CN117216402A - Original content recommendation method and device in game, electronic equipment and medium - Google Patents
Original content recommendation method and device in game, electronic equipment and medium Download PDFInfo
- Publication number
- CN117216402A CN117216402A CN202311284381.6A CN202311284381A CN117216402A CN 117216402 A CN117216402 A CN 117216402A CN 202311284381 A CN202311284381 A CN 202311284381A CN 117216402 A CN117216402 A CN 117216402A
- Authority
- CN
- China
- Prior art keywords
- content
- recent
- pool
- passive
- original
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 88
- 230000003068 static effect Effects 0.000 claims description 45
- 230000015556 catabolic process Effects 0.000 claims description 29
- 238000006731 degradation reaction Methods 0.000 claims description 29
- 238000004364 calculation method Methods 0.000 claims description 23
- 238000011156 evaluation Methods 0.000 claims description 15
- 238000005304 joining Methods 0.000 claims description 13
- 238000003860 storage Methods 0.000 claims description 12
- 230000007423 decrease Effects 0.000 claims description 11
- 238000012163 sequencing technique Methods 0.000 claims description 10
- 239000013589 supplement Substances 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 description 12
- 238000010586 diagram Methods 0.000 description 6
- 239000000047 product Substances 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 230000003993 interaction Effects 0.000 description 5
- 230000000750 progressive effect Effects 0.000 description 5
- 238000004891 communication Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000003247 decreasing effect Effects 0.000 description 3
- 238000013441 quality evaluation Methods 0.000 description 3
- 239000008280 blood Substances 0.000 description 2
- 210000004369 blood Anatomy 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000000593 degrading effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Abstract
The application provides an original content recommendation method, an original content recommendation device, electronic equipment and a medium in a game, wherein the method comprises the following steps: selecting recent original content from each first content pool of different levels according to a first preset recommendation proportion to recommend the recent original content to a user; determining recent inferior content according to feedback information of a user aiming at target recent original content, and removing the recent inferior content from a plurality of first content pools; the method comprises the steps of obtaining new recent original content to be supplemented to at least one target first content pool in a plurality of first content pools, recommending the new recent original content to a user according to recommendation rules, wherein the new recent original content is the recent original content which is not put into the plurality of first content pools. By adopting the original content recommendation method, the device, the electronic equipment and the medium in the game, the problems of cold start, excessive recommendation and narrow recommendation in the existing recommendation method are solved.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method, an apparatus, an electronic device, and a medium for recommending original content in a game.
Background
With the continued development of the game market, UGC (User Generated Content ) class games are gradually coming into the field of view of the user (player). Thanks to the evolution of game tools and engines, users can now more easily create and share their own created game content. In UGC games, as the number of UGC contents created by users is increased, the types are more and more abundant, a good UGC recommendation system can form a bridge between the users and the UGC contents, the interaction degree of the UGC games is increased, and how to recommend the proper UGC contents to different users becomes a vital link in the games. The existing recommended technical schemes are mainly divided into three types: the recommendation algorithm is a personalized recommendation mode, namely, content similar to the interest of a target user is found out by analyzing the behavior of the user, so that recommendation is performed.
However, the above recommendation method requires a large amount of user data to perform model training, and has serious cold start problems, and cannot effectively recommend against the newly registered user and the newly released content. Meanwhile, content similar to the interests of the user is recommended, and problems of over-recommendation and narrow recommendation are liable to occur.
Disclosure of Invention
In view of the above, the present application aims to provide a method, a device, an electronic device and a medium for recommending original content in a game, so as to solve the problems of cold start, over recommendation and narrow recommendation in the existing recommendation methods.
In a first aspect, an embodiment of the present application provides a method for recommending original content in a game, including:
selecting recent original content from each first content pool of different levels according to a first preset recommendation proportion to recommend the recent original content to a user;
determining recent inferior content according to feedback information of a user aiming at target recent original content, and removing the recent inferior content from a plurality of first content pools;
the method comprises the steps of obtaining new recent original content to be supplemented to at least one target first content pool in a plurality of first content pools, recommending the new recent original content to a user according to recommendation rules, wherein the new recent original content is the recent original content which is not put into the plurality of first content pools.
In a second aspect, an embodiment of the present application further provides an apparatus for recommending original content in a game, where the apparatus includes:
the first recommendation module is used for selecting recent original content from each first content pool with different grades according to a first preset recommendation proportion and recommending the recent original content to a user;
The content updating module is used for determining recent inferior content according to feedback information of a user aiming at target recent original content and removing the recent inferior content from a plurality of first content pools;
the second recommending module is used for acquiring new recent original content to be supplemented to at least one target first content pool in the first content pools, recommending the new recent original content to a user according to a recommending rule, wherein the new recent original content is the recent original content which is not put into the first content pools.
In a third aspect, an embodiment of the present application further provides an electronic device, including: the system comprises a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, the processor and the memory are communicated through the bus when the electronic device is running, and the machine-readable instructions are executed by the processor to perform the steps of the original content recommendation method in the game.
In a fourth aspect, embodiments of the present application also provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the original content recommendation method in a game as described above.
The embodiment of the application has the following beneficial effects:
according to the original content recommending method, device, electronic equipment and medium in the game, which are provided by the embodiment of the application, new recent original content can be directly obtained, and the new recent original content is recommended to a user, so that a newly registered user can obtain the new recent original content, and the new recent released content can be displayed, thereby avoiding cold start, simultaneously, the recent original content in different user preference intervals can be selected from a plurality of content pools to be recommended to the user, the recommending breadth is increased, and compared with the original content recommending method in the game in the prior art, the problems of cold start, excessive recommendation and narrow recommendation in the existing recommending method are solved.
In order to make the above objects, features and advantages of the present application more comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of an original content recommendation method in a game provided by an embodiment of the present application;
FIG. 2 is a schematic diagram of a process for updating recently original content in a plurality of first content pools according to an embodiment of the present application;
FIG. 3 illustrates a flow chart of calculating a favorites score provided by an embodiment of the present application;
fig. 4 is a schematic structural diagram of an apparatus for recommending original content in a game according to an embodiment of the present application;
fig. 5 shows a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. The components of the embodiments of the present application generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the application, as presented in the figures, is not intended to limit the scope of the application, as claimed, but is merely representative of selected embodiments of the application. Based on the embodiments of the present application, every other embodiment obtained by a person skilled in the art without making any inventive effort falls within the scope of protection of the present application.
The method for recommending original content in the game in one embodiment of the application can be operated on a local terminal device or a server. When the original content recommendation method in the game runs on a server, the method can be realized and executed based on a cloud interaction system, wherein the cloud interaction system comprises the server and the client device.
In an alternative embodiment, various cloud applications may be run under the cloud interaction system, for example: and (5) cloud game. Taking cloud game as an example, cloud game refers to a game mode based on cloud computing. In the running mode of the cloud game, a running main body of the game program and a game picture presentation main body are separated, storage and running of an original content recommendation method in the game are completed on a cloud game server, and the function of a client device is used for receiving and sending data and presenting game pictures, for example, the client device can be a display device with a data transmission function close to a user side, such as a mobile terminal, a television, a computer, a palm computer and the like; but the cloud game server which performs information processing is a cloud. When playing the game, the player operates the client device to send an operation instruction to the cloud game server, the cloud game server runs the game according to the operation instruction, codes and compresses data such as game pictures and the like, returns the data to the client device through a network, and finally decodes the data through the client device and outputs the game pictures.
In an alternative embodiment, taking a game as an example, the local terminal device stores a game program and is used to present a game screen. The local terminal device is used for interacting with the player through the graphical user interface, namely, conventionally downloading and installing the game program through the electronic device and running. The manner in which the local terminal device provides the graphical user interface to the player may include a variety of ways, for example, it may be rendered for display on a display screen of the terminal, or provided to the player by holographic projection. For example, the local terminal device may include a display screen for presenting a graphical user interface including game visuals, and a processor for running the game, generating the graphical user interface, and controlling the display of the graphical user interface on the display screen.
In a possible implementation manner, the embodiment of the application provides an original content recommendation method in a game, and a graphical user interface is provided through terminal equipment, wherein the terminal equipment can be the aforementioned local terminal equipment or the aforementioned client equipment in a cloud interaction system.
Referring to fig. 1, fig. 1 is a flowchart of an original content recommendation method in a game according to an embodiment of the present application. As shown in fig. 1, the method for recommending original content in a game according to the embodiment of the present application includes:
Step S101, selecting recent original content from each first content pool of different levels according to a first preset recommendation proportion, and recommending the recent original content to a user.
In this step, the first preset recommendation ratio may refer to a recommendation ratio between recently created contents respectively selected from the plurality of first content pools, the first preset recommendation ratio characterizing a recommendation weight of each first content pool.
The first content pools may refer to resource pools storing recently authored content, the recently authored content stored between the respective first content pools being different, one recently authored content being stored in only one of the plurality of first content pools.
Illustratively, the plurality of first content pools includes a first content pool D, a first content pool C, a first content pool B, a first content pool a, a first content pool S.
The user preference interval may refer to a user preference range corresponding to each first content pool, and the user preference interval corresponding to each first content pool is determined by recent original content stored in the first content pool.
Assuming that 100 recent original contents are stored in the first content pool C, the user preference of the 100 recent original contents is maximum value 10000 and minimum value 1000, and the user preference interval corresponding to the first content pool C is 1000 to 10000.
Recent original content may refer to original content that was released by a user during the last period of time, where the last period of time is a set time range, such as: the original content released by the user in the last 7 days is taken as the recent original content.
The original content can refer to original content of a user in a game, the user can share the content created by the user to other users for experience in UGC games, and the content available for sharing experience is called original content.
By way of example, the original content may be game level, map, copy, or pet, article, original by the user in the UGC game.
In the embodiment of the application, in order to timely recommend recently released original contents to a user, a certain number of recently released original contents can be randomly selected from the recently released original contents stored in the first content pool for each first content pool so as to be recommended to the user, so that the original contents recommended to the user are not only recently released contents, but also different recently released original contents are in different user preference intervals, the diversity of content recommendation is increased, and the problem of narrow recommendation is avoided. Meanwhile, even for newly registered users or new users who have just entered the game, newly released original content can be effectively recommended to them.
In an example, content information of recent original content selected from the plurality of first content pools may be recommended to the user in response to a selection operation of the user on the content recommendation interface, or the system may actively recommend content information of recent original content selected from the plurality of first content pools to the user when a specific condition is satisfied.
Taking the above example as an example, assuming that the first preset recommended proportions of the first content pool D, the first content pool C, the first content pool B, the first content pool a, and the first content pool S are respectively 10%, 20%, 35%, 25%, 10%, and when selecting the recent original content, a total of 100 recent original contents need to be selected for recommendation to the user, 100×10% =10 recent original contents are selected from the first content pool D, 100×20% =20 recent original contents are selected from the first content pool C, 100×35% =35 recent original contents are selected from the first content pool B, 100×25% =25 recent original contents are selected from the first content pool a, and 100×10% =10 recent original contents are selected from the first content pool S.
It should be noted that, the user preference interval corresponding to a first content pool is not fixed, and will change with the change of the recent original content stored in the first content pool.
Step S102, determining recent inferior content according to feedback information of a user aiming at target recent original content, and removing the recent inferior content from a plurality of first content pools.
In this step, the feedback information may refer to information of a feedback operation performed by the user on the target recent original content, where the feedback information is used to characterize the satisfaction degree of the user on the target recent original content, and the feedback information includes, but is not limited to praise, collection, sharing and play.
The recent bad content may refer to a preset number of recent original contents having the lowest user preference ranking among all the recent original contents stored in the plurality of first content pools.
In the embodiment of the application, the user acceptance of the recommended content is ensured while the newly issued original content in the near-term is recommended to the user, therefore, the user preference evaluation is carried out on the target recent original content according to the feedback operation of different users on the target recent original content, so that the recent inferior content is determined according to the user preference, the near-term inferior content is removed from the first content pool, and the repeated recommendation of the recent original content with lower user preference to the user is avoided.
Taking original content as a game level as an example, after recommending the game level released in the recent period to a user, part of the users can view the detailed information of the recommended game level in a game recommendation page, for example: the user can select whether to play the checkpoint according to own favorite preference, and can perform feedback operations such as praise, sharing, collection, evaluation and the like on the checkpoint after playing the checkpoint so as to express own favorites on the checkpoint. According to feedback operations of different users, comprehensive favorites of the users on a certain recent original content can be comprehensively judged, and according to the comprehensive favorites ranking of different recent original contents, recent inferior content is selected from all the recent original contents stored in the plurality of first content pools, so that the recent inferior content is removed from the plurality of first content pools.
Step S103, new recent original content is obtained to be supplemented to at least one target first content pool in the plurality of first content pools, and the new recent original content is recommended to the user according to the recommendation rule.
In this step, the new recent original content is recent original content that has not been put into the plurality of first content pools. The new recent original content is recently released original content that has not been recommended to the user.
The recommendation rules may refer to rules that recommend new recent original content, which are used to ensure that new recent original content will necessarily be recommended to the user.
In a first example, the recommended rule is a traffic decay rule, where the traffic decay rule refers to setting a basic traffic for a new recent original content, and exposing the new recent original content is not put in when the current remaining traffic is 0, each time the new recent original content is exposed once the current remaining traffic is equal to the basic traffic minus one. The flow attenuation rule is more suitable for scenes in which the user actively acquires recommended content.
In the second example, the recommendation rule is a push number rule, and the recommendation number rule refers to recommending the new recent original content according to the set recommendation number, and the remaining recommendation number is reduced by one as long as the new recent original content is recommended once, and is not recommended when the remaining recommendation number is 0, regardless of whether the new recent original content is exposed or not. The recommendation frequency rule is more suitable for scenes in which the user passively acquires the recommended content.
In the embodiment of the application, in order to timely update the recent original content in the plurality of first content pools so as to recommend the recent original content newly released by the user to the user, the newly released recent original content can be stored in the first resource library according to the sequence of release time. Then, the first content pool may be periodically detected, when new recent original content needs to be replenished, the earliest released recent original content is obtained from the first resource pool, the earliest released recent original content is added to the first content pool, the earliest released recent original content is removed from the first resource pool, or the earliest released recent original content is marked to be added to the first content pool, so that the earliest released recent original content is prevented from being repeatedly added to the first content pool.
Here, when new recent original content is added to the plurality of first content pools, it may be added to only one of the plurality of first content pools, or it may be added to the plurality of first content pools, depending on the user preference of the added new recent original content.
In addition, in the case where the plurality of first content pools join new recent original content and the plurality of first content pools are not updated, the user preference interval of the recent original content stored in the first content pool D may overlap with the user preference interval of the recent original content stored in the other first content pools, but the user preference intervals corresponding to different first content pools do not overlap after the plurality of first content pools are updated. Wherein, the plurality of first content pool updates means updating the recently original content stored in the first content pool with the new recently original content newly added into the first content pool to remove the recently inferior content stored in the first content pool.
In the embodiment of the present application, the plurality of first content pools includes a first active content pool D and a plurality of first passive content pools, where the plurality of first passive content pools are a first passive content pool C, a first passive content pool B, a first passive content pool a, and a first passive content pool S, respectively. After the recent inferior content is removed from the first content pool, new recent original content is added into the first content pool, and the newly added recent original content firstly enters the first active content pool and then enters a plurality of first passive content pools. The update process of recently created content in a plurality of first content pools is described below in connection with fig. 2. The active content pool is a content pool for actively acquiring original content. The passive content pool refers to a content pool in which original content is passively acquired.
Fig. 2 is a schematic diagram of a recent original content update procedure in a plurality of first content pools according to an embodiment of the present application.
As shown in fig. 2, in the present example, the at least one first target content pool includes a first active content pool D, and acquiring new recently released original content supplements to the at least one target first content pool of the plurality of first content pools includes: when the number of the recent original contents in the first active content pool D is smaller than the content pool number threshold corresponding to the first active content pool D, acquiring new recent original contents and adding the new recent original contents into the first active content pool D so that the number of the recent original contents in the first active content pool D is equal to the content pool number threshold corresponding to the first active content pool D.
In order to timely supplement the first active content pool D, detecting whether the number of recently created contents in the first active content pool D is smaller than the content pool number threshold 1600 of the first active content pool D every 30 seconds, if the number of recently created contents in the first active content pool D is smaller than 1600, scanning the first resource pool to obtain new recently created contents, and obtaining a difference value between the number of new recently created contents and the current number of recently created contents stored in the first active content pool to make the number of recently created contents in the first active content pool D equal to 1600. The time of detecting the first active content pool D and the content pool number threshold of the first active content pool D are set values, and a person skilled in the art may select specific values of the detection time and the content pool number threshold according to actual situations, which is not limited herein.
In an alternative embodiment, after acquiring new recent original content to be added to the first active content pool D, the method includes: steps a1 to a3.
Step a1, setting basic flow for new recent original content, and setting the new recent original content as an exposure state.
In order to enable new recent original content in the first active content pool D to be effectively recommended to the user, a base traffic is set for each new recent original content added to the first active content pool D, for example: setting the basic flow of each new recent original content as 500, setting the basic flow, setting each new recent original content as an exposure state, and recommending the original content in the exposure state to a user.
And a step a2, when the new recent original content is browsed by the user, subtracting one from the basic flow of the new recent original content in the exposure state.
Taking a new recent original content a as an example, after recommending the recent original content a to the user, if the time that the user browses the recent original content in the recommended page is greater than or equal to 1 second, determining that exposure is once, and at this time, subtracting one from the basic flow of the recent original content a in the exposure state.
Step a3, when the basic flow of the new recent original content is reduced to 0, changing the state of the new recent original content from the exposure state to the waiting state, wherein the recent original content in the waiting state is not recommended to the user.
When the basic flow of the recent original content a is reduced to 0 after the recent original content a is browsed 500 times, the state of the recent original content a is changed from the exposure state to the waiting state, and the recent original content a in the waiting state is not recommended to the user.
The purpose of the waiting period is that the playing time of some game levels may be long because the playing time of different game levels is different, and the feedback information is uploaded only after the playing is finished, so that the waiting period needs to be set to collect the feedback information of the player as completely as possible. In one example, the duration of the waiting period is 1 hour, that is, the number of plays, endorsements, collections, and shares of the recent original content a are accumulated and recorded on the recent original content a within 1 hour.
In an alternative embodiment, step S102 includes: steps b1 to b4.
Step b1, determining the number of recently authored content in a ranking state in the first active content pool.
After the original content is changed from the exposure state to the waiting state, calculating the accumulated time length of the recent original content entering the waiting state, ending the waiting state when the accumulated time length reaches a waiting time length threshold value, and changing the new recent original content from the waiting state to the ranking state.
Every time one original content enters the ranking state from the waiting state, the number of recent original contents in the ranking state in the first active content pool is counted, so that whether to start scoring the favorites of the original contents in the ranking state is determined according to the number. To ensure fairness, when the original content enters a ranking state, feedback information collection is no longer performed.
And b2, determining the preference score of the recent original content in the ranking state when the number of the recent original content in the ranking state reaches a first score number threshold.
When the number of the original contents in the ranking state in the first active content pool D reaches the first scoring number threshold value 100, a preference score calculation is performed to calculate a preference score of each recent original content in the ranking state. Wherein the preference score is determined by feedback information of the user for the original content, for example: the more the sharing number and the praise number are, the higher the preference score is; the fewer the sharing numbers and the endorsements, the lower the favorites score.
And b3, taking a first set number of recently created contents with top favorites scores in the first active content pool as recently good-quality contents, and transferring the recently good-quality contents to a plurality of first passive content pools.
The original content with higher preference score is upgraded to a first higher-level content pool, namely the first M original contents with higher preference score are used as near-term high-quality content, M near-term high-quality contents are put into a first passive content pool C, and M near-term high-quality contents are removed from a first active content pool D.
The plurality of first content pools are a first active content pool D, a first passive content pool C, a first passive content pool B, a first passive content pool A and a first passive content pool S in sequence from low to high in level, namely the first active content pool D is the first content pool with the lowest level, and the levels of the first passive content pools are all higher than those of the first active content pool. The level of the first content pool corresponds to the preference score of original content in the content pool, and the higher the preference score of the original content is, the higher the level of the first content pool is; the lower the favorites score of the original content, the lower the level of the first content pool.
And b4, removing the near-term inferior content from the plurality of first content pools by utilizing the near-term superior content.
And after M pieces of recent high-quality content are put into the first passive content pool and removed from the first active content pool, deleting the original content remained in the first active content pool D, and completing one-round updating of the first active content pool D. At the same time, near-term bad content in the first plurality of passive content pools is also determined to remove near-term bad content from the first plurality of passive content pools.
In an alternative embodiment, the plurality of first passive content pools have corresponding levels from high to low according to the recommendation level of the recently created content stored in each of the plurality of first passive content pools, and the removing the recently bad content from the plurality of first passive content pools by using the recently good content includes: steps c1 to c3.
Here, since the preference score of new recent original content added to the first passive content pool C may be higher than the preference scores of recent original content in other first passive content pools, it is determined for this purpose in a progressive upgrade manner which first passive content pool the newly added recent original content should be deposited into.
Because the recent original content has a progressive upgrading process, the scoring range of the favorites in each first content pool is not fixed, and in order to facilitate the description of the level of the first content pool, the level of the first content pool can be determined according to the recommendation degree of the recent original content stored in the first content pool, and the higher the recommendation degree is, the higher the level is. The higher the overall preference score, the higher the recommendation level; the lower the overall preference score, the lower the recommendation level.
Step c1, adding the first set quantity of recent high-quality content into the first passive content pool of the lowest level.
Taking the above example as an example, if M pieces of recent premium content are put into the plurality of first passive content pools, the first set number is M. When M pieces of recent high-quality content are put into the plurality of first passive content pools, the M pieces of recent high-quality content are put into the first passive content pool C first, that is, the first passive content pool of the lowest level, because the first passive content pool of the lowest level stores recent original content with the lowest grade of preference score in all the first passive content pools.
Step c2, starting from the first passive content pool with the lowest level added with the first set number of the recent high-quality content, progressively upgrading the recent high-quality content in each first passive content pool to the first passive content pool with the first level higher until the recent high-quality content in the first passive content pool with the second highest level is upgraded to the first passive content pool with the highest level.
Comparing the M near-term high-quality contents with near-term original contents stored in the first passive content pool C to determine which near-term high-quality contents in the first passive content pool C can be upgraded into the first passive content pool B, and selecting near-term high-quality contents which can be upgraded into the first passive content pool B from the first passive content pool C to be called first near-term high-quality contents.
Then, the first near-term high-quality content which is lifted into the first passive content pool B is compared with the near-term original content which is already stored in the first passive content pool B, which near-term high-quality content in the first passive content pool B can be lifted into the first passive content pool A is determined, and the near-term high-quality content which is selected from the first passive content pool B and can be lifted into the first passive content pool A is called second near-term high-quality content.
And so on until the third recent premium content in the first passive content pool a is upgraded into the first passive content pool S.
Step c3, starting from the first passive content pool with the highest level of the added recent high-quality content, progressively degrading the recent low-quality content in each first passive content pool to the first passive content pool with the lower level until the first set number of recent low-quality content is removed from the first passive content pool with the lowest level.
Determining a first recent inferior content in the first passive content pool S, transferring the first recent inferior content into the first passive content pool a, determining which recent inferior content in the first passive content pool a can be downgraded into the first passive content pool B, and calling the recent inferior content which can be downgraded into the first passive content pool B and is selected from the first passive content pool a as a second recent inferior content.
Then, determining which recent bad content in the first passive content pool B can be demoted to the first passive content pool C, and designating the recent bad content selected from the first passive content pool B that can be demoted to the first passive content pool C as third recent good content.
And so on until the fourth, later inferior content in the first passive content pool C is removed from the first passive content pool C. In the process of one round of upgrading and downgrading, a single recent high-quality content can only be upgraded once, and a single recent low-quality content can only be downgraded once, so that the situations that the single recent high-quality content is continuously upgraded in one round and the single low-quality content is continuously downgraded in one round are avoided.
In an alternative embodiment, step c2 comprises: step c21 to step c24.
And c21, selecting the first passive content pool with the lowest level as a first upgrading passive content pool.
When the recent high-quality content upgrading is executed, the first passive content pool C is first used as a first upgrading passive content pool.
Step c22, determining whether the first upgraded passive content pool is the highest level first passive content pool.
Step c23, if the first updated passive content pool is not the highest-level first passive content pool, selecting near-term high-quality content from near-term original content in the first updated passive content pool, and transferring the near-term high-quality content to the first passive content pool which is one level higher than the first updated passive content pool.
Since the first passive content pool C is not the highest-level first passive content pool, selecting a first recent candidate premium content from recently authored contents of the first passive content pool C when the condition is satisfied, selecting a first recent premium content from the first recent candidate premium content, and then transferring the first recent premium content from the first passive content pool C to the first passive content pool B.
If the first passive content pool is the highest level first passive content pool, the recent high-quality content promotion process ends and the recent low-quality content demotion process begins to be performed.
Step c24, taking the first passive content pool with the higher level as a new first upgrading passive content pool, and returning to execute the step of selecting the recently high-quality content from the recently original content in the first upgrading passive content pool.
After the first recent high-quality content is transferred to the first passive content pool B, the first passive content pool B is used as a new first upgrading passive content pool, and the second recent high-quality content is continuously selected from the first passive content pool B.
After the first recent high-quality content is transferred to the first passive content pool B, selecting a second recent candidate high-quality content from recent original content of the first passive content pool B when the condition is met, selecting the second recent high-quality content from the second recent candidate high-quality content, and transferring the second recent high-quality content from the first passive content pool B to the first passive content pool A.
Then, the first passive content pool A is used as a new first upgrading passive content pool, when the condition is met, third recent candidate high-quality content is selected from recent original content of the first passive content pool A, third recent high-quality content is selected from the third recent candidate high-quality content, and the third recent high-quality content is transferred from the first passive content pool A to the first passive content pool S.
Since the first passive content pool S is the highest-level first passive content pool, the recent high-quality content promotion process ends, and the recent low-quality content demotion process starts to be performed.
In an alternative embodiment, step c3 includes: steps c31 to c34.
And c31, selecting the first passive content pool with the highest level as a first degraded passive content pool.
When recent degradation of bad content is performed, the first passive content pool S is first taken as the first degraded passive content pool.
Step c32, determining whether the first degraded passive content pool is the lowest level first passive content pool.
And c33, if the first degraded passive content pool is not the first passive content pool with the lowest level, selecting the recent inferior content from the recent original content in the first degraded passive content pool, and transferring the recent inferior content to the first passive content pool with the lower level of the first degraded passive content pool.
The first passive content pool S receives third recent high-quality content upgraded by the first passive content pool A, and as the first passive content pool S is not the first passive content pool with the lowest level, when the condition is met, first recent candidate inferior content is selected from recent original content of the first passive content pool S, first recent inferior content is selected from the first recent candidate inferior content, and the first recent inferior content is downgraded from the first passive content pool S to the first passive content pool A.
And c34, taking the first passive content pool with the lower level as a new first degraded passive content pool, and returning to execute the step of selecting the recent inferior content from the recent original content in the first degraded passive content pool.
And taking the first passive content pool A as a new first degraded passive content pool, selecting a second recent candidate inferior content from recent original contents of the first passive content pool A when the condition is met, selecting the second recent inferior content from the second recent candidate inferior content, and transferring the second recent inferior content from the first passive content pool A to the first passive content pool B.
Taking the first passive content pool B as a new first degraded passive content pool, selecting a third recent candidate inferior content from recent original contents of the first passive content pool B when the condition is met, selecting the third recent inferior content from the third recent candidate inferior content, and transferring the third recent inferior content from the first passive content pool B to the first passive content pool C.
And taking the first passive content pool C as a new first degraded passive content pool, wherein the first passive content pool C is the first passive content pool with the lowest level, when the condition is met, selecting a fourth near-term candidate inferior content from the recently original content of the first passive content pool C, selecting the fourth near-term inferior content from the fourth near-term candidate inferior content, removing the fourth near-term inferior content from the first passive content pool C, and completing a round of removal process of the recently inferior content.
In an alternative embodiment, step c23 selects recently premium content from recently authored content in the first upgraded passive content pool, comprising: step c231 to step c233.
And step c231, when the evaluation condition is met, sorting according to the time sequence of the recent original content entering the first upgrading passive content pool, and selecting the top-ranked recent original content with a fixed proportion as the recent candidate high-quality content.
The rating condition may refer to a condition for scoring favorites. In the embodiment of the application, the evaluation condition is that the recently original content in the first upgrading passive content pool reaches the corresponding content pool quantity threshold value.
Taking the first upgrading passive content pool as a first passive content pool C as an example, recording the time when new recent original content enters the first passive content pool C, and sorting all the recent original content stored in the first passive content pool C according to the time sequence of entering the content pool when the recent original content in the first passive content pool C reaches the content pool quantity threshold of the first passive content pool C. The top 50% of recent original content is selected as the first recent candidate premium content.
Taking the first upgrading passive content pool as a first passive content pool B as an example, after transferring the first recent high-quality content to the first passive content pool B, recording the time when the first recent high-quality content enters the first passive content pool B, when the recent original content in the first passive content pool B reaches the content pool quantity threshold of the first passive content pool B, sorting all the recent original contents stored in the first passive content pool B according to the time sequence of entering the content pool, and selecting the recent original content with the ranking of 50% as a second recent candidate high-quality content.
Step c232, determining the current preference score of the recent candidate premium content.
And acquiring the latest feedback information of the recent candidate high-quality content, and calculating the current preference score of the recent candidate high-quality content according to the latest feedback information and the preference scoring formula.
And step c233, sorting the recent candidate high-quality contents according to the order of the current preference scores from high to low, and selecting the recent candidate high-quality contents with preset upgrade numbers corresponding to the first upgrade passive content pool in the first set number ranked at the front as final recent high-quality contents.
The preset upgrade number corresponding to each first passive content pool is different, taking the first set number as M as an example, the preset upgrade number corresponding to the first passive content pool C is M/2, the preset upgrade number corresponding to the first passive content pool B is M/4, and the preset upgrade number corresponding to the first passive content pool A is M/8.
Taking the first upgrading passive content pool as the first passive content pool C as an example, if 16 new recent original contents are added to the first passive content pool C, 8 first recent high-quality contents are selected from the first passive content pool C and transferred to the first passive content pool B, and the 8 first recent high-quality contents are selected as the recent candidate high-quality contents of the preset upgrading number corresponding to the first passive content pool C, which are ranked at the top, and are also the final recent high-quality contents selected from the first passive content pool C.
In an alternative embodiment, step c33 selects recently bad content from recently original content in the first degraded passive content pool, comprising: step c331 to step c333.
And c331, when the evaluation condition is met, sorting according to the time sequence of the recent original content entering the first degradation passive content pool, and selecting the recent original content with a fixed proportion and the top ranking as the recent candidate inferior content.
The rating condition may refer to a condition for scoring favorites. In the embodiment of the application, the evaluation condition is that the recently original content in the first degraded passive content pool reaches the corresponding content pool quantity threshold value.
Taking the first degraded passive content pool as a first passive content pool A for example, the first passive content pool A receives M/8 first recent bad content degraded by the first passive content pool S, and records the time when the first recent bad content enters the first passive content pool A. And when the recent original content in the first passive content pool A reaches the content pool quantity threshold value of the first passive content pool A, sequencing all the recent original content stored in the first passive content pool A according to the time sequence of entering the content pool. And selecting the first 50% of recent original content as second recent candidate inferior content.
Taking the first degraded passive content pool as a first passive content pool B for example, the first passive content pool B receives M/4 second recent inferior contents degraded by the first passive content pool A, records the time when second recent superior contents enter the first passive content pool B, and when the recent original contents in the first passive content pool B reach the threshold value of the number of the content pools of the first passive content pool B, sequences all the recent original contents stored in the first passive content pool B according to the time sequence of the entering content pools, and selects the recent original contents with the ranking of 50% as second recent candidate inferior contents.
Step c332, determining the current preference score of the recent candidate inferior content.
And acquiring the latest feedback information of the recent candidate inferior content, and calculating the current preference score of the recent candidate inferior content according to the latest feedback information and the preference scoring formula.
And step c333, sorting the recent candidate inferior contents according to the order of the current preference scores from high to low, and selecting the recent candidate inferior contents with the preset degradation number corresponding to the first degradation passive content pool in the first set number after ranking as the final recent inferior contents.
The preset degradation number corresponding to each first passive content pool is different, taking the first set number as M as an example, the preset degradation number corresponding to the first passive content pool S is M/8, the preset degradation number corresponding to the first passive content pool A is M/4, the preset degradation number corresponding to the first passive content pool B is M/2, and the preset degradation number corresponding to the first passive content pool C is M.
Taking the first degraded passive content pool as the first passive content pool C as an example, if 16 new recent original contents are added to the first passive content pool C, 16 fourth recent inferior contents are selected to be deleted from the first passive content pool C.
It can be seen that if all the first passive content pools are regarded as a whole, in the first passive content pool combination, every M pieces of recent original content are input, the M pieces of recent original content are eliminated, so that the stability of the quantity of the recent original content of the first passive content pool combination is ensured.
In an alternative embodiment, the value of the preset upgrade number corresponding to the plurality of first passive content pools gradually decreases with an increase in the level of the first passive content pools.
Taking the above example as an example, if the first set number is M, the preset upgrade number corresponding to the first passive content pool C is M/2, the preset upgrade number corresponding to the first passive content pool B is M/4, and the preset upgrade number corresponding to the first passive content pool a is M/8. It can be seen that the value of the preset number of upgrades is reduced by half each time the level of the first passive content pool is increased by one step, i.e. gradually as the level of the first passive content pool increases. It should be noted that the reduction range of the preset upgrade number is not fixed to be half, and those skilled in the art may select the reduction range of the preset upgrade number according to actual situations, which is not limited herein.
In an alternative embodiment, the magnitude of the preset degradation amount corresponding to the plurality of first passive content pools gradually increases with the decrease of the first passive content pool level.
Taking the above example as an example, if the first set number is M, the preset degradation number corresponding to the first passive content pool S is M/8, the preset degradation number corresponding to the first passive content pool a is M/4, the preset degradation number corresponding to the first passive content pool B is M/2, and the preset degradation number corresponding to the first passive content pool C is M. It can be seen that the value of the preset degradation amount doubles each time the level of the first passive content pool decreases by one step, i.e. gradually increases as the level of the first passive content pool decreases. It should be noted that, the magnitude of the increase of the preset degradation amount is not fixed to be doubled, and a person skilled in the art may select the magnitude of the increase of the preset degradation amount according to the actual situation, which is not limited herein.
It should be noted that, the preset upgrading number and the preset downgrading number of the same first passive content pool are equal, and the number of recent original content in the first passive content pool is stable, so that the pyramid content pool cannot be consumed to be empty by the exposure operation under the condition that the number of recent original content issued by a user is small and the number of exposure is large, and a stable recommendation function can be maintained.
FIG. 3 illustrates a flow chart of calculating a favorites score provided by an embodiment of the present application.
Determining a preference score of recent original content through step S201, step S202, step S203, step 204:
step S201, determining a static quality score according to static data of recent original content.
For each recent original content authored by a user, a criterion needs to be set to evaluate the popularity of the recent original content, and the more recent original content loved by the user is obtained, the more the recent original content should be upgraded into a higher-level content pool to obtain more exposure opportunities. Based on the method, an overall preference score calculation method is finally obtained from three angles of an creator, an evaluator and time by adopting a three-in-one evaluation method of static quality score, dynamic quality score and time attenuation coefficient.
The static quality score mainly considers static data carried by the recent original content just released, and is a numerical value evaluated from the perspective of heart blood paid by a user for creating the recent original content. The more the recent original content is paid out, the more likely the recent original content is to have higher quality. Specifically, the following 10 types of static features are considered, as shown in table 1 below. These static features all represent the cost of authoring recently authored content. Considering that importance degrees among the static features are different, the application learns the relation between the static features and future dynamic quality of the recent original content in a large amount of data of the recent original content by using a machine learning method, fits the static features together by using a deep neural network, and finally obtains a static quality score corresponding to the recent original content.
Table 1: static characteristic data table.
| Sequence number | Features (e.g. a character) | Description of the features |
| 1 | Time of manufacture | Seconds for authoring recently authored content |
| 2 | Component occupancy value | Sum of occupancy values of all components in recent original content |
| 3 | Capacity size | Disk space size occupied by recent original content file |
| 4 | Description Length | Length of recent original content description text |
| 5 | Total number of components | Sum of number of all components in recent original content |
| 6 | Component category number | Sum of types and numbers of all components in recent original content |
| 7 | Whether or not it is the initial name | Whether the creator itself has a name for the recent original content |
| 8 | Whether or not to contain music | Whether music is played in recent original content |
| 9 | Type(s) | What type of recent original content belongs to |
| 10 | Number of various types of components | Number of different components in recent original content |
As shown in Table 1, the static characteristics include production time, component occupancy value, capacity size, description length, total number of components, number of component categories, initial name, music contained, type and number of types of components. The type refers to the type of the recently created content, and taking the recently created content as a game level as an example, the type can be the types of racing, decryption and the like. Components refer to components used to make recent original content.
And evaluating ten static characteristics of the recent original content by using the deep neural network model, and determining a static quality score of the recent original content.
Step S202, determining a dynamic quality score according to feedback information of the user on the recently original content.
The dynamic quality score is used for representing whether the recent original content is fed back by the user after the user is exposed to the recent original content for a certain number of times. User feedback information includes, but is not limited to: play, praise, collect and share. Wherein, the playing can refer to whether the user plays after seeing the recent original content; the praise may refer to whether the user praise after playing the recent original content; the collection can refer to whether a user collects the recent original content in his own favorites or not after playing the recent original content; sharing may refer to whether a user plays recent original content before sharing to friends or other users.
Here, a dynamic quality score of the recent original content is determined based on the number of plays, praise, collection and share of the recent original content by different users. Because the exposure times possessed by different content pools are different, the dynamic characteristics are required to be normalized, the ratio of the total number of plays to the total number of exposures is taken as the play rate, the ratio of the total number of praise to the total number of plays is taken as the praise rate, the ratio of the total number of collection to the total number of plays is taken as the collection rate, and the ratio of the total number of shares to the total number of plays is taken as the share rate. The play rate, the praise rate, the collection rate and the share rate are four dynamic quality evaluation indexes for measuring the dynamic quality.
The four dynamic quality evaluation indexes are different in importance degree, and the importance degree is the sharing rate, the praise rate, the collection rate and the play rate from high to low in sequence, so that corresponding weights are set for each index according to the importance degree, and the sum of the weights of the four dynamic quality evaluation indexes is used as a dynamic quality score.
Illustratively, dynamic quality score = 1 x play rate +18 x praise rate +10 x collection rate +20 x share rate.
Step S203, determining a time attenuation coefficient according to the issued duration of the recent original content.
Time decay refers to the fact that after a recent original content is released, its popularity score gradually decreases with time until it is zero. In order to prevent the problem of aesthetic fatigue of users caused by a long-term overlord advanced content pool of classical recent original content. The time decay coefficient can be calculated using the following formula: time decay coefficient=max (1—number of days of recent original content release× 0.02,0), where 0.02 is a coefficient in the formula, and a person skilled in the art can select a specific value of the coefficient according to the actual situation.
And step S204, determining the preference score of the recent original content according to the static quality score, the dynamic quality score and the time attenuation coefficient.
Here, the preference score of the recent original content may be determined according to the static quality score, the dynamic quality score and the time attenuation coefficient, for example, the product of the static quality score, the dynamic quality score and the time attenuation coefficient may be used as the preference score of the recent original content, the product of the dynamic quality score and the time attenuation coefficient may be used as the preference score of the recent original content, the sum of the static quality score and the dynamic quality score may be calculated first, and then the product of the sum of the static quality score and the time attenuation coefficient may be used as the preference score of the recent original content.
In an alternative embodiment, determining a popularity score for recent original content based on a static quality score, a dynamic quality score, and a time decay factor, includes: steps e1 to e3.
And e1, determining the type and the level of a content pool where the recently original content is currently located.
Specifically, different favorites score calculation formulas are adopted according to different using scene characteristics.
In the low-level content pool, because the preset recommendation proportion corresponding to the recent original content is low, the exposure times are less, and the dynamic quality score has lower confidence, static quality scores are required to be introduced for balancing, so that the recent original content with more heart blood spent by an creator has higher probability to enter the high-level content pool.
In the advanced content pool, the number of exposure times is high due to the fact that the preset recommended proportion corresponding to the recent original content is high, so that sufficient user feedback data can be obtained, dynamic quality scoring is more reliable, and static quality scoring is not needed to be considered.
In summary, the first content pools are divided into the low-level content pool and the high-level content pool, and the first active content pool D and the first passive content pool C may be divided into the low-level content pool and the first passive content pool B, the first passive content pool a and the first passive content pool S may be divided into the high-level content pool.
And e2, if the current content pool is a first active content pool or a low-level first passive content pool, selecting a first favorites calculation formula, substituting the static quality score, the dynamic quality score and the time attenuation coefficient of the near-term original content into the first favorites calculation formula, and determining the favorites score of the near-term original content.
And taking the product of the static quality score, the dynamic quality score and the time attenuation coefficient of the near-term original content as the preference score of the near-term original content aiming at the first active content pool D and the first passive content pool C.
And e3, if the current content pool is a high-level first passive content pool, selecting a second favorites calculation formula, substituting the dynamic quality score and the time attenuation coefficient of the near-term original content into the second favorites calculation formula, and determining the favorites score of the near-term original content.
And taking the product of the dynamic quality score and the time attenuation coefficient of the near-term original content as the preference score of the near-term original content aiming at the first passive content pool B, the first passive content pool A and the first passive content pool S.
In an alternative embodiment, the content pool number threshold for a higher level first passive content pool of the plurality of first passive content pools is lower.
In one example, the first active content pool D has a content pool number threshold of 1600, the first passive content pool C has a content pool number threshold of 800, the first passive content pool B has a content pool number threshold of 400, the first passive content pool a has a content pool number threshold of 200, and the first passive content pool S has a content pool number threshold of 100. As the level of the content pools increases, the content pool number threshold gradually decreases, and the plurality of first content pools present a pyramid shape as the level increases.
In an alternative embodiment, the method further comprises: selecting historical hot content in different user preference intervals from each second content pool according to a second preset recommendation proportion to recommend the historical hot content to a user; determining inferior historical hot content according to feedback information of a user aiming at target historical hot content, and removing the inferior historical hot content from a plurality of second content pools; and acquiring new historical hot content to be supplemented into at least one target second content pool in the plurality of second content pools, and recommending the new historical hot content to the user according to recommendation rules.
Statistically, the proportion of users playing original content through the recommendation system is 40%, which means that 60% of the map is transmitted through other systems in the game, including but not limited to search systems, sharing systems, room systems and task systems.
In order to further utilize feedback data of all global systems, the application also utilizes a ranking list system to carry out global feedback collection, and feedback information of each game can be recorded to form a uniform favorites ranking list. Unlike the pyramid content pool recommendation method, the leaderboard feedback data is not limited to the recommendation system part, but extends to all feedback data.
In addition to adopting the recent original content progressive map pool, the application additionally creates a progressive map pool of historical hot content, namely a second content pool, wherein the main differences of the second content pool and the first content pool are as follows: and taking the full-service favorites ranking list as a source of scanning data, namely placing the historical hot content in the full-service favorites ranking list into a second resource library, placing the historical hot content in the second resource library into a second active content pool in a second content pool, and updating the historical hot content in the second content pool according to the same progressive method as the first content pool so as to remove the inferior historical hot content from the second content pool.
It should be noted that, when the historical hot content is obtained from the second resource library, the selection process adds randomness. Firstly, selecting a preset number of history hot contents which are ranked at the front from a second resource library as candidate history hot content sets, then randomly scrambling the history hot contents in the candidate history hot content sets, and sequentially selecting a required number of history hot contents from beginning to end of the randomly scrambled candidate history hot content sets to be added into a plurality of second content pools. And resetting the candidate historical hot content set at a set time point every day, returning to the head position again, and adding the newly selected historical hot content into a plurality of second content pools, wherein the same historical hot content can be scanned into the second content pools only once in the same day.
Compared with the first content pool, the quality of the initial original content in the second content pool is higher, and the influence of low-quality recommendation caused by the initial exposure of the recent original content can be compensated. Meanwhile, the user can review classical original content conveniently, and original content truly conforming to the favor of all users can appear in the recommendation list of the users for a long time.
Since the update process of the historical popular content in the second content pools is the same as the update process of the recent original content in the first content pools, the description thereof will not be repeated. It should be noted that, the plurality of first content pools and the plurality of second content pools may exist simultaneously, when a certain user opens the recommendation page, a first preset number of recent original contents are selected from the plurality of first content pools, a second preset number of historical popular contents are selected from the plurality of second content pools, and then the first preset number of recent original contents and the second preset number of historical popular contents are mixed together to be recommended to the user.
Compared with the original content recommending method in the game in the prior art, the method can directly acquire new recent original content and recommend the new recent original content to the user, so that the newly registered user can acquire the new recent original content, the new recently released content can be displayed, cold start is avoided, meanwhile, the recent original content in different user preference ranges can be selected from a plurality of content pools to be recommended to the user, the recommending breadth is increased, and the problems of cold start, excessive recommendation and narrow recommendation in the existing recommending method are solved.
Based on the same inventive concept, the embodiment of the application also provides an original content recommendation method device in the game corresponding to the original content recommendation method in the game, and because the principle of solving the problem by the device in the embodiment of the application is similar to that of the original content recommendation method in the game in the embodiment of the application, the implementation of the device can refer to the implementation of the method, and the repetition is omitted.
Referring to fig. 4, fig. 4 is a schematic structural diagram of an apparatus for recommending original content in a game according to an embodiment of the present application. As shown in fig. 4, the method and apparatus 300 for recommending original contents in a game include:
the first recommendation module 301 is configured to select, according to a first preset recommendation ratio, recent original content from each of the first content pools of different levels, and recommend the selected recent original content to a user;
a content updating module 302, configured to determine recent inferior content according to feedback information of a user for a target recent original content, and remove the recent inferior content from a plurality of first content pools;
the second recommending module 303 is configured to obtain new recent original content to be added to at least one target first content pool of the plurality of first content pools, and recommend the new recent original content to a user according to a recommendation rule, where the new recent original content is recent original content that has not been placed in the plurality of first content pools.
In one possible embodiment, the method further comprises: selecting historical hot content from each second content pool with different grades according to a second preset recommendation proportion to recommend the historical hot content to a user; determining inferior historical hot content according to feedback information of a user aiming at target historical hot content, and removing the inferior historical hot content from a plurality of second content pools; and acquiring new historical hot content to be supplemented into at least one target second content pool in the plurality of second content pools, and recommending the new historical hot content to the user according to recommendation rules.
In one possible embodiment, the at least one first target content pool comprises a first active content pool, obtaining new recent original content supplements to the at least one target first content pool of the plurality of first content pools, comprising: when the number of the recently created contents in the first active content pool is smaller than the content pool number threshold corresponding to the first active content pool, new recently created contents are acquired and added into the first active content pool, so that the number of the recently created contents in the first active content pool is equal to the content pool number threshold corresponding to the first active content pool.
In one possible embodiment, after acquiring new recent original content to be added to the first active content pool, the method comprises: setting basic flow for new recent original content, and setting the new recent original content as an exposure state; when the new recent original content is browsed by the user, subtracting one from the basic flow of the new recent original content in the exposure state; when the basic flow of the new recent original content is reduced to 0, changing the state of the new recent original content from the exposure state to the waiting state, wherein the recent original content in the waiting state is not recommended to the user; and calculating the accumulated time length of the new recent original content entering the waiting state, and changing the waiting state of the new recent original content into the ranking state when the accumulated time length reaches the threshold value of the waiting time length.
In one possible embodiment, the plurality of first content pools includes a plurality of first passive content pools, the near-term bad content is determined based on feedback information of a user for the target near-term original content, and the near-term bad content is removed from the plurality of first content pools, including: determining the number of recent original contents in a ranking state in a first active content pool; determining a preference score of the recent original content in the ranking state when the number of the recent original content in the ranking state reaches a first score number threshold; the method comprises the steps of taking a first set number of recently created contents with top favorites scores in a first active content pool as recently high-quality contents, and transferring the recently high-quality contents to a plurality of first passive content pools; the near-term poor content is removed from the first plurality of content pools using near-term good content.
In one possible embodiment, the plurality of first passive content pools are provided with corresponding levels from high to low according to respective stored recommendations of recently authored content, and the near-term poor quality content is removed from the plurality of first content pools by using recently good quality content, comprising: adding a first set number of recent premium content to a first passive content pool of a lowest level; progressively upgrading the recent premium content in each first passive content pool to a first passive content pool of a first level higher starting with the first passive content pool of the lowest level joining the first set number of recent premium content until the recent premium content in the first passive content pool of the next highest level is upgraded to the first passive content pool of the highest level; progressively downgrading recent bad content in each first passive content pool to a first passive content pool of a lower level, starting with the first passive content pool of the highest level joining recent good content, until a first set number of recent bad content is removed from the first passive content pool of the lowest level.
In one possible embodiment, starting with a first passive content pool joining a lowest level of a first set amount of recent premium content, progressively upgrading recent premium content in each first passive content pool to a first passive content pool of a higher level, comprising: selecting a first passive content pool with the lowest level as a first upgrading passive content pool; determining whether the first upgraded passive content pool is a highest level first passive content pool; if the first upgrading passive content pool is not the first passive content pool with the highest level, selecting near-term high-quality content from near-term original content in the first upgrading passive content pool, and transferring near-term high-quality content to the first passive content pool with the level higher than that of the first upgrading passive content pool; and taking the first passive content pool with the higher level as a new first upgrading passive content pool, and returning to execute the step of selecting the recently high-quality content from the recently original content in the first upgrading passive content pool.
In one possible embodiment, progressively downgrading recent bad content in each first passive content pool to a first passive content pool of a lower level starting with the first passive content pool joining the highest level of recent good content, comprising: selecting a first passive content pool with the highest level as a first degraded passive content pool; determining whether the first degraded passive content pool is a lowest level first passive content pool; if the first degraded passive content pool is not the first passive content pool with the lowest level, selecting recent inferior content from recent original content in the first degraded passive content pool, and transferring the recent inferior content to the first passive content pool with the lower level of the first degraded passive content pool; and taking the first passive content pool with the lower level as a new first degraded passive content pool, and returning to execute the step of selecting the recent inferior content from the recent original content in the first degraded passive content pool.
In one possible embodiment, selecting near-term premium content from near-term original content in the first upgraded passive content pool comprises: when the evaluation condition is met, sequencing according to the time sequence of the recent original content entering the first upgrading passive content pool, and selecting the recent original content with a fixed proportion and top ranking as the recent candidate high-quality content; determining a current preference score for recent candidate premium content; and sorting the recent candidate high-quality contents according to the order of the current preference scores from high to low, and selecting the recent candidate high-quality contents with preset upgrade numbers corresponding to the first upgrade passive content pool in the first set number ranked at the front as final recent high-quality contents.
In one possible embodiment, selecting recently bad content from recently original content in the first degraded passive content pool includes: when the evaluation condition is met, sequencing according to the time sequence of the recent original content entering the first degradation passive content pool, and selecting the recent original content with a fixed proportion and the top ranking as the recent candidate inferior content; determining a current preference score of the recent candidate inferior content; and sorting the recent candidate inferior contents according to the order of the current preference scores from high to low, and selecting the recent candidate inferior contents with preset degradation numbers corresponding to the first degradation passive content pool in the first set number after ranking as the final recent inferior contents.
In one possible embodiment, the numerical size of the preset upgrade number corresponding to the plurality of first passive content pools gradually decreases as the level of the first passive content pools increases.
In one possible embodiment, the magnitude of the number of predefined downgrades corresponding to the plurality of first passive content pools increases gradually with decreasing level of the first passive content pools.
In one possible embodiment, the popularity score of recent original content is determined by: determining a static quality score according to static data of recent original content; determining a dynamic quality score according to feedback information of a user on recent original content; determining a time attenuation coefficient according to the released duration of the recent original content; and determining the preference score of the recent original content according to the static quality score, the dynamic quality score and the time attenuation coefficient.
In one possible embodiment, determining a popularity score for recent original content based on a static quality score, a dynamic quality score, and a time decay factor, comprises: determining the type and the level of a content pool where the recently original content is currently located; if the current content pool is a first active content pool or a low-level first passive content pool, selecting a first favorites calculation formula, substituting a static quality score, a dynamic quality score and a time attenuation coefficient of the near-term original content into the first favorites calculation formula, and determining the favorites score of the near-term original content; and if the current content pool is a high-level first passive content pool, selecting a second favorites calculation formula, substituting the dynamic quality score and the time attenuation coefficient of the near-term original content into the second favorites calculation formula, and determining the favorites score of the near-term original content.
In one possible embodiment, the content pool number threshold for a higher level first passive content pool of the plurality of first passive content pools is lower.
Referring to fig. 5, fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the application. As shown in fig. 5, the electronic device 400 includes a processor 410, a memory 420, and a bus 430.
The memory 420 stores machine readable instructions executable by the processor 410, when the electronic device is running an original content recommendation method in a game as in the embodiments, the processor 410 communicates with the memory 420 via the bus 430, the processor 410 executes the machine readable instructions, the preamble of the processor 410 method item to perform the steps of:
selecting recent original content from each first content pool of different levels according to a first preset recommendation proportion to recommend the recent original content to a user;
determining recent inferior content according to feedback information of a user aiming at target recent original content, and removing the recent inferior content from a plurality of first content pools;
the method comprises the steps of obtaining new recent original content to be supplemented to at least one target first content pool in a plurality of first content pools, recommending the new recent original content to a user according to recommendation rules, wherein the new recent original content is the recent original content which is not put into the plurality of first content pools.
In one possible embodiment, the processor 410 is further configured to: selecting historical hot content from each second content pool with different grades according to a second preset recommendation proportion to recommend the historical hot content to a user; determining inferior historical hot content according to feedback information of a user aiming at target historical hot content, and removing the inferior historical hot content from a plurality of second content pools; and acquiring new historical hot content to be supplemented into at least one target second content pool in the plurality of second content pools, and recommending the new historical hot content to the user according to recommendation rules.
In one possible embodiment, the at least one first target content pool comprises a first active content pool, and the processor 410, when executing the obtaining of new recent original content supplements to the at least one target first content pool of the plurality of first content pools, is specifically configured to: when the number of the recently created contents in the first active content pool is smaller than the content pool number threshold corresponding to the first active content pool, new recently created contents are acquired and added into the first active content pool, so that the number of the recently created contents in the first active content pool is equal to the content pool number threshold corresponding to the first active content pool.
In one possible embodiment, the processor 410 is specifically configured to, after executing the acquisition of new recent original content to the first active content pool: setting basic flow for new recent original content, and setting the new recent original content as an exposure state; when the new recent original content is browsed by the user, subtracting one from the basic flow of the new recent original content in the exposure state; when the basic flow of the new recent original content is reduced to 0, changing the state of the new recent original content from the exposure state to the waiting state, wherein the recent original content in the waiting state is not recommended to the user; and calculating the accumulated time length of the new recent original content entering the waiting state, and changing the waiting state of the new recent original content into the ranking state when the accumulated time length reaches the threshold value of the waiting time length.
In one possible embodiment, the plurality of first content pools includes a plurality of first passive content pools, and the processor 410 is specifically configured to, when executing the determining of recent bad content according to feedback information of the user for the target recent original content, remove the recent bad content from the plurality of first content pools: determining the number of recent original contents in a ranking state in a first active content pool; determining a preference score of the recent original content in the ranking state when the number of the recent original content in the ranking state reaches a first score number threshold; the method comprises the steps of taking a first set number of recently created contents with top favorites scores in a first active content pool as recently high-quality contents, and transferring the recently high-quality contents to a plurality of first passive content pools; the near-term poor content is removed from the first plurality of content pools using near-term good content.
In a possible embodiment, the plurality of first passive content pools are provided with corresponding levels from high to low according to the recommendation level of the recently created content stored in each of the first passive content pools, and the processor 410 is specifically configured to, when executing the removal of the recently bad content from the plurality of first content pools by using the recently good content: adding a first set number of recent premium content to a first passive content pool of a lowest level; progressively upgrading the recent premium content in each first passive content pool to a first passive content pool of a first level higher starting with the first passive content pool of the lowest level joining the first set number of recent premium content until the recent premium content in the first passive content pool of the next highest level is upgraded to the first passive content pool of the highest level; progressively downgrading recent bad content in each first passive content pool to a first passive content pool of a lower level, starting with the first passive content pool of the highest level joining recent good content, until a first set number of recent bad content is removed from the first passive content pool of the lowest level.
In one possible embodiment, the processor 410 is further configured to, when executing the first passive content pools that are progressively updated with the recent premium content in each first passive content pool to a first passive content pool that is one level higher, starting with the first passive content pool that joins the lowest level of the first set amount of recent premium content: selecting a first passive content pool with the lowest level as a first upgrading passive content pool; determining whether the first upgraded passive content pool is a highest level first passive content pool; if the first upgrading passive content pool is not the first passive content pool with the highest level, selecting near-term high-quality content from near-term original content in the first upgrading passive content pool, and transferring near-term high-quality content to the first passive content pool with the level higher than that of the first upgrading passive content pool; and taking the first passive content pool with the higher level as a new first upgrading passive content pool, and returning to execute the step of selecting the recently high-quality content from the recently original content in the first upgrading passive content pool.
In one possible embodiment, the processor 410 is specifically configured to, when executing a first passive content pool that progressively downgrades recent bad content in each first passive content pool to a first passive content pool one level lower, starting with the first passive content pool that joins the highest level of recent good content: selecting a first passive content pool with the highest level as a first degraded passive content pool; determining whether the first degraded passive content pool is a lowest level first passive content pool; if the first degraded passive content pool is not the first passive content pool with the lowest level, selecting recent inferior content from recent original content in the first degraded passive content pool, and transferring the recent inferior content to the first passive content pool with the lower level of the first degraded passive content pool; and taking the first passive content pool with the lower level as a new first degraded passive content pool, and returning to execute the step of selecting the recent inferior content from the recent original content in the first degraded passive content pool.
In one possible embodiment, the processor 410, when executing selecting near-future premium content from near-future originals in the first upgraded passive content pool, is specifically configured to: when the evaluation condition is met, sequencing according to the time sequence of the recent original content entering the first upgrading passive content pool, and selecting the recent original content with a fixed proportion and top ranking as the recent candidate high-quality content; determining a current preference score for recent candidate premium content; and sorting the recent candidate high-quality contents according to the order of the current preference scores from high to low, and selecting the recent candidate high-quality contents with preset upgrade numbers corresponding to the first upgrade passive content pool in the first set number ranked at the front as final recent high-quality contents.
In one possible embodiment, the processor 410, when executing the selection of recent bad content from the recent original content in the first degraded passive content pool, is specifically configured to: when the evaluation condition is met, sequencing according to the time sequence of the recent original content entering the first degradation passive content pool, and selecting the recent original content with a fixed proportion and the top ranking as the recent candidate inferior content; determining a current preference score of the recent candidate inferior content; and sorting the recent candidate inferior contents according to the order of the current preference scores from high to low, and selecting the recent candidate inferior contents with preset degradation numbers corresponding to the first degradation passive content pool in the first set number after ranking as the final recent inferior contents.
In one possible embodiment, the numerical size of the preset upgrade number corresponding to the plurality of first passive content pools gradually decreases as the level of the first passive content pools increases.
In one possible embodiment, the magnitude of the number of predefined downgrades corresponding to the plurality of first passive content pools increases gradually with decreasing level of the first passive content pools.
In one possible embodiment, the processor 410 determines a popularity score for recently authored content by: determining a static quality score according to static data of recent original content; determining a dynamic quality score according to feedback information of a user on recent original content; determining a time attenuation coefficient according to the released duration of the recent original content; and determining the preference score of the recent original content according to the static quality score, the dynamic quality score and the time attenuation coefficient.
In one possible embodiment, the processor 410 is specifically configured to, when executing determining the preference score for recent original content based on the static quality score, the dynamic quality score, and the time decay factor: determining the type and the level of a content pool where the recently original content is currently located; if the current content pool is a first active content pool or a low-level first passive content pool, selecting a first favorites calculation formula, substituting a static quality score, a dynamic quality score and a time attenuation coefficient of the near-term original content into the first favorites calculation formula, and determining the favorites score of the near-term original content; and if the current content pool is a high-level first passive content pool, selecting a second favorites calculation formula, substituting the dynamic quality score and the time attenuation coefficient of the near-term original content into the second favorites calculation formula, and determining the favorites score of the near-term original content.
In one possible embodiment, the content pool number threshold for a higher level first passive content pool of the plurality of first passive content pools is lower.
By the method, the new recent original content is directly acquired, and is recommended to the user, so that the newly registered user can acquire the new recent original content, the new recently released content can be displayed, cold start is avoided, meanwhile, the recent original content in different user preference intervals can be selected from a plurality of content pools to be recommended to the user, the recommendation breadth is increased, and the problems of cold start, excessive recommendation and narrow recommendation in the existing recommendation method are solved.
The embodiment of the application also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor when the computer program is executed by the processor, and the processor executes the following steps:
selecting recent original content from each first content pool of different levels according to a first preset recommendation proportion to recommend the recent original content to a user;
determining recent inferior content according to feedback information of a user aiming at target recent original content, and removing the recent inferior content from a plurality of first content pools;
the method comprises the steps of obtaining new recent original content to be supplemented to at least one target first content pool in a plurality of first content pools, recommending the new recent original content to a user according to recommendation rules, wherein the new recent original content is the recent original content which is not put into the plurality of first content pools.
In one possible embodiment, the processor is further configured to: selecting historical hot content from each second content pool with different grades according to a second preset recommendation proportion to recommend the historical hot content to a user; determining inferior historical hot content according to feedback information of a user aiming at target historical hot content, and removing the inferior historical hot content from a plurality of second content pools; and acquiring new historical hot content to be supplemented into at least one target second content pool in the plurality of second content pools, and recommending the new historical hot content to the user according to recommendation rules.
In one possible embodiment, the at least one first target content pool comprises a first active content pool, and the processor, when executing the obtaining of new recent original content supplements to the at least one target first content pool of the plurality of first content pools, is specifically configured to: when the number of the recently created contents in the first active content pool is smaller than the content pool number threshold corresponding to the first active content pool, new recently created contents are acquired and added into the first active content pool, so that the number of the recently created contents in the first active content pool is equal to the content pool number threshold corresponding to the first active content pool.
In one possible embodiment, the processor is specifically configured to, after executing the acquisition of new recent original content to join the first active content pool: setting basic flow for new recent original content, and setting the new recent original content as an exposure state; when the new recent original content is browsed by the user, subtracting one from the basic flow of the new recent original content in the exposure state; when the basic flow of the new recent original content is reduced to 0, changing the state of the new recent original content from the exposure state to the waiting state, wherein the recent original content in the waiting state is not recommended to the user; and calculating the accumulated time length of the new recent original content entering the waiting state, and changing the waiting state of the new recent original content into the ranking state when the accumulated time length reaches the threshold value of the waiting time length.
In one possible embodiment, the plurality of first content pools includes a plurality of first passive content pools, and the processor, when executing determining recent bad content from feedback information of the user for the target recent original content, is specifically configured to: determining the number of recent original contents in a ranking state in a first active content pool; determining a preference score of the recent original content in the ranking state when the number of the recent original content in the ranking state reaches a first score number threshold; the method comprises the steps of taking a first set number of recently created contents with top favorites scores in a first active content pool as recently high-quality contents, and transferring the recently high-quality contents to a plurality of first passive content pools; the near-term poor content is removed from the first plurality of content pools using near-term good content.
In a possible embodiment, the plurality of first passive content pools are provided with corresponding levels from high to low according to the recommendation level of the recently original content stored in each of the plurality of first passive content pools, and the processor is specifically configured to, when executing the removal of the recently inferior content from the plurality of first content pools using the recently superior content: adding a first set number of recent premium content to a first passive content pool of a lowest level; progressively upgrading the recent premium content in each first passive content pool to a first passive content pool of a first level higher starting with the first passive content pool of the lowest level joining the first set number of recent premium content until the recent premium content in the first passive content pool of the next highest level is upgraded to the first passive content pool of the highest level; progressively downgrading recent bad content in each first passive content pool to a first passive content pool of a lower level, starting with the first passive content pool of the highest level joining recent good content, until a first set number of recent bad content is removed from the first passive content pool of the lowest level.
In one possible embodiment, the processor, when executing the first passive content pools starting from the lowest level joining the first set amount of recent premium content, progressively upgrades the recent premium content in each first passive content pool to a first passive content pool one level higher, is specifically to: selecting a first passive content pool with the lowest level as a first upgrading passive content pool; determining whether the first upgraded passive content pool is a highest level first passive content pool; if the first upgrading passive content pool is not the first passive content pool with the highest level, selecting near-term high-quality content from near-term original content in the first upgrading passive content pool, and transferring near-term high-quality content to the first passive content pool with the level higher than that of the first upgrading passive content pool; and taking the first passive content pool with the higher level as a new first upgrading passive content pool, and returning to execute the step of selecting the recently high-quality content from the recently original content in the first upgrading passive content pool.
In one possible embodiment, the processor, when executing the first passive content pools that progressively downgrade recent bad content in each first passive content pool to a first passive content pool one level lower, is specifically configured to: selecting a first passive content pool with the highest level as a first degraded passive content pool; determining whether the first degraded passive content pool is a lowest level first passive content pool; if the first degraded passive content pool is not the first passive content pool with the lowest level, selecting recent inferior content from recent original content in the first degraded passive content pool, and transferring the recent inferior content to the first passive content pool with the lower level of the first degraded passive content pool; and taking the first passive content pool with the lower level as a new first degraded passive content pool, and returning to execute the step of selecting the recent inferior content from the recent original content in the first degraded passive content pool.
In a possible embodiment, the processor, when executing selecting recent premium content from the recent original content in the first upgraded passive content pool, is specifically configured to: when the evaluation condition is met, sequencing according to the time sequence of the recent original content entering the first upgrading passive content pool, and selecting the recent original content with a fixed proportion and top ranking as the recent candidate high-quality content; determining a current preference score for recent candidate premium content; and sorting the recent candidate high-quality contents according to the order of the current preference scores from high to low, and selecting the recent candidate high-quality contents with preset upgrade numbers corresponding to the first upgrade passive content pool in the first set number ranked at the front as final recent high-quality contents.
In one possible embodiment, the processor, when executing selecting recent bad content from the recent original content in the first degraded passive content pool, is specifically configured to: when the evaluation condition is met, sequencing according to the time sequence of the recent original content entering the first degradation passive content pool, and selecting the recent original content with a fixed proportion and the top ranking as the recent candidate inferior content; determining a current preference score of the recent candidate inferior content; and sorting the recent candidate inferior contents according to the order of the current preference scores from high to low, and selecting the recent candidate inferior contents with preset degradation numbers corresponding to the first degradation passive content pool in the first set number after ranking as the final recent inferior contents.
In one possible embodiment, the numerical size of the preset upgrade number corresponding to the plurality of first passive content pools gradually decreases as the level of the first passive content pools increases.
In one possible embodiment, the magnitude of the number of predefined downgrades corresponding to the plurality of first passive content pools increases gradually with decreasing level of the first passive content pools.
In one possible embodiment, the processor determines a popularity score for recently authored content by: determining a static quality score according to static data of recent original content; determining a dynamic quality score according to feedback information of a user on recent original content; determining a time attenuation coefficient according to the released duration of the recent original content; and determining the preference score of the recent original content according to the static quality score, the dynamic quality score and the time attenuation coefficient.
In one possible embodiment, the processor, when executing determining the preference score of the recent original content based on the static quality score, the dynamic quality score, and the time decay factor, is specifically configured to: determining the type and the level of a content pool where the recently original content is currently located; if the current content pool is a first active content pool or a low-level first passive content pool, selecting a first favorites calculation formula, substituting a static quality score, a dynamic quality score and a time attenuation coefficient of the near-term original content into the first favorites calculation formula, and determining the favorites score of the near-term original content; and if the current content pool is a high-level first passive content pool, selecting a second favorites calculation formula, substituting the dynamic quality score and the time attenuation coefficient of the near-term original content into the second favorites calculation formula, and determining the favorites score of the near-term original content.
In one possible embodiment, the content pool number threshold for a higher level first passive content pool of the plurality of first passive content pools is lower.
By the method, the new recent original content is directly acquired, and is recommended to the user, so that the newly registered user can acquire the new recent original content, the new recently released content can be displayed, cold start is avoided, meanwhile, the recent original content in different user preference intervals can be selected from a plurality of content pools to be recommended to the user, the recommendation breadth is increased, and the problems of cold start, excessive recommendation and narrow recommendation in the existing recommendation method are solved.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.
In the several embodiments provided by the present application, it should be understood that the disclosed systems, devices, and methods may be implemented in other manners. The above-described apparatus embodiments are merely illustrative, for example, the division of the units is merely a logical function division, and there may be other manners of division in actual implementation, and for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some communication interface, device or unit indirect coupling or communication connection, which may be in electrical, mechanical or other form.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer readable storage medium executable by a processor. Based on this understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
Finally, it should be noted that: the above examples are only specific embodiments of the present application, and are not intended to limit the scope of the present application, but it should be understood by those skilled in the art that the present application is not limited thereto, and that the present application is described in detail with reference to the foregoing examples: any person skilled in the art may modify or easily conceive of the technical solution described in the foregoing embodiments, or perform equivalent substitution of some of the technical features, while remaining within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application. Therefore, the protection scope of the application is subject to the protection scope of the claims.
Claims (18)
1. A method for recommending original content in a game, comprising:
selecting recent original content from each first content pool of different levels according to a first preset recommendation proportion to recommend the recent original content to a user;
determining recent inferior content according to feedback information of a user aiming at target recent original content, and removing the recent inferior content from a plurality of first content pools;
And acquiring new recent original content to be supplemented to at least one target first content pool in the first content pools, and recommending the new recent original content to a user according to a recommendation rule, wherein the new recent original content is the recent original content which is not put into the first content pools.
2. The method according to claim 1, wherein the method further comprises:
selecting historical hot content from each second content pool with different grades according to a second preset recommendation proportion to recommend the historical hot content to a user;
determining inferior historical hot content according to feedback information of a user aiming at target historical hot content, and removing the inferior historical hot content from a plurality of second content pools;
and acquiring new historical hot content to be supplemented into at least one target second content pool in the plurality of second content pools, and recommending the new historical hot content to a user according to recommendation rules.
3. The method of claim 1, wherein the at least one first target content pool comprises a first active content pool, wherein the obtaining new recent original content supplements at least one target first content pool of the plurality of first content pools, comprises:
When the number of the recently created contents in the first active content pool is smaller than the content pool number threshold corresponding to the first active content pool, new recently created contents are acquired and added into the first active content pool, so that the number of the recently created contents in the first active content pool is equal to the content pool number threshold corresponding to the first active content pool.
4. A method according to claim 3, characterized in that after the acquisition of new recent original content is added to the first active content pool, it comprises:
setting basic flow for the new recent original content, and setting the new recent original content as an exposure state;
when the new recent original content is browsed by a user, subtracting one from the basic flow of the new recent original content in an exposure state;
when the basic flow of the new recent original content is reduced to 0, changing the state of the new recent original content from an exposure state to a waiting state, wherein the recent original content in the waiting state is not recommended to a user;
calculating the accumulated time length of the new recent original content entering the waiting state, and changing the waiting state into the ranking state when the accumulated time length reaches the waiting time length threshold value.
5. The method of claim 1, wherein the plurality of first content pools includes a plurality of first passive content pools, wherein the determining recent bad content from feedback information of a user for the target recent original content, removing the recent bad content from the plurality of first content pools, comprises:
determining the number of recent original contents in a ranking state in the first active content pool;
determining a preference score of the recent original content in the ranking state when the number of the recent original content in the ranking state reaches a first score number threshold;
taking a first set number of recently created content with top preference scores in the first active content pool as recently good content, and transferring the recently good content into the plurality of first passive content pools;
and removing the near-term inferior content from the plurality of first content pools by utilizing the near-term superior content.
6. The method of claim 5, wherein the plurality of first passive content pools have corresponding levels from high to low according to respective stored recommendations of recently authored content, wherein the removing recently bad content from the plurality of first content pools using the recently good content comprises:
Adding a first set number of recent premium content to a first passive content pool of a lowest level;
progressively upgrading the recent premium content in each first passive content pool to a first passive content pool of a first level higher starting with the first passive content pool of the lowest level joining the first set number of recent premium content until the recent premium content in the first passive content pool of the next highest level is upgraded to the first passive content pool of the highest level;
progressively downgrading recent bad content in each first passive content pool to a first passive content pool of a lower level, starting with the first passive content pool of the highest level joining recent good content, until a first set number of recent bad content is removed from the first passive content pool of the lowest level.
7. The method of claim 6, wherein progressively upgrading recent premium content in each first passive content pool to a first passive content pool one level higher starting with the first passive content pool joining the lowest level of the first set amount of recent premium content, comprising:
selecting a first passive content pool with the lowest level as a first upgrading passive content pool;
Determining whether the first upgraded passive content pool is a highest level first passive content pool;
if the first upgrading passive content pool is not the highest-level first passive content pool, selecting recent high-quality content from recent original content in the first upgrading passive content pool, and transferring the recent high-quality content to a first passive content pool which is one level higher than the first upgrading passive content pool;
and taking the first passive content pool with the higher level as a new first upgrading passive content pool, and returning to the step of selecting the recently high-quality content from the recently original content in the first upgrading passive content pool.
8. The method of claim 6, wherein progressively downgrading recent bad content in each first passive content pool to a first passive content pool one level lower, starting with the first passive content pool joining the highest level of recent good content, comprises:
selecting a first passive content pool with the highest level as a first degraded passive content pool;
determining whether the first degraded passive content pool is a lowest level first passive content pool;
if the first degraded passive content pool is not the first passive content pool with the lowest level, selecting recent inferior content from recent original content in the first degraded passive content pool, and transferring the recent inferior content to the first passive content pool with the lower level of the first degraded passive content pool;
And taking the first passive content pool with the lower level as a new first degraded passive content pool, and returning to the step of selecting the recent inferior content from the recent original content in the first degraded passive content pool.
9. The method of claim 7, wherein selecting near-future premium content from near-future original content in the first upgraded passive content pool comprises:
when the evaluation condition is met, sequencing according to the time sequence of the recent original content entering the first upgrading passive content pool, and selecting the recent original content with a fixed proportion and the top ranking as the recent candidate high-quality content;
determining a current preference score for the recent candidate premium content;
and sequencing the recent candidate high-quality contents according to the order of the current preference scores from high to low, and selecting the recent candidate high-quality contents with preset upgrading numbers corresponding to the first upgrading passive content pool in the first set number ranked at the front as final recent high-quality contents.
10. The method of claim 8, wherein the selecting recent bad content from the recent original content in the first degraded passive content pool comprises:
When the evaluation condition is met, sorting according to the time sequence of the recent original content entering the first degradation passive content pool, and selecting the recent original content with a fixed proportion and the top ranking as the recent candidate inferior content;
determining a current preference score of the recent candidate inferior content;
and sequencing the recent candidate inferior contents according to the sequence from high to low of the current preference scores, and selecting the recent candidate inferior contents with preset degradation numbers corresponding to the first degradation passive content pool in the first set number after ranking as final recent inferior contents.
11. The method of claim 9, wherein the value of the number of preset upgrades corresponding to the plurality of first passive content pools decreases progressively as the level of the first passive content pools increases.
12. The method of claim 10, wherein the magnitude of the number of predefined downgrades corresponding to the plurality of first passive content pools increases gradually as the level of the first passive content pools decreases.
13. The method of claim 1, wherein the popularity score of recently authored content is determined by:
Determining a static quality score according to the static data of the recent original content;
determining a dynamic quality score according to feedback information of a user on the recent original content;
determining a time attenuation coefficient according to the released duration of the recent original content;
and determining the preference score of the recent original content according to the static quality score, the dynamic quality score and the time attenuation coefficient.
14. The method of claim 13, wherein said determining a popularity score for said recently authored content based on said static quality score, said dynamic quality score, and said time decay factor comprises:
determining the type and the level of a content pool where the recent original content is currently located;
if the current content pool is a first active content pool or a low-level first passive content pool, selecting a first favorites calculation formula, substituting the static quality score, the dynamic quality score and the time attenuation coefficient of the recent original content into the first favorites calculation formula, and determining the favorites score of the recent original content;
and if the current content pool is a high-level first passive content pool, selecting a second favorites calculation formula, substituting the dynamic quality score and the time attenuation coefficient of the recent original content into the second favorites calculation formula, and determining the favorites score of the recent original content.
15. The method of claim 6, wherein a content pool number threshold for a higher level first passive content pool of the plurality of first passive content pools is lower.
16. A recent original content recommendation device for a user in a game, comprising:
the first recommendation module is used for selecting recent original content from each first content pool with different grades according to a first preset recommendation proportion and recommending the recent original content to a user;
the content updating module is used for determining recent inferior content according to feedback information of a user aiming at target recent original content and removing the recent inferior content from a plurality of first content pools;
the second recommending module is used for acquiring new recent original content to be supplemented to at least one target first content pool in the first content pools, recommending the new recent original content to a user according to a recommending rule, wherein the new recent original content is the recent original content which is not put into the first content pools.
17. An electronic device, comprising: a processor, a storage medium and a bus, the storage medium storing machine-readable instructions executable by the processor, the processor and the storage medium communicating over the bus when the electronic device is running, the processor executing the machine-readable instructions to perform the steps of the method of user recent original content recommendation in a game as claimed in any one of claims 1 to 15.
18. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, performs the steps of the user recent original content recommendation method in a game according to any of claims 1 to 15.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311284381.6A CN117216402A (en) | 2023-10-07 | 2023-10-07 | Original content recommendation method and device in game, electronic equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311284381.6A CN117216402A (en) | 2023-10-07 | 2023-10-07 | Original content recommendation method and device in game, electronic equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117216402A true CN117216402A (en) | 2023-12-12 |
Family
ID=89042394
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311284381.6A Pending CN117216402A (en) | 2023-10-07 | 2023-10-07 | Original content recommendation method and device in game, electronic equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117216402A (en) |
-
2023
- 2023-10-07 CN CN202311284381.6A patent/CN117216402A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6378850B1 (en) | Server and program | |
| US20140200959A1 (en) | Predicting future performance of games | |
| JP6145387B2 (en) | User matching method and system | |
| JP7443486B2 (en) | Content item recommendations | |
| US20220164407A1 (en) | Custom Compilation Videos | |
| CN107087235A (en) | Media content recommendations method, server and client | |
| US20210142387A1 (en) | Adaptive rewarding for content personalization | |
| CN104903889A (en) | Method for organising content | |
| US10449454B2 (en) | System and method for determining events of interest in a multi-player online game | |
| CN1980333A (en) | Method and device for generating media demo | |
| KR102060000B1 (en) | Information processing system, information processing method and program, server, and information processing terminal | |
| CN109947438B (en) | Game installation method based on machine learning, electronic equipment and readable storage medium | |
| US20160381158A1 (en) | Automatic Invitation Delivery System | |
| US20150302009A1 (en) | Adaptive Media Library for Application Ecosystems | |
| CN109922357A (en) | The method and device of video recommendations | |
| JP2021037060A (en) | System, method, and program for providing predetermined game, and method for creating classification of decks | |
| CN117216402A (en) | Original content recommendation method and device in game, electronic equipment and medium | |
| CN111659125A (en) | Game-based friend recommendation method and device and computer-readable storage medium | |
| CN109635969A (en) | Push the method and device of resource transfers offer | |
| CN108136261B (en) | Server and non-transitory computer-readable medium | |
| CN110851724B (en) | Article recommendation method based on self-media number grade and related products | |
| CN110574066B (en) | Server device and recording medium | |
| CN115103712A (en) | Personalized game notifications | |
| CN113626681A (en) | Account information processing method and device, electronic equipment and storage medium | |
| KR102544763B1 (en) | Method for recommending counselors tailored to the counselee and device for the same |
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 |