CN112381234A

CN112381234A - Resource distribution method, device, equipment and computer readable storage medium

Info

Publication number: CN112381234A
Application number: CN202011241207.XA
Authority: CN
Inventors: 刘爱宾
Original assignee: Beijing Dajia Internet Information Technology Co Ltd
Current assignee: Beijing Dajia Internet Information Technology Co Ltd
Priority date: 2020-11-09
Filing date: 2020-11-09
Publication date: 2021-02-19
Anticipated expiration: 2040-11-09
Also published as: CN112381234B

Abstract

The disclosure provides a resource distribution method, a resource distribution device and a computer-readable storage medium, relates to the technical field of computers, and is used for improving the matching degree of a distribution result of information resources and a business target. The method comprises the following steps: acquiring object characteristics of each target object in the target object group; estimating the object characteristics of each target object and the second relevance of each information resource according to the object characteristics of each history object in the history object group and the first relevance of each information resource in the resource allocation set corresponding to the history object group; determining the information resources mapped by each target object according to the estimated second relevance, and determining a resource allocation set corresponding to the target object group according to the information resources mapped by each target object; and distributing each information resource for each target object according to the mapping relation between each target object and each information resource in the resource allocation set corresponding to the target object group. The method improves the accuracy of the acquired resource diversity corresponding to the target object group.

Description

Resource distribution method, device, equipment and computer readable storage medium

Technical Field

The present disclosure relates to the field of computer technologies, and in particular, to a resource distribution method, apparatus, device, and computer-readable storage medium.

Background

In some scenarios, a plurality of different information resources need to be distributed to a plurality of objects, and at present, a plurality of information resources are often randomly distributed to each object, so that the information resources distributed by each object may have a phenomenon that the information resources are seriously inconsistent with the business target of the scenario, for example, the business target is the balance of the total amount of the information resources distributed to each object, or an object with less information resources currently held distributes more information resources, and the like, thereby affecting the willingness of each object to operate through the distributed information resources, and therefore, when a plurality of information resources are distributed to a plurality of objects, how to improve the matching degree of the distribution result of the information resources and the business target becomes a problem to be considered.

Disclosure of Invention

The embodiment of the disclosure provides a resource allocation method, a resource allocation device, a resource allocation apparatus and a computer-readable storage medium, which are used for improving the matching degree between the distribution result of information resources and a service target.

In a first aspect of the present disclosure, a resource distribution method is provided, including:

acquiring object characteristics of each target object in a target object group, wherein the object characteristics represent willingness values of the target objects to operate through distributed information resources;

estimating the object characteristics of each target object and the second relevance of each information resource according to the object characteristics of each history object in a history object group and the first relevance of each information resource in a resource allocation set corresponding to the history object group, wherein the resource allocation set corresponding to the history object group comprises the mapping relation between each history object in the history object group and each information resource, and

determining the information resources mapped by each target object according to the estimated second relevance, and determining the resource allocation set corresponding to the target object group according to the information resources mapped by each target object;

and distributing the information resources to the target objects according to the mapping relation between the target objects and the information resources in the resource allocation set corresponding to the target object group.

In a possible implementation manner, the step of estimating a second degree of association between the object feature of each target object and each information resource in the resource allocation set corresponding to the history object group according to the object feature of each history object in the history object group and the first degree of association between each information resource in the resource allocation set corresponding to the history object group, determining the information resource mapped to each target object according to the estimated second degree of association, and determining the resource allocation set corresponding to the target object group according to the information resource mapped to each target object includes:

and inputting the object characteristics of each target object by adopting a trained resource allocation model, and obtaining a resource allocation set corresponding to the target object group output by the resource allocation model, wherein the resource allocation model is obtained by training by adopting the object characteristics of each historical object in the historical object group and the resource allocation set corresponding to the historical object group as training samples based on a machine learning method.

In one possible implementation, the trained resource allocation model is obtained by:

extracting object characteristics of each historical object in the historical object group in each training sample and a resource allocation set corresponding to the historical object group;

and adjusting a resource allocation model of historical training or an initially created resource allocation model through machine learning according to the mapping relation between the object characteristics of the historical objects and the information resources in the resource allocation set corresponding to the historical object group to obtain the trained resource allocation model.

In a possible implementation manner, the resource allocation set includes K information resources, where K is an integer greater than 1; the step of inputting the object characteristics of each target object by using the trained resource allocation model and obtaining the resource allocation set corresponding to the target object group output by the resource allocation model includes:

inputting the object characteristics of each target object into the trained resource allocation model, performing M rounds of resource allocation operations, and determining the distribution sequence of the M information resources, wherein M is a positive integer not greater than K;

determining the information resources mapped by each target object according to the mapping relation between the distribution sequence of the M information resources and the distribution sequence corresponding to each target object;

and determining a resource allocation set corresponding to the target object group according to the information resources mapped by each target object.

In one possible implementation manner, the ith round of resource allocation operation, where i is a positive integer smaller than M, includes:

determining the object characteristics of each target object and target resources as current resource prediction characteristics, wherein the target resources comprise information resources of which the distribution sequence is determined currently;

estimating a fourth degree of association between the current resource characteristic and each information resource based on a third degree of association between the historical resource prediction characteristic and each information resource; and

determining the information resource of the current resource prediction feature mapping according to the estimated fourth relevance;

determining the distribution sequence of the information resources mapped by the current resource prediction characteristics as the i, and determining the information resources mapped by the current resource prediction characteristics as the target resources.

In a possible implementation manner, the object group includes a first number of objects, the resource allocation set includes a second number of information resources, and the second number is greater than a preset multiple of the first number.

In one possible implementation, the information resource includes a game resource in a game, the target object includes a game player, and the object feature includes at least one of:

a player rating of the game player;

a resource value of a game resource owned by the game player;

information of a failure of the game player in a historical game session.

In a second aspect of the present disclosure, a resource distribution apparatus is provided, including:

the characteristic acquisition unit is configured to acquire object characteristics of each target object in a target object group, wherein the object characteristics represent willingness values of the target objects to operate through distributed information resources;

the resource allocation unit is specifically used for estimating a second degree of association between the object feature of each target object and each information resource according to the object feature of each history object in the history object group and the first degree of association between each information resource in the resource allocation set corresponding to the history object group; the resource allocation set corresponding to the historical object group comprises the mapping relation between each historical object in the historical object group and each information resource, the information resource mapped by each target object is determined according to the estimated second relevance, and the resource allocation set corresponding to the target object group is determined according to the information resource mapped by each target object;

and the resource distribution unit is configured to execute the mapping relation between each target object and each information resource in the resource distribution set corresponding to the target object group and distribute each information resource for each target object.

In one possible implementation, the resource allocation unit is specifically configured to perform:

In a possible implementation manner, the resource allocation set includes K information resources, where K is an integer greater than 1; the resource allocation unit is specifically configured to perform:

In a possible implementation manner, the resource allocation unit is specifically configured to perform an ith round of resource allocation operation according to the following manner, where i is a positive integer smaller than M:

a player rating of the game player;

a resource value of a game resource owned by the game player;

information of a failure of the game player in a historical game session.

In a third aspect of the present disclosure, a computer device is provided, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, the method of any one of the first aspect and one possible implementation manner is implemented.

In a fourth aspect of the present disclosure, a computer-readable storage medium is provided, which stores computer instructions that, when executed on a computer, cause the computer to perform the method according to any one of the first aspect and one of the possible embodiments.

The scheme of the present disclosure brings at least the following beneficial effects:

in the embodiment of the disclosure, the object characteristics of each object in the object group and the second relevance of each information resource are estimated through the object characteristics of each history object in the history object group and the first relevance of each information resource in the corresponding resource allocation set, further determining a resource allocation set corresponding to the target object group according to the second association degree, distributing information resources for each target object according to the corresponding resource allocation set, estimating the second association degree of each target object and each information resource according to the first association degree of each history object and each information object, since the first degree of association of each history object and each information object is highly matched with the business objective, the second degree of association between each target object and each information object thus obtained is also high in matching degree with the business objective, therefore, the matching degree of the resource allocation set corresponding to the obtained target object group and the service target is improved.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the disclosure and are not to be construed as limiting the disclosure.

Fig. 1 is a schematic flowchart of a resource distribution method according to an exemplary embodiment of the present disclosure;

FIG. 2 is a flowchart illustrating a training process of a resource allocation model according to an exemplary embodiment of the present disclosure;

FIG. 3 is a flowchart illustrating a training process of a resource allocation model according to an exemplary embodiment of the present disclosure;

FIG. 4 is a flowchart illustrating a training process of a resource allocation model according to an exemplary embodiment of the present disclosure;

FIG. 5 is a flowchart illustrating a training process of a resource allocation model according to an exemplary embodiment of the present disclosure;

FIG. 6 is a process diagram of a round of resource allocation operations provided by an exemplary embodiment of the present disclosure;

fig. 7 is a schematic flowchart of obtaining a resource allocation set corresponding to a target object group according to an exemplary embodiment of the present disclosure;

FIG. 8 is a process diagram of a round of resource allocation operations provided by an exemplary embodiment of the present disclosure;

fig. 9 is a schematic structural diagram of a resource allocation model provided in an exemplary embodiment of the present disclosure;

fig. 10 is a schematic diagram illustrating a principle of obtaining a resource allocation set corresponding to a target object group according to an exemplary embodiment of the present disclosure;

fig. 11 is a schematic diagram of a distribution process of an information resource according to an exemplary embodiment of the disclosure;

fig. 12 is a diagram illustrating a structure of a resource distribution apparatus according to an exemplary embodiment of the present disclosure;

fig. 13 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present disclosure.

Detailed Description

In order to make the technical solutions of the present disclosure better understood by those of ordinary skill in the art, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.

It should be noted that the terms "first," "second," and the like in the description and claims of the present disclosure and in the above-described drawings are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the disclosure described herein are capable of operation in sequences other than those illustrated or otherwise described herein.

In order to facilitate better understanding of the technical solutions of the present disclosure by those skilled in the art, the following technical terms related to the present disclosure are explained.

Object, history object and target object: the object in the embodiments of the present disclosure may be, but is not limited to, any one of a user, an account, an avatar, a game player, and the like. The history objects are objects of history records, and the history object group is a combination formed by a plurality of history objects; the target object is an object of the current information resource to be distributed, and the target object group is a combination formed by a plurality of target objects.

Information resources: the information resource in the embodiment of the present disclosure is an electronic resource, which may be, but is not limited to, a media stream resource, a game resource, electronic money, and the like, the media stream resource may be, but is not limited to, at least one resource including an audio stream, a video stream, and the like, and the game resource may be, but is not limited to, at least one resource including a game item, a game gold, a arena place in a game, a treasure place, and the like.

Object characteristics: some feature data related to the information resource in the feature data of the object, such as game resource when the information resource, and the object feature when the target object is a game player, may include at least one of the following features: the player level of the game player, the resource value of the game resource owned by the game player, and the failure information of the game player in the historical game block.

Service scenario and service objective: the service scene in the embodiment of the present disclosure may include, but is not limited to, a game scene, a scene for auditing media stream resources, and the like; the service target refers to a target on a service corresponding to a service scene, and a person skilled in the art can set the service target according to actual requirements, for example, when the service scene is a game scene, the service target can be set such that the total power value of game items (information resources) distributed by each game player of a plurality of game players is the same or similar, or the service target can be set such that the total amount of game gold coins distributed by each game player is the same.

Resource allocation set: the method comprises the mapping relation between each object and each information resource in an object group, wherein the object group comprises at least one object.

The following explains the design concept of the present disclosure.

In some scenes, a plurality of different information resources need to be distributed to a plurality of objects, and at present, the plurality of information resources are often randomly distributed to each object in a random distribution mode, so that the information resources distributed by each object may have a phenomenon that the information resources are seriously inconsistent with the service target of the scene, and the willingness of each object to operate through the distributed information resources is influenced; if the business objective is to balance the total amount of the information resources distributed to each object, the information resources are distributed to each object in a random distribution mode, so that the total amount of the information resources distributed by each object may have a relatively large difference, and further the willingness of the object with low total amount of the information resources to operate through the information resources is seriously influenced; when the business objective is to distribute more information resources to the currently-held objects with less information resources, and distribute the information resources to each object in a random distribution manner, the objects with less information resources may be distributed with less information resources, so that the willingness of the objects to operate through the distributed information resources is seriously affected, and therefore how to improve the matching degree of the distribution result of the information resources and the business objective becomes a problem to be considered.

In view of this, the present disclosure designs a resource distribution method, device, apparatus, and computer-readable storage medium, in the embodiments of the present disclosure, by learning object features of each history object in a history object group and a first degree of association of each information resource in a resource allocation set corresponding to the history object group, estimating object features of each target object in a target object group and a second degree of association of each information resource, further determining information resources mapped by each target object according to the estimated second degree of association, and distributing the information resources mapped by each target object to a corresponding target object.

Further, in order to facilitate learning of the first association degree of each information resource in the resource allocation set corresponding to each history object in the history object group, a machine learning model may be constructed as the resource allocation model through machine learning, the first association degree may be learned through training of the resource allocation model, and the resource allocation set corresponding to the target object group output by the trained resource allocation model may be obtained by inputting the object feature of each target object in the target object group into the trained resource allocation model.

The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings:

first, a service scenario and a service target of the embodiment of the present disclosure are exemplarily explained.

In the service scene of the embodiment of the present disclosure, one target object group may include, but is not limited to, Q target objects, including K information resources, and the service scene may, but is not limited to, distribute the K information resources to the Q target objects in the object group; q is a positive integer greater than 1, K is a positive integer, and several examples of service scenarios are given below, as follows:

service scenario a 1: game scene

The information resources in this example are game resources (such as at least one resource that may include, but is not limited to, game gold, game props, game arenas, treasure places, etc.), the target objects are game players, and the example includes Q1 game players and K1 game resources; the game scenario may be, but is not limited to: distributing the K1 game resources to the Q1 game players, wherein the same number of game resources may be distributed to each game player of the Q1 game players, or different numbers of game resources may be distributed to different game players of the Q1 game players, and the like, and those skilled in the art can set the resources according to actual needs.

The business target corresponding to the game scene may include, but is not limited to, business target B1 or business target B2:

business objective B1:

so that the difference of the total value of the game resources distributed by different game players in the Q1 game players is smaller than the first threshold value.

Specifically, the business objective B1 may be: so that the difference of the total value of the game resources distributed by any two game players in the Q1 game players is smaller than a first threshold value; the business objective B1 may also be: so that the difference between the maximum value and the minimum value of the total value of the game resources distributed by the game players in the Q1 game players is smaller than a first threshold value, and the like, the following gives specific examples of several business objectives B1:

when the game resource is a game item, the total value of the game resource may be, but is not limited to, the total power value of the game item, and the business object B1 may be, but is not limited to: so that the difference in the total power value of any two game players of the Q1 distributed resultant game items is less than the first threshold value.

When the game resource is game gold, the total value of the game resource may be, but is not limited to, the total value of the game gold, and the business object B1 may be, but is not limited to: so that the difference in the total value of the game gold distributed by any two game players among the Q1 game players is less than the first threshold value.

Where the game resource is a cache location, the total value of the game resource may be, but is not limited to, the total value of the cache for the game at the cache location (i.e., the total value of the cache for the game placed at the cache location), and the business objective B1 may be, but is not limited to: so that the difference in the total value of the treasure deposit from the treasure place distributed by any two of the Q1 game players is less than the first threshold.

Where the game resource is a game entry, the total value of the game resource may be, but is not limited to, the total value of the game award for the game entry (i.e., the total value of the game award that the game player wins at the game entry), and the business objective B1 may be, but is not limited to: so that the difference of the total value of the game prizes of the game arenas distributed by any two game players in the Q1 game players is smaller than the first threshold value.

Business objective B2:

the game player Q is herein taken as the game player Q who has the lowest player rating or has the least game money or the most number of game losses among the Q1 game players_PThe business objective B2 may be: make game player Q_PThe total value of the game resources distributed is higher than that of other game players, i.e., the game players Q other than the game player Q among the Q1 game players_PAnd (4) a player of the game.

Specifically, the business object B2 may be, but is not limited to being: make game player Q_PThe difference value of the total value of the game resources distributed by the lowest-level game player is larger than a second threshold value, or the game player Q is enabled_PThe difference value between the total value of the game resources distributed by the highest-level game player is larger than a third threshold value; the lowest-level game player is a game player whose total value of game resources distributed by the other game players is the lowest, and the highest-level game player is a game player whose total value of game resources distributed by the other game players is the highest; specific examples of several business objectives B2 are given below:

when the game resource is a game item, the total value of the game resource may be, but is not limited to, the total power value of the game item, and the business object B2 may be, but is not limited to: make game player Q_P1The total power value of the property of the game property distributed is higher than that of other game players, and the other game players are game players Q except the game player Q among the Q2 game players_P1Other game player, game player Q_P1The game player with the lowest player rating or having the fewest game coins among the Q2 game players.

When the game resource is game gold, the total value of the game resource may be, but is not limited to, the total value of the game gold, and the business object B2 may be, but is not limited to: make game player Q_P2The total value of the game money and the lowest-level game player Q are distributed_q1The difference value of the total value of the game resources obtained by distribution is larger than a second threshold value, and the game player Q_P2For the game player with the lowest player level or the game player with the least amount of money for the game among the Q2 game players, the lowest game player Q_q1The game player with the smallest total value of the game gold distributed among the Q1 game players.

As an example, the game is a poker game, the information resource is electronic poker cards, and the game comprises K2 electronic poker cards, each electronic poker card has a card value, and the card values of different electronic poker cards can be the same or different; in this case, the game scenario may be: distributing K2 electronic playing cards to Q1 game players, each game player distributing K3 electronic playing cards and finally K4 playing cards, wherein K2 ═ Q1 × K3+ K4, Q1 and K2 are positive integers greater than 1, and K3 and K4 are positive integers less than K2.

In a game scenario corresponding to the poker game, the business object B1 may be, but is not limited to: so that the total size of the card values of the electronic playing cards dealt by different game players among the Q1 game players differs by less than a first threshold value.

In a game scenario corresponding to the poker game, the business object B2 may be, but is not limited to: make game player Q_P3The total value of the distributed electronic playing cards is higher than that of other game players, namely the game player Q_P3May be the game player with the lowest player rating or having the least amount of game money or the most number of historical session losses among the Q1 game players.

Service scenario a 2: reviewing scenarios of media stream resources

The information resources in this example are media stream resources (such as may include, but are not limited to, at least one of video streams, audio streams, etc.), including Q2 auditors and K5 media stream resources; the scenario may be, but is not limited to being: distributing K5 media stream resources to Q2 auditors to audit the media stream resources, wherein the same number of media stream resources can be distributed to each auditor in Q2 auditors, or different numbers of media stream resources can be distributed to different auditors in Q2 auditors, and the like, and a person skilled in the art can set the resources according to actual requirements.

The business objective corresponding to the scenario may include, but is not limited to, business objective C1 or business objective C2:

business objective C1:

the difference value of the total value of the media stream resource distributed by different auditors in the Q2 auditors is smaller than the fourth threshold, and the total value of the media stream resource may be, but is not limited to, the total value of the play duration of the media stream resource, such as the normal play duration of a video stream or the normal play duration of an audio stream.

Specifically, the business objective C1 may be: enabling the difference value of the total value of the media stream resources distributed by any two auditors in the Q2 auditors to be smaller than a fourth threshold value; the business objective C1 may also be: and the difference value between the maximum value and the minimum value of the total value of the media stream resource distributed by the auditors in the Q2 auditors is smaller than a fourth threshold value and the like.

Business objective C2:

the auditor with the lowest auditor grade among the Q2 auditors is taken as the auditor Q_QThe business objective C2 may be: so that the auditor Q_QThe total value of the media stream resource obtained by distribution is higher than that of other auditors, wherein the other auditors are Q auditors except for the auditor Q2 auditors_QAn auditor outside.

Specifically, the business objective C2 may be, but is not limited to being: so that the auditor Q_QThe difference value between the total value of the media stream resources obtained by distribution and the total value of the media stream resources obtained by distribution of the lowest auditor is larger than a fifth threshold value, or the auditor Q is enabled to be_QThe difference value between the total value of the media stream resources obtained by distribution and the total value of the media stream resources obtained by distribution of the highest auditor is larger than a sixth threshold value; the lowest auditor is the total media stream resource distributed by the other auditorsAnd the auditor with the lowest value, wherein the highest auditor is the auditor with the highest total value of the media stream resources distributed by the other auditors.

Referring to fig. 1, an embodiment of the present disclosure provides a resource distribution method, which specifically includes the following steps:

step S101, obtaining object characteristics of each object in the object group, wherein the object characteristics represent intention values of the objects to operate through distributed information resources.

As an embodiment, in view of improving the accuracy of information resources distributed to each target object, the information resources include game resources in a game scene, the target object includes a game player, and the object feature may include, but is not limited to, at least one of the following features:

a player rating of the game player, wherein the player rating may be determined based on, but not limited to, at least one of a game gold held by the game player, a time at which the game player registered with the game account, winning information or losing information of the game player in a historical session.

A resource value of a game resource owned by a game player; specifically, if the game resource is a game item, the resource value of the game resource may be, but is not limited to, the power value of the game item; when the game resource is game gold, the resource value of the game resource can be but is not limited to the value of the game gold; when the game resource is a treasure place, the resource value of the game resource can be but is not limited to the value of a game treasure placed in the treasure place; when the game resource is a game arena, the resource value of the game resource can be, but is not limited to, a game award obtained after the game player wins the game arena.

Information of the game player's failure in the historical game block; the losing information may include, but is not limited to, at least one of losing number information or losing time information of the game player; the number-of-losing information may be the total number of losing or the number of consecutive losing of the game player in the history block, and the number-of-losing information may be the total number of losing or the number of consecutive losing of the game player in the history block in the set history period, and the like.

Step S102, estimating the object characteristics of each target object and the second relevance of each information resource according to the object characteristics of each history object in a history object group and the first relevance of each information resource in the resource allocation set corresponding to the history object group; the resource allocation set includes mapping relationships between the objects and the information resources.

Step S103, determining the information resource mapped by each target object according to the estimated second association degree, and determining the resource allocation set corresponding to the target object group according to the information resource mapped by each target object.

Step S104, distributing each information resource to each target object according to the mapping relationship between each target object and each information resource in the resource allocation set corresponding to the target object group.

As an embodiment, the step S102 and the step S103 specifically include:

and inputting the object characteristics of each target object by adopting a trained resource allocation model, and obtaining a resource allocation set corresponding to the target object group output by the resource allocation model, wherein the resource allocation model is obtained by adopting the object characteristics of each historical object in the historical object group and the resource allocation set corresponding to the historical object group as training samples and training the object characteristics of each historical object in the historical object group and the resource allocation set corresponding to the historical object group based on a machine learning method.

As an embodiment, as shown in fig. 2, the above-mentioned trained resource allocation model can be obtained by, but is not limited to, the following ways:

step S201, extracting object characteristics of each historical object in a historical object group in each training sample and a resource allocation set corresponding to the historical object group;

step S202, adjusting a resource allocation model of historical training or an initially created resource allocation model through machine learning according to the object characteristics of the historical objects and the mapping relation of the information resources in the resource allocation set corresponding to the historical object group to obtain the trained resource allocation model.

The manner in which the trained resource allocation model is obtained is further described below.

The first model acquisition mode: the initially created resource allocation model is adjusted.

Referring to fig. 3, the method specifically includes the following steps:

step S301, training samples are acquired, and a resource allocation model initially created by machine learning.

The training sample in the embodiment of the present disclosure may be an object feature of each historical object in a historical object group obtained from a server of a game and a resource allocation set corresponding to the historical object group, where the training sample may be historical data within a set time duration before a current time, such as historical data within a month before the current time, or may be historical data of a set number before the current time, such as 1 ten thousand historical data before the current time, where the current time may be understood as a time when an initial resource allocation model is created or a time when the current resource allocation model is triggered to be obtained, and the historical data may include the object feature of each historical object in the historical object group and the resource allocation set corresponding to the historical object group.

Step S302, extracting the object characteristics of each historical object in the historical object group in each training sample and the resource allocation set corresponding to the historical object group.

Step S303, adjusting the initially created resource allocation model through machine learning according to the mapping relationship between the object features of the historical objects and the information resources in the resource allocation set corresponding to the historical object group, so as to obtain the trained resource allocation model.

After the initially created resource allocation model is adjusted to obtain the current resource allocation model, the current resource allocation model may be used to perform a prediction operation on the object group to be predicted, so as to obtain a resource allocation set corresponding to the object group to be predicted, and in the process of predicting the object group to be predicted by using the current resource allocation model, the current resource allocation model may be adjusted to obtain a new resource allocation model, for example, a time period is set, and the current machine learning model is periodically adjusted.

Here, after the current resource allocation model is used to predict the resource allocation set corresponding to the target group, the current resource allocation model may be regarded as a historically trained resource allocation model, and the historically trained resource allocation model may be adjusted to obtain the current resource allocation model, or the current resource allocation model may be adjusted to obtain the resource allocation model to be used next time.

The second model acquisition mode: and adjusting the resource allocation model of the historical training.

Referring to fig. 4, the method specifically includes the following steps:

step S401, updating the training sample based on the historical data before the current time.

It should be noted that the current time here may be understood as a time for triggering the resource allocation model to update, for example, a time interval between the current time and a time when the last trained resource allocation model is obtained reaches the set time period, or a time for triggering the resource allocation model to update due to some reason, such as service upgrade, etc.

Specifically, the historical data within a set time length before the current time may be determined as the training sample, for example, the historical data generated between the current time and the time when the last training resource allocation model is obtained may be determined as the training sample; the historical data of a second set amount before the current time can be determined as sample data; the historical browsing data within a set time length before the current time and the training sample used by the last training resource allocation model can be determined as the training sample of the step, for example, the training sample used by the last training resource allocation model and the historical data generated between the current time and the time of obtaining the last training resource allocation model are determined as the training samples.

Step S402, extracting the object characteristics of each history object in the history object group and the resource allocation set corresponding to the history object group in the updated training sample.

Step S403, adjusting the resource allocation model of the historical training by machine learning according to the mapping relationship between the object features of each historical object in the historical object group and each information resource in the resource allocation set corresponding to the historical object group in the updated training sample, so as to obtain the trained resource allocation model.

As an embodiment, when a service scene includes K information resources, in step S203, step S303, or step S403, the initially created resource allocation model or the historically trained resource allocation model may be adjusted by the updated training sample based on the following manner, so as to obtain the trained resource allocation model.

As shown in fig. 5, the method specifically includes the following steps:

step S501, inputting object characteristics of each historical object in a historical object group in a training sample into a resource allocation model to be trained, performing M rounds of resource allocation operations, and determining the distribution sequence of M information resources in K information resources; m is a positive integer not greater than K.

Wherein the resource allocation model to be trained is the initially created resource allocation model or a historically trained resource allocation model.

Step S502 is to estimate the information resource mapped to each history object in the history object group based on the mapping relationship between the distribution order of the M information resources and the distribution order corresponding to each history object in the history object group.

Step S503, determining an estimated resource allocation set corresponding to the history object group according to the information resource mapped by each history object in the history object group.

Step S504, based on the estimated deviation degree of the resource allocation set and the resource allocation set in the training sample, the resource allocation model to be trained is adjusted to obtain a trained resource allocation model.

And step S505, when the deviation degree of the resource allocation set in the training sample based on the estimated resource allocation set is determined to be smaller than the deviation degree threshold, outputting the trained resource allocation model.

As shown in fig. 6, as an embodiment, the ith (i is a positive integer smaller than M) round of resource allocation operations in the M rounds of resource allocation operations in step S501 may specifically, but not limited to, include the following steps:

step S601, determining the object feature of each history object in the history object group and the target resource as the current resource prediction feature, where the target resource includes the information resource whose distribution order is currently determined.

Step S602, estimating a fourth degree of association between the current resource feature and each of the K information resources based on a third degree of association between the historical resource prediction feature and each of the K information resources; and determining the information resource of the current resource prediction feature mapping according to the estimated fourth relevance.

Specifically, the current resource feature and each of the K information resources have a fourth degree of association, and the information resource with the largest fourth degree of association may be, but is not limited to, determined as the information resource mapped by the current resource prediction feature.

Step S603, determining the distribution order of the information resources mapped by the current resource prediction feature as the i, and determining the information resources mapped by the current resource prediction feature as the target resources.

After step S603, the round (i + 1) of resource allocation operation is entered, where i is 1, the target resource in step S601 may be, but is not limited to, set to a null value or specific identification information to represent that there is no information resource for which the distribution order is currently determined.

As an example, in step S502, a mapping relationship between the distribution order of the M information resources and the distribution order corresponding to each history object in the history object group, that is, a distribution order corresponding to each history object may be predetermined, such as allocating different distribution orders of the M information resources determined by M resource allocation operations to different history objects; for ease of understanding, a specific example is given here:

here, taking the example that Q is 3 and M is 51, that is, 3 history objects are included in one history object group, 51 rounds of resource allocation operations are performed in step S501 to determine the distribution order of 51 information resources; specifically, but not limited to, the distribution order determined by the 1 st to 17 th resource allocation operations may be mapped to the distribution order of the history object C1, the distribution order determined by the 18 th to 34 th resource allocation operations may be mapped to the distribution order of the history object C2, and the distribution order determined by the 35 th to 51 th resource allocation operations may be mapped to the distribution order of the history object C3, where the history objects C1 to C3 are 3 history objects in the history object group.

In the above process, the distribution order determined by the 1 st, 4 th, 7 th, 10 th, 13 th, 16 th, 19 th, 22 th, 25 th, 28 th, 31 th, 34 th, 37 th, 40 th, 43 th, 46 th and 49 th resource allocation operations may be mapped to the distribution order of the history object C1; mapping the distribution sequence determined by the 2 nd, 5 th, 8 th, 11 th, 14 th, 17 th, 20 th, 23 th, 26 th, 29 th, 32 th, 35 th, 38 th, 41 th, 44 th, 47 th and 50 th rounds of resource allocation operation to the distribution sequence of the history object C2; the distribution order determined by the 3 rd, 6 th, 9 th, 12 th, 15 th, 18 th, 21 th, 24 th, 27 th, 30 th, 33 th, 36 th, 39 th, 42 th, 45 th, 48 th, 51 th resource allocation operation is mapped to the distribution order of the history object C3.

The distribution sequence determined by the 1-3, 10-12, 19-21, 28-30, 37-39 and 46-47 resource allocation operations can be mapped to the distribution sequence of the history object C1; mapping the distribution sequence determined by the 4 th-6 th, 13 th-15 th, 22 th-24 th, 31 th-33 th, 40 th-42 th and 48 th-49 th resource allocation operations to the distribution sequence of the historical object C2; the distribution order determined by the 7 th-9 th, 16 th-18 th, 25 th-27 th, 34 th-36 th, 43 th-45 th and 50 th-51 th resource allocation operations is mapped to the distribution order of the history object C3.

It should be noted that, according to other manners, a person skilled in the art may also set a mapping relationship between the distribution order of the M information resources and the distribution order corresponding to each history object in the history object group, which is not limited herein.

As an example, in step S503, if M is the same as K, the mapping relationship between each history object and the information resource in the history object group obtained in step S502 may be determined as an estimated resource allocation set corresponding to the history object group, where for convenience of understanding, the resource allocation set is represented by the following table 1:

table 1: resource allocation aggregation

K1 to k in Table 1_MThe identification information of M information resources, and 1 to M are distribution sequences of the information resources.

As an embodiment, when there are different information resources with the same resource value in the M information resources, the resource allocation set may also be set in the form of table 2 below.

Table 2: resource allocation aggregation

Each resource value in table 2 (including the above-mentioned resource value 1 to resource value 6) includes three different information resources of types 1 to 3, and numerals 1 to 18 represent the distribution order of the information resources corresponding to the resource values and types.

If M is a value smaller than K, the mapping relationship between each history object and the information resource in the history objects obtained in step S502 and the information resources other than the information resources whose distribution order is determined by the M rounds of resource allocation operations in the K information resources may be determined as the pre-estimated resource allocation set corresponding to the history object group; in step S504, the other information resources may be randomly distributed to the history objects in the history object group, or distributed to specific history objects in the history object group, and a person skilled in the art may set a distribution manner of the other information resources according to actual needs, where for convenience of understanding, the resource allocation set is represented in table 3 below:

table 3: resource allocation aggregation

K1 to k in Table 3_MThe identification information of M information resources, and 1 to M are distribution sequences of the information resources.

As an embodiment, when there are different information resources with the same resource value in the M information resources, the resource allocation set may be set in the form of table 4 below.

Table 4: resource allocation aggregation

Each resource value in table 4 (including the above-mentioned resource value 1 to resource value 5) includes three different information resources of types 1 to 3, and numerals 1 to 15 represent the distribution order of the information resources corresponding to the resource values and types.

After obtaining the trained resource distribution model according to the above-mentioned manner, the resource allocation set corresponding to the target object group may be obtained as follows, specifically, please refer to fig. 7, which includes the following steps:

step S701, inputting the object characteristics of each target object in the target object group into the trained resource allocation model, performing M rounds of resource allocation operations, and determining a distribution order of the M information resources, where M is a positive integer not greater than K.

Step S702 is to determine the information resource mapped to each target object according to the mapping relationship between the distribution order of the M information resources and the distribution order corresponding to each target object.

Step S703 is to determine a resource allocation set corresponding to the target object group according to the information resource mapped by each target object.

As an embodiment, referring to fig. 8, in step S701, the resource allocation operation of the ith (i is a positive integer smaller than M) round in the resource allocation operations of the M rounds specifically includes the following processes:

step S801 determines the object characteristics of each target object and the target resources including information resources whose distribution order is currently determined as the current resource prediction characteristics.

Step S802, estimating a fourth degree of association between the current resource feature and each information resource based on a third degree of association between the historical resource prediction feature and each information resource; and determining the information resource of the current resource prediction feature mapping according to the estimated fourth relevance.

Specifically, the current resource feature and each of the information resources have a fourth degree of association, and the information resource with the maximum fourth degree of association may be, but is not limited to, determined as the information resource mapped by the current resource prediction feature.

Step S803, determining the distribution order of the information resources mapped by the current resource prediction feature as the i, and determining the information resources mapped by the current resource prediction feature as the target resources.

As an embodiment, considering that the game scenario may be that the same number of game resources are distributed to each game player of the above Q1 game players, and the scenario of auditing the media stream resources may be that the same number of media stream resources are distributed to each auditor of the above Q2 auditors, so the object group in the embodiment of the present disclosure may include a first number of objects Y1, and the resource allocation set may include a second number of information resources Y2, where Y2 is greater than a preset multiple of Y1, where a person skilled in the art may set the preset multiple according to actual needs, such as setting the preset multiple to 3, 5, 16, 17, etc., such as setting Y2 ≧ 17 × Y1, etc.

A specific example of distributing information resources by using the resource distribution method provided by the embodiment of the present disclosure is given below.

The method is applied to a game scene in a poker game, wherein a target object group in the game scene comprises 3 game players (objects), the game scene comprises 54 electronic poker cards as information resources, the resource value of the information resources is the card value of the electronic poker cards, and the game scene comprises the following steps: distributing 54 electronic playing cards to 3 game players in the target object group, distributing 17 electronic playing cards by each game player, and finally, taking the remaining 3 electronic playing cards as other information resources.

The business goals in this example are: the number of the game players in the target object group is increased, that is, it can be understood that the business objective is to make the difference of the total size of the card values of the electronic playing cards distributed by different game players among 3 game players in the target object group smaller than the first threshold, so that the winning probability of the 3 game players tends to be consistent, and the number of the game players in the 3 game players is increased.

The structural diagram of the resource allocation model in this example can refer to the structural diagram of the neural network illustrated in fig. 9, the basic structure of the resource allocation model is composed of nonlinear change units, and the resource allocation model has strong nonlinear mapping capability, and can well learn the nonlinear relationship between the input and the output of the resource allocation model.

The example mainly includes a training process and a using process of the resource allocation model, specifically:

and (I) training a resource allocation model.

Firstly, extracting the game matching data of each game player (namely the historical object group) from the historical data of the online game player historical match, filtering out the game matching data of the game players for multiple times of continuous match as a training sample of a resource allocation model, wherein the input data of the resource allocation model comprises the object characteristics of 3 game players in the historical object group, and the output data is a resource allocation set corresponding to the historical object group, wherein the object characteristics of the 3 game players can be but are not limited to game gold coins, continuous losing times, player grades of the game players and the like of the game players; specifically, refer to the above steps S201 to S202, steps S301 to S303, steps S401 to 403, and steps S501 to S505, and the repetition will not be described.

In the process of outputting the resource allocation set corresponding to the history object group by the resource allocation model in this example, 54 rounds of resource allocation operations need to be performed, in each round of resource allocation operations, the input data of the resource allocation model includes object features of 3 game players in the history object group and current target resources (i.e., electronic playing cards whose distribution sequence is currently determined), and the output data of the resource allocation model includes the distribution sequence of the screened electronic playing cards in the round of resource allocation operations, which may be specifically referred to steps S601 to S603, and the repeated points are not described again.

In this example, the resource allocation set is set as an array with a length of 54 (i.e. the array includes 54 elements, each element corresponding to an electronic playing card), and a specific example of the obtained resource allocation set is provided below, as shown in table 5.

Table 5: resource allocation aggregation

The first row in table 5 represents the card value of the electronic playing card, the first column represents the type of the electronic playing card, the numbers 0-53 in the second row to the 4 th row represent the distribution sequence of the electronic playing cards, wherein the electronic playing cards with the distribution sequence of 0-16 can be set to map one game player, the electronic playing cards with the distribution sequence of 17-33 can be set to map one game player, the electronic playing cards with the distribution sequence of 34-50 map one game player, the electronic playing cards with the distribution sequence of 51-53 are other game resources, and other game resources can be distributed to the designated objects in the object group, and the process of obtaining the table 5 is shown in fig. 10 without being described in more detail.

And (II) a use process of the resource allocation model.

Referring to fig. 11, specifically, when dealing is started after a game is started, whether a distribution process of control information resources is triggered or not may be determined, if the distribution process of control information resources is triggered, object features of 3 game players (target objects) in a target object group are extracted, the object features of the 3 game players in the target object group are input into a trained resource allocation model, 54 rounds of resource allocation operations are performed by the trained resource allocation model to obtain a distribution sequence of 54 electronic playing cards, a resource allocation set corresponding to the target object group is obtained from the distribution sequence of the 54 electronic playing cards, and the 54 electronic playing cards are distributed to the 3 game players in the target object group according to the resource allocation set corresponding to the target object, where the specific process may be referred to the above steps S101 to S103, steps S701 to S703, and steps S801 to S803, the description is not repeated here; if the distribution process of the control information resource is not triggered, 54 electronic playing cards can be randomly distributed to 3 game players in the target object group.

In the process, the number of times of continuous game-playing of the game players in the target object group can be increased, and according to the statistical result of the ABtest data, after the resource distribution method provided by the embodiment of the disclosure is adopted to distribute the electronic poker cards for the game players in the target object group, the continuous game-playing duration of the game players is increased by about 10%.

As shown in fig. 12, based on the same inventive concept, an embodiment of the present disclosure further provides a resource distribution apparatus 1200, including:

a characteristic obtaining unit 1201 configured to perform obtaining an object characteristic of each target object in a target object group, where the object characteristic represents an intention value of the target object to operate through distributed information resources;

a resource allocation unit 1202, configured to estimate, according to an object feature of each history object in a history object group and a first relevance degree of each information resource in a resource allocation set corresponding to the history object group, a second relevance degree of the object feature of each target object and each information resource; the resource allocation set corresponding to the history object group comprises mapping relations between the history objects in the history object group and the information resources, the information resources mapped by the target objects are determined according to the estimated second relevance, and the resource allocation set corresponding to the target object group is determined according to the information resources mapped by the target objects;

a resource distributing unit 1203, configured to execute mapping relationships between the target objects and the information resources in the resource allocation sets corresponding to the target object groups, and distribute the information resources to the target objects.

As an embodiment, the resource allocation unit 1202 is specifically configured to perform:

and adjusting a resource allocation model of historical training or an initially created resource allocation model through machine learning according to the mapping relation between the object features of the historical objects and the information resources in the resource allocation set corresponding to the historical object group to obtain the trained resource allocation model.

As an embodiment, the resource allocation set includes K information resources, where K is an integer greater than 1; the above-mentioned resource allocation unit 1202 is specifically configured to perform:

As an embodiment, the resource allocation unit 1202 is specifically configured to perform the ith round of resource allocation operation as follows, where i is a positive integer smaller than M:

estimating a fourth degree of association between the current resource feature and each information resource based on a third degree of association between the historical resource prediction feature and each information resource; and

and determining the distribution sequence of the information resources mapped by the current resource prediction characteristics as the i, and determining the information resources mapped by the current resource prediction characteristics as the target resources.

As an embodiment, the object group includes a first number of objects, and the resource allocation set includes a second number of information resources, where the second number is greater than a preset multiple of the first number.

As an embodiment, the information resource includes a game resource in a game, the target object includes a game player, and the object feature includes at least one of:

a player rating of the game player;

a resource value of a game resource owned by the game player;

the information of the game player's failure in the historical game.

As shown in fig. 13, the present disclosure provides an electronic device 1300 comprising a processor 1301, a memory 1302 for storing the processor-executable instructions described above;

wherein the processor is configured to execute any one of the resource distribution methods of the present disclosure.

In an exemplary embodiment, a storage medium comprising instructions, such as a memory comprising instructions, executable by a processor of the electronic device to perform the method is also provided. Alternatively, the storage medium may be a non-transitory computer readable storage medium, for example, which may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for resource distribution, comprising:

estimating second relevance of the object characteristics of each target object and each information resource according to the object characteristics of each history object in a history object group and the first relevance of each information resource in a resource allocation set corresponding to the history object group; the resource allocation set corresponding to the history object group comprises the mapping relation between each history object in the history object group and each information resource, and

2. The method according to claim 1, wherein the step of estimating the object feature of each target object and the second degree of association of each information resource by the object feature of each history object in the history object group and the first degree of association of each information resource in the resource allocation set corresponding to the history object group, determining the information resource mapped by each target object according to the estimated second degree of association, and determining the resource allocation set corresponding to the target object group according to the information resource mapped by each target object comprises:

3. The method of claim 2, wherein the trained resource allocation model is obtained by:

4. The method according to claim 2 or 3, wherein the resource allocation set comprises K information resources, wherein K is an integer greater than 1; the step of inputting the object characteristics of each target object by using the trained resource allocation model and obtaining the resource allocation set corresponding to the target object group output by the resource allocation model includes:

5. The method of claim 4, wherein an ith round of resource allocation operations, the i being a positive integer less than the M, comprises:

6. A method according to any of claims 1-3, wherein a first number of objects is included in said set of objects, and a second number of information resources is included in said resource allocation set, said second number being larger than a preset multiple of said first number.

7. A method according to any one of claims 1 to 3, wherein the information resource comprises a game resource in a game, the target object comprises a game player, and the object characteristics comprise at least one of:

a player rating of the game player;

a resource value of a game resource owned by the game player;

information of a failure of the game player in a historical game session.

8. A resource distribution apparatus, comprising:

9. An electronic device comprising a processor, a memory for storing instructions executable by the processor;

wherein the processor is configured to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon computer instructions which, when executed on a computer, cause the computer to perform the method of any one of claims 1-7.