquiring a first estimated heat and a second estimated heat of each recommended content, wherein the first estimated heat is determined according to a plurality of dimension characteristics of the recommended content, the plurality of dimension characteristics comprise display position characteristics, and the second estimated heat is determined according to the rest dimension characteristics of the recommended content except the display position characteristics; and calculating position deviation weight according to the first estimated heat and the second estimated heat, and obtaining the real heat generated in the last update period of the recommended content when a new update period is entered. Based on the change value of the real heat of the first two periods, a self-adaptive heat attenuation mode is provided, so that the heat is rapidly increased, the heat attenuation is reduced, the heat increase speed is obviously slowed down, and the heat attenuation is enhanced. And generating a recommendation list based on the heat calculation. The method and the device solve the problem that the recommendation list cannot be updated effectively in the process of updating the recommendation list based on the change of the heat.
Description
Technical Field
The present disclosure relates to the field of data processing technologies, and in particular, to a content recommendation method and a server based on heat.
Background
The content recommendation system is used for recommending related content to a user or guiding the user to use a specific function based on big data operation and user preference. Wherein, the content recommendation comprises media resource recommendation, book recommendation, commodity recommendation, music recommendation and the like. Taking the media asset recommendation scenario as an example, the content recommendation system may collect the number of times each media asset is selected by the user, such as the number of times the media asset is clicked by the user, the number of times the media asset is collected by the user, the number of times the media asset is shared by the user, the number of times the media asset is reviewed by the user, and the like. The number of times the asset is selected by the user is used as the heat. And sorting according to the heat degree of each media asset to form a recommendation list, and updating the recommendation list based on the heat degree change of the media asset, so that the media asset recommendation list is pushed to the user for further selection by the user.
However, after the plurality of assets are simply sorted according to the heat, if a certain asset is selected by the user more times, the heat of the assets of the asset is higher, and then the assets are placed in the front of the recommendation list, so that the probability of being selected is increased. Thus, the heat of the assets is not its true heat, but includes the heat of the assets caused by the difference in display positions. And further, in the process of updating the recommendation list based on the heat change of the media resource, the content recommendation system cannot effectively update the recommendation list, so that the use experience of a user is reduced.
Disclosure of Invention
The application provides a content recommendation method based on heat, which aims to solve the problem that a content recommendation system cannot effectively update a recommendation list and reduce the use experience of a user in the process of updating the recommendation list based on heat change.
In one aspect, the present application provides a content recommendation method based on popularity, applied to a server, including:
acquiring a first estimated heat and a second estimated heat of each recommended content, wherein the first estimated heat is determined according to a plurality of dimension characteristics of the recommended content, the plurality of dimension characteristics comprise display position characteristics, and the second estimated heat is determined according to the rest dimension characteristics of the recommended content except the display position characteristics;
calculating a position deviation weight according to the first estimated heat and the second estimated heat, and removing a heat error in the heat generated by the recommended content in the last update period by using the position deviation weight when a new update period is entered, so as to obtain the real heat generated by the recommended content in the last update period, wherein the heat error is generated by the display position characteristics of the recommended content;
and generating a recommended content list based on the real heat generated by each recommended content in the last updating period, wherein the recommended content list is used for being sent to the terminal equipment when a content recommendation request of the terminal equipment is received.
In another aspect, the present application provides a server including a memory and a processor, the memory storing program instructions, the processor executing the program instructions to perform the steps of:
acquiring a first estimated heat and a second estimated heat of each recommended content, wherein the first estimated heat is determined according to a plurality of dimension characteristics of the recommended content, the plurality of dimension characteristics comprise display position characteristics, and the second estimated heat is determined according to the rest dimension characteristics of the recommended content except the display position characteristics;
calculating a position deviation weight according to the first estimated heat and the second estimated heat, and removing a heat error in the heat generated by the recommended content in the last update period by using the position deviation weight when a new update period is entered, so as to obtain the real heat generated by the recommended content in the last update period, wherein the heat error is generated by the display position characteristics of the recommended content;
and generating a recommended content list based on the real heat generated by each recommended content in the last updating period, wherein the recommended content list is used for being sent to the terminal equipment when a content recommendation request of the terminal equipment is received.
As can be seen from the above technical solution, the embodiments of the present application provide a content recommendation method and a server based on heat, where the method first obtains a first estimated heat and a second estimated heat of each recommended content, calculates a position deviation weight according to the first estimated heat and the second estimated heat, and when a new update period is entered, removes a heat error generated by a display position feature of the recommended content in the heat generated by the recommended content in the previous update period by using the position deviation weight, so as to obtain a real heat generated by the recommended content in the previous update period. Based on the change value of the real heat in two adjacent historic periods, a self-adaptive heat attenuation mode is provided, so that the heat is rapidly increased, the heat attenuation is reduced, the heat rising speed is obviously slowed down, and the heat attenuation is enhanced. And generating a recommended content list based on the heat calculation mode, wherein the recommended content list is used for being sent to the terminal equipment when a content recommendation request of the terminal equipment is received. Therefore, according to the content recommendation method provided by the embodiment of the application, the heat error caused by the display position difference of the recommended content can be removed, so that the recommendation list can be updated according to the real heat change of the recommended content, the recommendation list can be effectively updated, and a more accurate recommended content list is provided for a user.
Drawings
For a clearer description of the technical solutions of the present application, the drawings that are required to be used in the embodiments will be briefly described, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without the inventive effort.
Fig. 1 exemplarily shows a usage scenario of a terminal device provided in an embodiment of the present application;
fig. 2 schematically illustrates a user interface of a terminal device provided in an embodiment of the present application;
fig. 3 is a schematic diagram schematically illustrating a media resource page interface in a terminal device according to an embodiment of the present application;
FIG. 4 is a schematic diagram schematically illustrating a media details page interface in a terminal device according to an embodiment of the present application;
FIG. 5 is a schematic diagram schematically illustrating another media details page interface in a terminal device according to an embodiment of the present application;
fig. 6 illustrates a flowchart of a content recommendation method performed by an embodiment of the present application.
Detailed Description
For purposes of clarity and implementation of the present application, the following description will make clear and complete descriptions of exemplary implementations of the present application with reference to the accompanying drawings in which exemplary implementations of the present application are illustrated, it being apparent that the exemplary implementations described are only some, but not all, of the examples of the present application.
It should be noted that the brief description of the terms in the present application is only for convenience in understanding the embodiments described below, and is not intended to limit the embodiments of the present application. Unless otherwise indicated, these terms should be construed in their ordinary and customary meaning.
The terms "first," second, "" third and the like in the description and in the claims and in the above-described figures are used for distinguishing between similar or similar objects or entities and not necessarily for limiting a particular order or sequence, unless otherwise indicated. It is to be understood that the terms so used are interchangeable under appropriate circumstances. The terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a product or apparatus that comprises a list of elements is not necessarily limited to all elements explicitly listed, but may include other elements not expressly listed or inherent to such product or apparatus. The term "module" refers to any known or later developed hardware, software, firmware, artificial intelligence, fuzzy logic, or combination of hardware or/and software code that is capable of performing the function associated with that element.
The terminal device in the embodiment of the application may include any one of a mobile phone, a tablet computer, a wearable device, a notebook computer and a display device. The present application will be described below with reference to a display device as an example.
Fig. 1 is a schematic diagram of a usage scenario of a display device according to an embodiment. As shown in fig. 1, a user may operate the display device 200 through the smart device 300 and the control apparatus 100.
In some embodiments, the control apparatus 100 may be a remote controller, and the communication between the remote controller and the display device includes infrared protocol communication or bluetooth protocol communication, and other short-range communication modes, etc., and the display device 200 is controlled by a wireless or other wired mode. The wireless mode can be direct connection or non-direct connection, and can be routed or not routed. The user may control the display device 200 by inputting user instructions through keys on a remote control, voice input, control panel input, etc. Such as: the user can input corresponding control instructions through volume up-down keys, channel control keys, up/down/left/right movement keys, voice input keys, menu keys, on-off keys, etc. on the remote controller to realize the functions of the control display device 200.
In some embodiments, the smart device 300 may also be used to control the display device 200. For example, the application running on the display device 200 is controlled using the smart device 300, or the display device 200 is controlled using the application running on the smart device 300. The application program, by configuration, can provide various controls to the user in an intuitive User Interface (UI) on a screen associated with the smart device.
In some embodiments, the smart device 300 and the display device 200 may also be used for communication of data.
In some embodiments, the display device 200 may also perform control in a manner other than the control apparatus 100 and the smart device 300, for example, the voice command control of the user may be directly received through a module configured inside the display device 200 device for acquiring voice commands, or the voice command control of the user may be received through a voice control apparatus configured outside the display device 200 device.
In some embodiments, the display device 200 is also in data communication with a server 400. The display device 200 may be permitted to make communication connections via a Local Area Network (LAN), a Wireless Local Area Network (WLAN), and other networks. The server 400 may provide various contents and interactions to the display device 200. The server 400 may be a cluster, or may be multiple clusters, and may include one or more types of servers.
In some embodiments, software steps performed by one step execution body may migrate on demand to be performed on another step execution body in data communication therewith. For example, software steps executed by the server may migrate to be executed on demand on a terminal device in data communication therewith, and vice versa.
In some embodiments, a "user interface" is a media interface for interaction and exchange of information between an application or operating system and a user that enables conversion between an internal form of information and a form acceptable to the user. A commonly used presentation form of the user interface is a graphical user interface (Graphic User Interface, GUI), which refers to a user interface related to computer operations that is displayed in a graphical manner. It may be an interface element such as an icon, a window, a control, etc. displayed in a display screen of the electronic device, where the control may include at least one of a visual interface element such as an icon, a button, a menu, a tab, a text box, a dialog box, a status bar, a navigation bar, a Widget, etc.
In general, a terminal device obtains corresponding data from a server by transmitting a data request to the server. For example, a content recommendation request is acquired from a server to acquire recommended content information from the server, where the recommended content information may be a recommended content list based on a hotness ranking, the recommended content list including content identifications, content names, and the like of a plurality of recommended contents. The user can obtain the latest updated recommended content list from the terminal equipment (such as a mobile phone) so as to conveniently select the content needed by the user.
Fig. 2 is a schematic diagram of a user interface of a terminal device (e.g., smart device 300 of fig. 1) according to some embodiments of the present application. In some implementations, a user may open a corresponding application by touching an application icon on the user interface, or may open a corresponding folder by touching a folder icon on the user interface.
In some embodiments, each application corresponds to a content recommendation system for recommending relevant content to a user or directing the user to use specific functions based on big data calculations and user preferences. The content recommendation system may be a server with which the terminal device requests recommended content, or a server cluster made up of a plurality of servers. In addition, the content recommendation system may also include a media asset library, a user database, and the like. Content recommendations include video recommendations, book recommendations, merchandise recommendations, music recommendations, and the like. For convenience of explanation, the following embodiments will take a media recommendation scenario in a video application as an example, and describe the functions of the content recommendation system. The content recommendation system may collect the number of times each media asset is selected by the user, such as the number of times the media asset is clicked by the user, the number of times the media asset is collected by the user, the number of times the media asset is shared by the user, the number of times the media asset is reviewed by the user, and the like.
In some embodiments, the content recommendation system uses the number of times the media asset is selected by the user as the heat of the media asset, sorts the heat of the media asset in the media asset library to form a recommended media asset list, and updates the recommended media asset list based on the heat change of the media asset. When receiving a media asset recommendation request sent by the terminal equipment, sending the latest recommended media asset list to the terminal equipment, so as to push the latest media asset recommendation list to the user for selection by the user. It should be noted that, the method is not limited to the use of the number of times the media asset is selected by the user to represent the heat of the media asset, but the number of times the media asset is clicked by the user, the number of times the media asset is collected by the user, the number of times the media asset is shared by the user and the number of times the media asset is reviewed by the user can be respectively given different weights, and the heat of the media asset is represented according to the corresponding weighted calculation result, so that the method can be designed according to practical situations.
Illustratively, when the user selects the video application, the controller controls the display to display a video application home page. The video application homepage displays interactive areas corresponding to one or more functions of the video application. As shown in fig. 3, a search box 330, a navigation bar 310, and a content display area 320 located below the navigation bar may be displayed in the interaction region, the content display area 320 including a plurality of media asset controls, such as "movie a", "movie B", and the like. The user can operate the control needing to be interacted through the control device, so that interaction with the interaction area is performed. The terminal device receives the user-entered interactions and responds to the detected interactions by performing corresponding operations.
When the terminal equipment receives the input operation of a user on a certain media resource control, details of the media resource and a recommendation list related to the media resource are displayed, wherein the recommendation list comprises a plurality of recommendation media resources. The user's operations on the assets include, but are not limited to, entering asset information for an asset in the search box, selecting an asset control in the content display area 320, etc. For ease of illustration, the user-selected assets will be referred to as target assets. The media asset information may be a name of the target media asset, a type of the target media asset, a time of delivery of the target media asset, a participating actor of the target media asset, and the like. The method includes the steps that after a user operates any one target media asset, the user is triggered to enter a corresponding target media asset detail page, and a plurality of associated recommended media assets recommended by a content recommendation system according to media asset information of the target media asset are displayed in a recommendation list in the target media asset detail page, wherein the plurality of recommended media assets are arranged according to a preset heat degree sequence so as to be convenient for the user to select.
FIG. 4 is a schematic diagram of an exemplary target movie details page interface, which may specifically be a movie details page entered into "movie A" after a user clicks on "movie A" in the content display area 320 of FIG. 3. As shown in fig. 4, the target media asset detail page includes a first display area 420 and a recommendation list 410, the first display area 420 including a plurality of media asset information controls, such as "movie a", "7.2 score", "action", "love" and "profile", etc. The recommendation list 410 includes a plurality of recommended assets associated with a target asset, wherein the plurality of recommended assets are arranged in a certain order to form the recommendation list. The user can watch by clicking any one of the recommended assets in the recommended list.
In specific implementation, the server receives a content recommendation request sent by the terminal device, where the content recommendation request is used to request the server to return recommended content associated with the target content, for example, the content recommendation request may be a media asset recommendation request, and the recommended content is recommended media asset. Next, the server obtains a recommended content set in response to the request, the recommended content set including a plurality of recommended content associated with the target content. A number of times that the plurality of recommended content is selected by the user is collected. The times selected by the user comprise the times of clicking the content by the user, the times of collecting the content by the user, the times of sharing the content by the user, the times of commenting the content by the user and the like. The initial popularity of each recommended content is determined based on the number of times the content is selected by the user. And sequencing the initial heat of each recommended content to obtain a recommended list. For example, the recommended content in the recommendation list is sequentially recommended media asset 1, recommended media asset 2, recommended media asset 3 and recommended media asset 4 according to the corresponding initial heat from high to low. The server transmits the recommendation list to the terminal device, and then the terminal device displays the recommendation list. The server transmits the recommendation result to the terminal equipment in the form of a recommendation list, and the terminal equipment displays the recommendation result to the user.
In some embodiments, when the recommendation list is displayed on the interface in the terminal device, the plurality of recommendation contents may be arranged in rows or columns. Further, the recommendation list returned to the terminal device by the server includes content information and display order information corresponding to each recommended content, and the display order information is, for example, arrangement order information of each recommended content in the recommendation list. If the recommended content is sequenced from high to low according to the corresponding initial heat, a recommended media resource 1, a recommended media resource 2, a recommended media resource 3 and a recommended media resource 4 form a recommended list, and the display position sequence information of the corresponding recommended media resource 1 is the media resource with the highest heat. Further, the terminal device may display a plurality of recommended contents according to the display position order information. Meanwhile, the user may perform a selection operation on the recommended content, and for example, the user may select any one of the plurality of recommended contents according to his own preference or personal habits.
In some embodiments, the recommendation list may be displayed in a media asset leaderboard format, including, but not limited to, based solely on the target media assets entered by the user. For example, when the user does not determine to select any target media assets, historical search details and a recommendation list for searching are displayed before media asset information is entered in the search box. The recommended list includes a plurality of recommended assets, and the recommended assets are not assets related to the target assets, but are any plurality of assets in the current asset library.
In specific implementation, the server receives a content recommendation request sent by the terminal device, where the content recommendation request is used to request the server to return a recommendation list. The server collects the number of times that all recommended content in the current media asset library is selected by the user. The initial popularity of each recommended content is determined based on the number of times the content is selected by the user. And sequencing the initial heat of each recommended content to obtain a recommended list, and sending the recommended list to the terminal equipment by the server.
In some embodiments, before entering the first update period, the number of times that each of the assets in the asset library is selected by the user is obtained in a preset time, and the number of times that each of the assets is selected by the user is characterized as a corresponding initial heat. And carrying out heat sequencing according to the initial heat corresponding to each media asset to generate a recommendation list. For example, before the update period is not entered, each media asset in the media asset library is issued in a manner of not being subjected to heat sorting, the number of times of each media asset selected by the user in one month is collected, and the number of times of each media asset selected by the user corresponding to the number of times is determined to be the initial heat. Wherein the plurality of assets in the asset library comprises various types of assets, such as: a plurality of media assets of the types of television drama, movies, variety, etc.
Referring to fig. 6, the content recommendation method based on heat provided by the present application may include the following steps:
s1, acquiring first estimated heat and second estimated heat of each recommended content, wherein the first estimated heat is determined according to a plurality of dimension features of the recommended content, the plurality of dimension features comprise display position features, and the second estimated heat is determined according to the rest dimension features of the recommended content except the display position features.
Each recommended content includes a plurality of dimension features, for example, referring to table 1, and the dimension features corresponding to the media include media delivery time, media type, media score, media display position, etc. The media asset display location is used to display the location of the media asset in the recommendation list. Exemplary, the display position of the media asset 1 is the first position, the display position of the media asset 2 is the second position, and the display position of the media asset 3 is the third position. In general, the recommended content may have a hotness error due to a display position, and a user of the recommended content with the display position in front may observe the recommended content preferentially, and the corresponding click rate may be relatively high. Therefore, the recommendation list is updated according to the real heat change of the recommended content by removing the heat error generated by the display position, and a more accurate recommendation list is provided for the user.
| Media assets/features | Media asset location xp | Media asset time x1 | Media asset type x2 | Media asset scoring x3 | Whether click on y |
| Media resource 1 | 1 | 2019 | Suspense doubt | 8.5 | Is that |
| Media resource 2 | 2 | 2018 | Xian Xia (fairy) | 6.5 | Is that |
| Media assets 3 | 3 | 2021 | Police robber | 7.2 | Whether or not |
TABLE 1
In some implementations, S1 includes: firstly, obtaining a weight coefficient corresponding to each dimension characteristic.
Illustratively, as shown in Table 1, the media asset delivery time is set to X1, the media asset type is set to X2, the media asset score is set to X3, and the media asset display position is set to Xp. Taking the example of the media asset 1, the media asset delivery time X1 may be a difference from the current time, for example, the delivery time is 2019, and the media asset delivery time X1 is 2021-2019=2. And similarly, carrying out normalization processing according to the production time X1, the media resource score X3 and the media resource display position Xp corresponding to the plurality of media resources, and carrying out unified dimension processing by using a first formula to obtain a characteristic value corresponding to each media resource characteristic. The first formula is as follows:
wherein x is min Is the minimum value of the corresponding feature, x max Is the maximum of the corresponding features.
Further, the media asset type X2 can be processed by a one-hot method, such as suspense type [1, 0], swordsman type [0,1,0], and police type [0, 1]; therefore, the data samples corresponding to media 1 are (xp: 0, x1:0.3, x2: [1, 0], x3:1, y: 1). Substituting a large number of data samples obtained by the process into a second formula for gradient descent training, and obtaining the weight coefficient corresponding to each feature.
The second formula is as follows:
y＝w 1 x 1 +w 2 x 2 +......+w p x p ；
wherein w1 is a weight coefficient corresponding to X1; w2 is a weight coefficient corresponding to X2; wp is the weight coefficient corresponding to Xp.
Further, after the weight coefficient corresponding to each feature is obtained, when the formula is verified, a click threshold value can be set to be 1, a data sample (xp: 0, x1:0.3, x2: [1, 0] and x3: 1) corresponding to the media asset 1 and the weight coefficient corresponding to each feature can be calculated to obtain a click prediction value y, and when the click prediction value is 1 or close to 1, clicking is predicted, otherwise, clicking is not performed.
In some implementations, S1 further includes: and carrying out weighted summation on all the dimension characteristics corresponding to the recommended content according to the weight coefficient corresponding to each dimension characteristic to obtain a first predicted value, and determining the first estimated heat according to the first predicted value. And carrying out weighted summation on the other dimensional characteristics according to the weight coefficients corresponding to the other dimensional characteristics except the display position characteristics to obtain a second predicted value, and determining a second estimated heat according to the second predicted value.
Illustratively, a plurality of dimensional characteristics of each recommended content, such as a media asset delivery time, a media asset type, a media asset score, a media asset display location, and the like, are obtained. And carrying all the dimension characteristics of the recommended content and the weight coefficient corresponding to each characteristic into a second formula to obtain a first predicted value y1 of the recommended content.
Acquiring multiple dimension characteristics of each recommended content, such as media asset delivery time, media asset type, media asset score, etc., it should be noted that the media asset display position is not included. And carrying the rest dimension characteristics of the recommended content which does not contain the media asset display position characteristics and the weight coefficient corresponding to each characteristic into a third formula to obtain a second predicted value y2 of the recommended content.
The third formula is as follows:
y＝w 1 x 1 +w 2 x 2 +......+w p-1 x p-1 ；
wherein w1 is a weight coefficient corresponding to X1; w2 is a weight coefficient corresponding to X2; wp-1 is the weight coefficient corresponding to Xp-1.
Further, the first predicted heat p1 and/or the second predicted heat p2 are obtained by substituting the first predicted value or the second predicted value into the fourth formula. The fourth formula is as follows:
wherein P is the estimated heat; y is the predicted value.
S2, calculating position deviation weight according to the first estimated heat and the second estimated heat, and removing heat errors in the heat generated in the last update period of the recommended content by using the position deviation weight when a new update period is entered, so as to obtain real heat generated in the last update period of the recommended content, wherein the heat errors are heat errors generated by display position characteristics of the recommended content;
the first estimated heat is compared with the second estimated heat to obtain a position deviation weight, and when a new updating period is entered, the actual heat generated in the last updating period of the recommended content is determined according to the position deviation weight. If the position deviation weight corresponding to the media asset 3 is p 2/p1=0.8. The preset update period is set to be one day, that is, each time of day, the heat ranking is performed according to the actual heat of each recommended content including the media asset 3. Wherein the true heat of each recommended content is determined by the true heat generated by the last update period. I.e. the heat ranking on the fourth day is based on the actual heat generated by each recommended content on the third day.
In specific implementation, acquiring heat generated by the recommended content in the last update period; and multiplying the position deviation weight by the heat generated by the recommended content in the last updating period to obtain the real heat generated by the recommended content in the last updating period. For example, if the number of times the asset 3 is selected by the user in the third day is 3000, the true heat generated in the third day is 3000×0.8=2400.
And S3, generating a recommended content list based on the real heat generated by each recommended content in the last updating period, wherein the recommended content list is used for being sent to the terminal equipment when a content recommendation request of the terminal equipment is received.
In some implementations, S3 includes: the initial heat of each recommended content is obtained, wherein the initial heat is the heat of each recommended content before entering the first updating period.
Illustratively, the number of times each recommended content is selected by the user during the historical time is taken as the initial heat before entering the first update period. Before entering the first updating period, collecting and counting 5000 times of total selection of the media assets 3 within one month of the user, and obtaining 5000 initial heat of the media assets 3.
In some implementations, S3 further includes: and adding the initial heat of the same recommended content and the real heat generated in the last updating period to obtain the current accumulated heat of the recommended content.
For example, based on the initial heat of the asset 3 being 5000 and the actual heat generated on the third day being 3000×0.8=2400, the current accumulated heat of the asset 3 is 5000+2400=7400.
In some implementations, S3 further includes: acquiring a first real heat degree generated by the recommended content in a first history update period and a second real heat degree generated by the recommended content in a second history update period, wherein the first history update period is a last update period, and the second history update period is a last update period of the first history update period; calculating the difference value between the first real heat and the second real heat and the ratio of the difference value to the second real heat to obtain the heat change trend of the recommended content; if the ratio is larger than a preset threshold, the heat change trend of the recommended content is a climbing trend; if the ratio is smaller than or equal to a preset threshold, the heat change trend of the recommended content is a decay trend. It should be noted that, the preset threshold is a specified positive number, and the magnitude of the preset threshold can be adjusted according to the time requirement.
Illustratively, the decay in heat, typically over time, is nonlinear. That is, when the change in heat over time is a climbing change, the attenuation intensity can be reduced. Conversely, when the change in heat over time begins to saturate or decline, the decay intensity may be enhanced. In practical situations, the freshness of the assets is continuously reduced with the lapse of time, but the corresponding freshness duration of each asset is different. For example, the selection times of the high-quality media resource users are more, so that the fresh time length of the high-quality media resource users can be prolonged and adjusted, and the high-quality media resource or more conversion can be realized. Meanwhile, in the process of rapidly rising the heat of the high-quality medium resource, the degree of decay of the heat along with time can be reduced. In contrast, non-quality media, without too great conversion potential over time, can increase the intensity of the time decay.
Taking the fourth day of updating heat ranking as an example, if the first real heat generated by the recommended content in the third day is obtained, the number of times that the media assets are selected by the user is 3000. And meanwhile, the second real heat generated by the recommended content on the whole day of the next day, namely the number of times of media resource selection by the user is 1000. Setting the preset threshold to be 1, and then obtaining the heat change trend of the recommended content to be (3000-1000)/1000= 2>1. Namely, the change trend of the recommended content is a rising trend, and the degree of heat decay with time can be adjusted slowly.
If the first real heat generated by the recommended content on the whole day of the third day is obtained, the number of times of media resource selection by the user is 1000. And meanwhile, the second real heat generated by the recommended content on the whole day of the next day, namely the number of times of media resource selection by the user is 2000. Setting a preset threshold value to be 1, and then obtaining the heat change trend of the recommended content to be (1000-2000)/2000= -0.5<1. Namely, the change trend of the recommended content is a decay trend, and the degree of heat decay with time can be adjusted to be larger.
In some implementations, S3 further includes: according to the heat increment of the recommended content in two adjacent history updating periods, determining a heat change trend of the recommended content and a heat attenuation function corresponding to the heat change trend, wherein the heat change trend comprises a climbing trend and an attenuation trend, the heat attenuation function is used for determining the heat attenuation intensity of the recommended content, and the heat attenuation intensity determined by the heat attenuation function corresponding to the climbing trend is smaller than the heat attenuation intensity determined by the heat attenuation function corresponding to the attenuation trend; and adjusting the current accumulated heat of the recommended content by using a heat decay function corresponding to the heat change trend of the recommended content to obtain the current real heat corresponding to each recommended content. The corresponding heat decay function is used for multiplying the obtained current accumulated heat, so that the current accumulated heat is subjected to decay intensity adjustment by using the heat decay function, and the current real heat of the medium resource 3 is obtained. And respectively calculating the current true heat of each media resource 1, 2, 3 and 4 in the recommendation list, and sequencing the calculated results according to the current true heat to generate a recommendation content list.
Illustratively, the decay function is:
f(t)＝e kw(t-t0) ；
wherein t is 0 Is the current time; t is the starting time of the content entering the content library; w is a weight parameter of the time decay intensity; k is a correction weight parameter for w, and the value of k is in direct proportion to the difference value between the ratio and the preset threshold value. In this example, the use is an exponential function, and the attenuation intensity is small at the initial stage and becomes larger as time passes. Can be set according to the actual situation.
And performing heat sequencing according to the current real heat corresponding to each recommended content to generate a recommended content list, wherein the recommended content list is used for being sent to the terminal equipment when a content recommendation request of the terminal equipment is received. The recommended content list obtained by the method in the embodiment can avoid the influence of the recommended display position on the click rate of the recommended content in practical application. For example, the probability of being observed by the user at each position in the recommendation list is different, the recommended content user with the display position in front can be preferentially observed, and finally, the corresponding click rate is naturally relatively high. Meanwhile, in the process of updating the recommendation list based on the content heat change, the recommendation list is effectively updated.
Fig. 5 is a schematic diagram of another exemplary media detail page interface, which may specifically be a media detail page entered into "movie a" after a user clicks on "movie a" in the content display area 320 of fig. 3. As shown in fig. 5, the target media asset detail page includes a first display area 500 and a recommendation list 510, and the recommendation list 510 includes a plurality of recommended media assets associated with the target media asset, wherein the plurality of recommended media assets are updated recommendation lists after ranking based on heat. Specifically, when a user sends a content recommendation request to a server at any time in an update period, the server returns an updated recommended content list to the terminal device, and the terminal device displays the updated recommended content list for the user to select.
Further, the present application provides a server, including a memory and a processor, where the memory stores program instructions, and the processor executes, by executing the program instructions, other program steps including execution or configuration of the processor in each embodiment of the terminal device, which are not described herein.
According to the technical scheme, the first estimated heat and the second estimated heat of each recommended content are obtained based on the heat-based content recommendation method provided by the embodiment of the application. And calculating position deviation weight according to the first estimated heat and the second estimated heat, and removing heat errors in the heat generated by the recommended content in the last updating period by using the position deviation weight when a new updating period is entered, so as to obtain the real heat generated by the recommended content in the last updating period, wherein the heat errors are heat errors generated by the display position characteristics of the recommended content. And generating a recommended content list based on the real heat generated by each recommended content in the last updating period, wherein the recommended content list is used for being sent to the terminal equipment when a content recommendation request of the terminal equipment is received. Therefore, the difference caused by the display positions can be removed, the recommendation list is updated according to the real heat change, the recommendation list is effectively updated, and a more accurate recommendation list is provided for the user.
The same and similar parts of the embodiments in this specification are referred to each other, and are not described herein. In a specific implementation, the present invention further provides a computer storage medium, where the computer storage medium may store a program, where the program may include some or all of the steps in each embodiment of the content recommendation method provided by the present invention when the program is executed. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (random access memory, RAM), or the like.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application. The illustrative discussions above are not intended to be exhaustive or to limit the embodiments to the precise forms disclosed above. Many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles and the practical application, to thereby enable others skilled in the art to best utilize the embodiments and various embodiments with various modifications as are suited to the particular use contemplated.
Claims (8)
1. A content recommendation method based on heat, applied to a server, comprising:
acquiring a plurality of dimension characteristics of each recommended content, wherein the plurality of dimension characteristics comprise display position characteristics;
acquiring a weight coefficient corresponding to each dimension characteristic;
weighting and summing all the dimension characteristics corresponding to the recommended content according to the weight coefficient corresponding to each dimension characteristic to obtain a first estimated value;
weighting and summing the other dimension characteristics according to the weight coefficients corresponding to the other dimension characteristics except the display position characteristics to obtain a second predicted value;
substituting the first predicted value or the second predicted value into the following formula to determine a first predicted heat or a second predicted heat:
wherein P is the estimated heat; y is a predicted value;
calculating a position deviation weight according to the first estimated heat and the second estimated heat, and removing a heat error in the heat generated by the recommended content in the last update period by using the position deviation weight when a new update period is entered, so as to obtain the real heat generated by the recommended content in the last update period, wherein the heat error is generated by the display position characteristics of the recommended content;
and generating a recommended content list based on the real heat generated by each recommended content in the last updating period, wherein the recommended content list is used for being sent to the terminal equipment when a content recommendation request of the terminal equipment is received.
2. The method of claim 1, wherein removing a heat error in a heat generated by the recommended content in a last update period using the position deviation weight to obtain a true heat generated by the recommended content in the last update period comprises:
acquiring the heat generated by the recommended content in the last updating period;
and multiplying the position deviation weight by the heat generated by the recommended content in the last updating period to obtain the real heat generated by the recommended content in the last updating period.
3. The method of claim 2, wherein generating the list of recommended content based on the true heat generated by each recommended content at the last update period comprises:
acquiring initial heat of each recommended content, wherein the initial heat is the heat of each recommended content before entering a first updating period;
calculating the current real heat of each recommended content according to the initial heat of each recommended content and the real heat generated by each recommended content in the last updating period;
and ranking each recommended content according to the current true hotness to generate the recommended content list.
4. The method of claim 3, wherein calculating the current true heat of each recommended content based on the initial heat of each recommended content and the true heat generated by each recommended content in the last update period comprises:
adding the initial heat of the same recommended content and the real heat generated in the last updating period to obtain the current accumulated heat of the recommended content;
according to the heat increment of the recommended content in two adjacent history updating periods, determining a heat change trend of the recommended content and a heat attenuation function corresponding to the heat change trend, wherein the heat change trend comprises a climbing trend and an attenuation trend, the heat attenuation function is used for determining the heat attenuation intensity of the recommended content, and the heat attenuation intensity determined by the heat attenuation function corresponding to the climbing trend is smaller than the heat attenuation intensity determined by the heat attenuation function corresponding to the attenuation trend;
and adjusting the current accumulated heat of the recommended content by using a heat decay function corresponding to the heat change trend of the recommended content to obtain the current real heat corresponding to each recommended content.
5. The method of claim 4, wherein determining a trend of heat change of the recommended content based on heat increases generated by the recommended content during two adjacent history update periods comprises:
acquiring a first real heat degree generated by the recommended content in a first historical update period and a second real heat degree generated by the recommended content in a second historical update period, wherein the first historical update period is the last update period, and the second historical update period is the last update period of the first historical update period;
calculating a difference value between the first real heat and the second real heat and a ratio of the difference value to the second real heat to obtain a heat change trend of the recommended content;
if the ratio is larger than a preset threshold, the heat change trend of the recommended content is the climbing trend; and if the ratio is smaller than or equal to the preset threshold, the heat change trend of the recommended content is the attenuation trend.
6. The method of claim 5, wherein the decay function is:
f(t)＝e kw(t-t0) ；
wherein t is 0 Is the current time; t is the starting time of the content entering the content library; w is a weight parameter of the time decay intensity; k is a correction weight parameter for w, and the value of k is in direct proportion to the difference value between the ratio and the preset threshold value.
7. The method of claim 1, wherein the positional deviation weight is a ratio of the first estimated heat to a second estimated heat.
8. A server comprising a memory and a processor, said memory having stored therein program instructions, said processor executing the steps of:
acquiring a plurality of dimension characteristics of each recommended content, wherein the plurality of dimension characteristics comprise display position characteristics;
acquiring a weight coefficient corresponding to each dimension characteristic;
weighting and summing all the dimension characteristics corresponding to the recommended content according to the weight coefficient corresponding to each dimension characteristic to obtain a first estimated value;
weighting and summing the other dimension characteristics according to the weight coefficients corresponding to the other dimension characteristics except the display position characteristics to obtain a second predicted value;
substituting the first predicted value or the second predicted value into the following formula to determine a first predicted heat or a second predicted heat:
wherein P is the estimated heat; y is a predicted value;
calculating a position deviation weight according to the first estimated heat and the second estimated heat, and removing a heat error in the heat generated by the recommended content in the last update period by using the position deviation weight when a new update period is entered, so as to obtain the real heat generated by the recommended content in the last update period, wherein the heat error is generated by the display position characteristics of the recommended content;
and generating a recommended content list based on the real heat generated by each recommended content in the last updating period, wherein the recommended content list is used for being sent to the terminal equipment when a content recommendation request of the terminal equipment is received.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110921287.1A CN113626709B (en) | 2021-08-11 | 2021-08-11 | Content recommendation method and server based on heat |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110921287.1A CN113626709B (en) | 2021-08-11 | 2021-08-11 | Content recommendation method and server based on heat |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113626709A CN113626709A (en) | 2021-11-09 |
| CN113626709B true CN113626709B (en) | 2023-05-05 |
Family
ID=78384599
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110921287.1A Active CN113626709B (en) | 2021-08-11 | 2021-08-11 | Content recommendation method and server based on heat |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113626709B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114302242B (en) * | 2022-01-25 | 2023-10-31 | 聚好看科技股份有限公司 | Media asset recommendation method, display equipment and server |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106227883A (en) * | 2016-08-05 | 2016-12-14 | 北京聚爱聊网络科技有限公司 | The temperature of a kind of content of multimedia analyzes method and apparatus |
| CN108062573A (en) * | 2017-12-29 | 2018-05-22 | 广东欧珀移动通信有限公司 | Model training method and device |
| CN109217957A (en) * | 2018-09-11 | 2019-01-15 | 广东翼卡车联网服务有限公司 | A kind of radio station push program method |
| CN109446419A (en) * | 2018-10-17 | 2019-03-08 | 武汉斗鱼网络科技有限公司 | A kind of method and device for recommending video |
| CN111079016A (en) * | 2019-12-17 | 2020-04-28 | 北京奇艺世纪科技有限公司 | Short video recommendation method and device and electronic equipment |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10089401B2 (en) * | 2016-01-12 | 2018-10-02 | Fox Broadcasting Company | Method and pluggable system for trend-based allocation of media assets between global and local storage |
-
2021
- 2021-08-11 CN CN202110921287.1A patent/CN113626709B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106227883A (en) * | 2016-08-05 | 2016-12-14 | 北京聚爱聊网络科技有限公司 | The temperature of a kind of content of multimedia analyzes method and apparatus |
| CN108062573A (en) * | 2017-12-29 | 2018-05-22 | 广东欧珀移动通信有限公司 | Model training method and device |
| CN109217957A (en) * | 2018-09-11 | 2019-01-15 | 广东翼卡车联网服务有限公司 | A kind of radio station push program method |
| CN109446419A (en) * | 2018-10-17 | 2019-03-08 | 武汉斗鱼网络科技有限公司 | A kind of method and device for recommending video |
| CN111079016A (en) * | 2019-12-17 | 2020-04-28 | 北京奇艺世纪科技有限公司 | Short video recommendation method and device and electronic equipment |
Non-Patent Citations (2)
| Title |
|---|
| Zhe Zhao等.Recommending what video to watch next: a multitask ranking system.RecSys '19: Proceedings of the 13th ACM Conference on Recommender Systems.2019,43-51. * |
| 马胜超.基于深度学习的推荐算法研究.中国优秀硕士学位论文全文数据库 (信息科技辑).2021,(第4期),I138-939. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113626709A (en) | 2021-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110781391B (en) | Information recommendation method, device, equipment and storage medium | |
| US20230394073A1 (en) | Real-time personalized suggestions for communications between participants | |
| KR101634773B1 (en) | Schedule management system and method using calendar | |
| US9189551B2 (en) | Method and apparatus for category based navigation | |
| JP5927290B2 (en) | Phrase-based communication system | |
| EP1679828B1 (en) | Method and system for prioritizing tasks made available by devices in a network | |
| US20210390087A1 (en) | Application suggestion features | |
| WO2018075334A1 (en) | Collaborator recommendation using collaboration graphs | |
| CN108153463B (en) | Application interface display control method and device, storage medium and mobile terminal | |
| US9922110B2 (en) | Information processing apparatus, information processing method, and program | |
| US10126918B2 (en) | Customized guided workflow element arrangement | |
| US20140059041A1 (en) | Graphical User Interface for Interacting with Automatically Generated User Profiles | |
| WO2018208806A1 (en) | System and method for arranging application icons on a mobile device | |
| JP5472118B2 (en) | Operation support method, operation support system, operation support apparatus, and operation support program | |
| JP6335210B2 (en) | Message service providing method for message service connected to search service, message server and user terminal for executing message service providing method | |
| CN111079016A (en) | Short video recommendation method and device and electronic equipment | |
| CN115455280A (en) | Recommendation list determining method and server | |
| CN113626709B (en) | Content recommendation method and server based on heat | |
| US20170212651A1 (en) | Short cut links in a graphical user interface | |
| CN114817736A (en) | Management method, device and equipment for display information in comprehensive energy service platform | |
| JP6890746B2 (en) | Information processing device, information processing method | |
| CA2842213C (en) | Method and apparatus for category based navigation | |
| JP2024502609A (en) | Providing ambient information and associated systems and devices based on learned user context and interactions | |
| CN113378068B (en) | Content recommendation method and server | |
| JP6539772B1 (en) | INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, PROGRAM, 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 |