CN109821244A

CN109821244A - Game data processing method, device, storage medium and electronic device

Info

Publication number: CN109821244A
Application number: CN201910054620.6A
Authority: CN
Inventors: 陶建容; 李�浩; 巩琳霞; 冯潞潞; 沈乔治; 范长杰
Original assignee: Netease Hangzhou Network Co Ltd
Current assignee: Netease Hangzhou Network Co Ltd
Priority date: 2019-01-21
Filing date: 2019-01-21
Publication date: 2019-05-31
Anticipated expiration: 2039-01-21
Also published as: CN109821244B

Abstract

The invention discloses a kind of game data processing method, device, storage medium and electronic devices.This method comprises: the first game camp of prediction and the second game camp, which carry out battle operation, obtains the first object probability of the first battle result, and obtain base values corresponding with first object probability；According to first object object when the goal behavior data for carrying out generating when battle operation in preceding switch game, determine target contribution rate, wherein, first object object is any object in the first game camp, target contribution rate is the ratio that first object object is the contribution that the second battle result is made, and the second battle result is the result that the first game camp and the second game camp carry out that battle operation obtains in working as preceding switch game；Based on base values and target contribution rate, the rewards and punishments result of first object object is determined.Through the invention, achieved the effect that distinguish rewards and punishments of the target object in battle class game.

Description

Game data processing method, device, storage medium and electronic device

Technical field

The present invention relates to data processing fields, are situated between in particular to a kind of game data processing method, device, storage Matter and electronic device.

Background technique

Currently, in scene of game, can have vying each other between object corresponding with player and cooperate, need to The ability of the corresponding object of player is evaluated.

When determining the capacity variance of object corresponding with player, can be determined by artificial rule, for example, passing through The game experiences people abundant such as planning define dependency rule, can be in every innings of end of match, it is specified that triumph side obtains corresponding prize It encourages, and the losing side is accordingly punished.In this way in every innings of end of match, the prize of the different players in triumph side or the losing side It is all identical for punishing.If being matched to two side's wide margins, for example, the side A and the wide margin of the side B, a possibility that side A wins, is not Greatly, when the side A fails, too many score can be also detained to the side A.

Since performance quality, the ability power of each player in triumph side or the losing side are different, if to target object Battle class game in rewards and punishments differentiation is not added, the reward that all players in triumph side obtain is all identical, all players of the losing side by The punishment arrived is also identical, then is inequitable and unreasonable for strong for the wherein ability, player that does very well.

Aiming at the problem that differentiation is not added to rewards and punishments of the target object in battle class game in the prior art, not yet mention at present Effective solution scheme out.

Summary of the invention

The main purpose of the present invention is to provide a kind of game data processing method, device, storage medium and electronic device, At least to solve fighting the technical issues of differentiation is not added in the rewards and punishments in class game to target object.

To achieve the goals above, according to an aspect of the invention, there is provided a kind of game data processing method.The party Method include: predict the first game camp and the second game camp carry out battle operate obtain the first battle result first object it is general Rate, and obtain base values corresponding with first object probability, wherein base values is appointed for adjusting in the first game camp An object carries out the rewards and punishments result that battle operation obtains；Battle operation is carried out in working as preceding switch game according to first object object When the goal behavior data that generate, determine target contribution rate, wherein first object object is any right in the first game camp As target contribution rate is the ratio that first object object is the contribution that the second battle result is made, and the second battle result is first Game camp and the second game camp carry out the result that battle operation obtains in working as preceding switch game；Based on base values and mesh Contribution rate is marked, determines the rewards and punishments result of first object object.

In an alternate embodiment of the invention, second battle result be triumph result in the case where, to first object object according to Rewards and punishments result is rewarded, in the case where the second battle result is failure result, to first object object according to rewards and punishments result It is punished.

In an alternate embodiment of the invention, battle operation is carried out in the first game camp of prediction and the second game camp obtain first Before the first object probability for fighting result, this method further include: obtain the first game camp and fought in historical game play Operate the first historical behavior data generated, the second game camp carries out the second history that battle operation generates in historical game play Behavioral data and the first game camp and the second game camp carry out the history battle knot that battle operation generates in historical game play Fruit；Structure is fought by the first historical behavior data, the second historical behavior data and history to be trained first object model； Predict that the first game camp and the second game camp carry out battle operation and obtain the first object probability of the first battle result to include: The first game camp is obtained when the first current behavior data and the second game for carrying out battle operation generation in preceding switch game Camp is when the second current behavior data for carrying out battle operation generation in preceding switch game；By the first current behavior data and Two current behavior data are input to trained first object model, obtain the first object probability of the first battle result.

In an alternate embodiment of the invention, it is generated according to first object object when carrying out battle operation in preceding switch game Goal behavior data determine that target contribution rate includes: multiple first objects and the second game battle array obtained in the first game camp Multiple second objects in battalion, wherein multiple first objects include first object object；From more in addition to first object object In a first object and multiple second objects, it is randomly selected the second target object of first object quantity, and by first object The second target object matching of quantity is the teammate of first object object, and the will be removed in multiple first objects and multiple second objects Third target object except one target object and the second target object matches as the opponent of first object object；By the first mesh The goal behavior data for marking the goal behavior data of object, the goal behavior data of the second target object and third target object are defeated Enter into trained first object model, obtains the second destination probability of the first battle result；Second destination probability is determined For target contribution rate.

In an alternate embodiment of the invention, after the second target object of first object quantity is randomly selected, this method is also Comprise determining that multiple objective results that the second target object of first object quantity is randomly selected；Under every kind of objective result, The second destination probability is obtained, the second destination probability of the second destination number is obtained, wherein the second destination number is multiple target knots The quantity of fruit；It includes: that the second destination probability of the second destination number is made even that second destination probability, which is determined as target contribution rate, , target contribution rate is obtained.

In an alternate embodiment of the invention, according to first object object when carried out in preceding switch game battle operation when generation Goal behavior data, before determining target contribution rate, this method further include: obtain the first game camp carrying out battle operation When the interbehavior data that generate, wherein goal behavior data include interbehavior data；Is established based on interbehavior data Two object modules, wherein the node in the second object module is used to indicate the object in the first game camp, the second object module In path be used to indicate between object corresponding with the node on path and interbehavior occur；Exist according to first object object When the goal behavior data for carrying out generating when battle operation in preceding switch game, determine that target contribution rate includes: in the second target In model, the third destination number in first object path and the 4th destination number of the second destination path are obtained, wherein the first mesh Mark path by first node corresponding with first object object and in the first game camp in addition to first object object The corresponding second node of object, the second destination path pass through second node；By the ratio of third destination number and the 4th destination number Value, is determined as target contribution rate.

In an alternate embodiment of the invention, according to first object object when carried out in preceding switch game battle operation when generation Goal behavior data, before determining target contribution rate, this method further include: obtain the first game camp and the second game camp The historical behavior data that battle operation generates are carried out in historical game play；Clustering processing is carried out to historical behavior data, is obtained more The sub- historical behavior data of a type；Obtain the corresponding contribution to the history of rate of sub- historical behavior data of multiple types, wherein history Contribution rate is that the sub- historical behavior data of each type are the ratio for the contribution that the second battle result is made；By multiple types and Contribution to the history of rate is trained third object module；Battle operation is carried out in working as preceding switch game according to first object object When the goal behavior data that generate, determine that target contribution rate comprises determining that the target type of goal behavior data；By target type Trained third object module is inputted, target contribution rate is obtained.

In an alternate embodiment of the invention, it is generated according to first object object when carrying out battle operation in preceding switch game Goal behavior data determine that target contribution rate includes: to obtain the contribution rate of multiple objects in the first game camp, wherein every The contribution rate of a object is the ratio that each object is the contribution that the second battle result is made；Pass through the contribution rate pair of multiple objects Target contribution rate is normalized, the target contribution rate that obtains that treated.

In an alternate embodiment of the invention, it is being based on base values and target contribution rate, is determining the rewards and punishments knot of first object object After fruit, this method further include: in the case where goal behavior data fit goal condition, repaired by corresponding with goal condition Positive value amendment rewards and punishments result.

It in an alternate embodiment of the invention, should before correcting rewards and punishments result by correction value corresponding with goal behavior data Method further includes at least one of: so that first object object is continuously available the number of the second battle result in goal behavior data In the case where more than or equal to targeted number, goal behavior data fit goal condition is determined；It indicates and works as in goal behavior data In the case where there is goal behavior in preceding switch game, goal behavior data fit goal condition is determined.

In an alternate embodiment of the invention, it is being based on base values and target contribution rate, is determining the rewards and punishments knot of first object object After fruit, this method further include: in the case that preceding switch game be in the deciding grade and level stage, by rewards and punishments result and target factor it Product, is determined as the target rewards and punishments of first object object as a result, wherein, target factor with first object object triumph number of fields or Person fail number of fields increase and increase.

In an alternate embodiment of the invention, obtaining base values corresponding with first object probability includes: in the second battle knot In the case that fruit is triumph result, first foundation index corresponding with first object probability is obtained, wherein first object probability Bigger, first foundation index is smaller；In the case where the second battle result is failure result, obtain opposite with first object probability The second base values answered, wherein first object probability is bigger, and the second base values is bigger.

In an alternate embodiment of the invention, obtaining base values corresponding with first object probability includes: to obtain and the first mesh Mark the corresponding basic score of probability, wherein basic score carries out battle behaviour for adjusting any object in the first game camp Make obtained rewards and punishments score；Based on base values and target contribution rate, determine that the rewards and punishments result of first object object includes: to be based on Basic score and target contribution rate determine the rewards and punishments score of first object object, wherein fighting result second is triumph result In the case where, increase rewards and punishments score to the raw score of first object object point, in the feelings that the second battle result is failure result Under condition, rewards and punishments score is reduced to the raw score of first object object point.

To achieve the goals above, according to another aspect of the present invention, a kind of game data processing unit is additionally provided.It should Device includes: processing unit, obtains the first battle for predicting that the first game camp and the second game camp carry out battle operation As a result first object probability, and obtain base values corresponding with first object probability, wherein base values is for adjusting Any object carries out the rewards and punishments result that battle operation obtains in first game camp；First determination unit, for according to the first mesh Mark object determines target contribution rate when the goal behavior data for carrying out generating when battle operation in preceding switch game, wherein the One target object is any object in the first game camp, and target contribution rate is that first object object is that the second battle result is done The ratio of contribution out, the second battle result are the first game camp and the second game camp when carrying out in preceding switch game pair The result that war operation obtains；Second determination unit determines first object object for being based on base values and target contribution rate Rewards and punishments result, wherein in the case where the second battle result is triumph result, first object object is carried out according to rewards and punishments result Reward punishes first object object according to rewards and punishments result in the case where the second battle result is failure result.

To achieve the goals above, according to another aspect of the present invention, a kind of storage medium is additionally provided.The storage medium In be stored with computer program, wherein computer program be arranged to operation when execute the embodiment of the present invention game data at Reason method.

To achieve the goals above, according to another aspect of the present invention, a kind of electronic device is additionally provided.The electronic device Including memory and processor, which is characterized in that be stored with computer program in memory, processor is arranged to operation and calculates Machine program is to execute the game data processing method of the embodiment of the present invention.

Through the invention, the first game camp of prediction and the second game camp carry out battle operation and obtain the first battle result First object probability, and obtain corresponding with first object probability base values, wherein base values is for adjustment first Any object carries out the rewards and punishments result that battle operation obtains in game camp；According to first object object when in preceding switch game The goal behavior data generated when battle operation are carried out, determine target contribution rate, wherein first object object is the first game battle array Any object in battalion, target contribution rate are the ratio that first object object is the contribution that the second battle result is made, second pair War result is the result that the first game camp and the second game camp carry out that battle operation obtains in working as preceding switch game；It is based on Base values and target contribution rate determine the rewards and punishments result of first object object, wherein fighting result second is triumph result In the case where, first object object is rewarded according to rewards and punishments result, in the case where the second battle result is failure result, First object object is punished according to rewards and punishments result.Due to the destination probability according to battle result, determining and destination probability Base values, and determine contribution rate of the target object in one innings of game, then obtained according to base values and contribution rate Rewards and punishments solve as a result, to carry out distinguishing rewards and punishments to target object and are fighting the rewards and punishments in class game not to target object The technical issues of adding differentiation, has reached the technical effect distinguished to rewards and punishments of the target object in battle class game.

Detailed description of the invention

The attached drawing constituted part of this application is used to provide further understanding of the present invention, schematic reality of the invention It applies example and its explanation is used to explain the present invention, do not constitute improper limitations of the present invention.In the accompanying drawings:

Fig. 1 is a kind of hardware block diagram of the mobile terminal of game data processing method according to an embodiment of the present invention；

Fig. 2 is a kind of flow chart of game data processing method according to an embodiment of the present invention；

Fig. 3 is a kind of stream of the method for ability scoring based on player's multi-mode behavior expression according to an embodiment of the present invention Cheng Tu；

Fig. 4 is a kind of schematic diagram of winning rate prediction model according to an embodiment of the present invention；

Fig. 5 is a kind of schematic diagram of Rating Model based on figure according to an embodiment of the present invention；And

Fig. 6 is a kind of schematic diagram of game data processing unit according to an embodiment of the present invention.

Specific embodiment

It should be noted that in the absence of conflict, the features in the embodiments and the embodiments of the present application can phase Mutually combination.The present invention will be described in detail below with reference to the accompanying drawings and embodiments.

In order to make those skilled in the art more fully understand application scheme, below in conjunction in the embodiment of the present application Attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is only The embodiment of the application a part, instead of all the embodiments.Based on the embodiment in the application, ordinary skill people Member's every other embodiment obtained without making creative work, all should belong to the model of the application protection It encloses.

It should be noted that the description and claims of this application and term " first " in above-mentioned attached drawing, " Two " etc. be to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that using in this way Data be interchangeable under appropriate circumstances, so as to embodiments herein described herein.In addition, term " includes " and " tool Have " and their any deformation, it is intended that cover it is non-exclusive include, for example, containing a series of steps or units Process, method, system, product or equipment those of are not necessarily limited to be clearly listed step or unit, but may include without clear Other step or units listing to Chu or intrinsic for these process, methods, product or equipment.

Embodiment of the method provided by the embodiment of the present application can be in mobile terminal, terminal or similar operation It is executed in device.For running on mobile terminals, Fig. 1 is a kind of game data processing method according to an embodiment of the present invention Mobile terminal hardware block diagram.As shown in Figure 1, mobile terminal 10 may include that one or more (only shows one in Fig. 1 It is a) (processor 102 can include but is not limited to the processing of Micro-processor MCV or programmable logic device FPGA etc. to processor 102 Device) and memory 104 for storing data, in an alternate embodiment of the invention, above-mentioned mobile terminal can also include for leading to The transmitting device 106 and input-output equipment 108 of telecommunication function.It will appreciated by the skilled person that knot shown in FIG. 1 Structure is only to illustrate, and does not cause to limit to the structure of above-mentioned mobile terminal.For example, mobile terminal 10 may also include than in Fig. 1 Shown more perhaps less component or with the configuration different from shown in Fig. 1.

Memory 104 can be used for storing computer program, for example, the software program and module of application software, such as this hair The corresponding computer program of one of bright embodiment game data processing method, processor 102 are stored in storage by operation Computer program in device 104 realizes above-mentioned method thereby executing various function application and data processing.Memory 104 may include high speed random access memory, may also include nonvolatile memory, and such as one or more magnetic storage device dodges It deposits or other non-volatile solid state memories.In some instances, memory 104 can further comprise relative to processor 102 remotely located memories, these remote memories can pass through network connection to mobile terminal 10.The example of above-mentioned network Including but not limited to internet, intranet, local area network, mobile radio communication and combinations thereof.

Transmitting device 106 is used to that data to be received or sent via a network.Above-mentioned network specific example may include The wireless network that the communication providers of mobile terminal 10 provide.In an example, transmitting device 106 includes a Network adaptation Device (Network Interface Controller, referred to as NIC), can be connected by base station with other network equipments to It can be communicated with internet.In an example, transmitting device 106 can for radio frequency (Radio Frequency, referred to as RF) module is used to wirelessly be communicated with internet.

A kind of a kind of game data processing method for running on above-mentioned mobile terminal is provided in the present embodiment, and Fig. 2 is A kind of flow chart of game data processing method according to an embodiment of the present invention.As shown in Fig. 2, the process includes the following steps:

Step S202, the first game camp of prediction and the second game camp carry out battle operation and obtain the first battle result First object probability, and obtain base values corresponding with first object probability.

Above-mentioned steps S202 of the present invention provide technical solution in, prediction the first game camp and the second game camp into Row battle operation obtains the first object probability of the first battle result, and obtains basis corresponding with first object probability and refer to Mark, wherein base values is used to adjust any object in the first game camp and carries out the rewards and punishments result that battle operation obtains.

In this embodiment, target game scene can be battle class scene of game, for example, being Massively Multiplayer Online role Game for play (Massive Multiplayer Online Role Playing Game, referred to as MORPG) scene, sport are competing Skill game (Sports Game, referred to as SPG) scene, more online online competitive game (Multi target object (Online of people Battle Arena, referred to as MOBA) scene etc..First game camp and the second game camp are to carry out in target game scene The both sides of operation are fought, the first game camp and the second game camp include object corresponding with multiple game players, That is, the first game camp and the second game camp include the game role controlled by multiple game players.

In target game scene, multiple objects in the first game camp are the relationship cooperated with each other, the second game battle array Multiple objects in battalion are the relationship cooperated with each other, multiple in the multiple objects and the second game camp in the first game camp Object is the relationship vied each other.Due to the difference of the abilities such as the game operation of game player, consciousness, in the first game camp Performance of multiple objects in target game scene in multiple objects and the second game camp is also just different, can pass through first Performance of the multiple objects in multiple objects and the second game camp in corresponding playing method or bout in game camp, To identify capacity variance.

The embodiment predicts that the first game camp and the second game camp carry out battle operation and obtain the first battle result First object probability, this first battle result namely prediction battle win-or-lose result, can be with respect to the second game camp for For triumph battle as a result, can also be unsuccessfully battle as a result, first object probability can be used for for opposite second game camp Predict the capabilities gap in the first game camp and the second game camp.When the first battle result is triumph battle result, first Destination probability is winning rate, that is, the first game camp is indicated a possibility that defeating the second game camp by first object probability；? First battle result is when unsuccessfully fighting result, and first object probability is failure rate, that is, the first game camp is defeated by the second trip A possibility that play camp, is indicated by first object probability.

The first mesh of the first battle result is obtained predicting that the first game camp and the second game camp fight operating After marking probability, base values corresponding with first object probability is obtained, the base values is for adjusting the first game camp Middle any object carries out the rewards and punishments that battle operation obtains as a result, can be the first game camp and the match victory or defeat of the second game camp Basic score, that is, to add and subtract basic score base_score.The embodiment can pass through first object probability and coefficient of rewards and punishment Determine base values, which is used to adjust the size of base values, can rule of thumb preset.When basis refers to It is designated as adding and subtracting basic timesharing, coefficient of rewards and punishment can be positive or negative points coefficient.

For example, the first game camp is the side A in target game scene, and the second game camp is in scene of game The side B, the first battle result are triumph battle as a result, first object probability is R (value of the R between 0-1), that is, the side A wins the side B A possibility that be R, score base is added and subtracted based on the base values of the side A_scoreA, base_scoreAIt can be calculated by the following formula It arrives:

Wherein, M is positive or negative points coefficient.

When the side A defeats the side B, basis plus-minus score base_scoreAFor M (1-R), when the side A is defeated by the side B, basis plus-minus Score base_scoreAFor MR.If R > 0.5, this competition A can power it is stronger, it is contemplated that can win, increase it is point just corresponding reduce, and R is bigger, and increasing point is fewer；If the side A is defeated, deduction is more；If R < 0.5, this competition A can power it is weaker, it is contemplated that meeting Defeated, deduction is just corresponding to be reduced, and R is smaller, and deduction is fewer, if A team wins, increasing point is more.

In an alternate embodiment of the invention, the calculating on the plus-minus basis point on the plus-minus basis point and above-mentioned side A of the side B of the embodiment Mode is identical, and details are not described herein again.

Step S204, according to first object object when the goal behavior for carrying out generating when battle operation in preceding switch game Data determine target contribution rate.

In the technical solution that above-mentioned steps S204 of the present invention is provided, basis corresponding with first object probability is being obtained After index, according to first object object in the goal behavior data generated when carrying out battle in preceding switch game and operating, really Set the goal contribution rate, wherein first object object is any object in the first game camp, and target contribution rate is first object Object is the ratio for the contribution that the second battle result is made, and the second battle result is that the first game camp and the second game camp exist Obtained result is operated when carrying out battle in preceding switch game.

In this embodiment, the first game camp includes multiple objects, and first object object is any in multiple objects Object, to need to predict the object of the battle ability in target game scene.Due to the game operation of the first mesh object, with the As soon as the differences such as the consciousness of the corresponding game player of target object, performance of the first object object in target game scene is not yet Together.First object object is obtained when the goal behavior data for carrying out generating when battle operation in preceding switch game, the target line For data namely when the statistical data in preceding switch game, it can serve to indicate that first object object when in preceding switch game Behavior expression, for example, the goal behavior data be to secondary attack, backboard, grab, the relevant data such as nut cap.

In this embodiment, the second battle result is the first game camp and the second game camp when in preceding switch game Carry out that battle operation obtains as a result, can be triumph battle result, or unsuccessfully fight result.Since different objects exist Fight capability in target game scene is not identical, then different objects are also different for the contribution of the second battle result. If differentiation is not added to the rewards and punishments of multiple objects in the first game camp, put on an equal footing, then strong for ability, pair that does very well As corresponding game player is inequitable.Thus, the embodiment according to first object object in the preceding switch game into The goal behavior data generated when row battle operation, determine target contribution rate, the target contribution rate namely performance contribution proportion, tribute Degree is offered, is the ratio that first object object is the contribution that the second battle result is made, for example, the second battle result is triumph pair War is as a result, the ability of first object object is stronger, then first object object gets over the ratio for the contribution that the second battle result is made Greatly, the ability of first object object is weaker, then the ratio for the contribution that first object object makes the second battle result is smaller.

In an alternate embodiment of the invention, which is input to goal behavior data in object module, obtains target contribution Rate.Wherein, object module is used to determine the ratio of contribution that object makes battle result, can by historical behavior data and History result of the match is trained to obtain.

The embodiment can by according to the corresponding object of each player when carried out in preceding switch game battle operation when The goal behavior data of generation, to determine target contribution rate, to distinguish the corresponding object of every player when preceding switch game In contribution rate.

Step S206 is based on base values and target contribution rate, determines the rewards and punishments result of first object object.

In the technical solution that above-mentioned steps S206 of the present invention is provided, it is based on base values and target contribution rate, determines the The rewards and punishments result of one target object.

In this embodiment, due in the preceding switch competition game, the performance situation of the corresponding object of different players is not Together, in the reward to each object in triumph side and also should not be identical to the punishment of each object in the losing side, for example, Bonus point to each object in triumph side and also should not be identical to the deduction of each object in the losing side.It is obtaining and the After the corresponding base values of one destination probability and target contribution rate, it is based on base values and target contribution rate, determines first The rewards and punishments of target object as a result, the rewards and punishments result for first object object to be rewarded or is punished, can for rewards and punishments score, Rewards and punishments blood volume, rewards and punishments gold coin, rewards and punishments rank, rewards and punishments holy water, rewards and punishments weapon etc., herein with no restrictions.

As an alternative embodiment, in the case where the second battle result is triumph result, to first object pair As being rewarded according to rewards and punishments result, in the case where the second battle result is failure result, to first object object according to prize Result is punished to be punished.

In the case where the second battle result is triumph result, first object object is encouraged according to above-mentioned rewards and punishments result It encourages, when the target contribution rate that first object object is triumph result contribution is big, that is, first object is working as preceding switch by object Do very well in game relative to other objects, then can give first object object some more rewards, when first object object is victory The target contribution rate hour of sharp result contribution, that is, first object object is in working as preceding switch game relative to other Object tables It is existing poor, it can give first object object less reward；In the case that second battle result is failure result, when first object pair When big as the target contribution rate to failure result, that is, first object object is when relatively other Object tables in preceding switch game It is existing poor, then it can give first object object some more punishment, when target contribution rate of the first object object to failure result is small, That is, first object object relatively other objects in working as preceding switch game do very well, then it is few can to give first object object A little punishment.

The rewards and punishments result of first game camp of the embodiment and other target objects in the second game camp can also be with It is determined by the method for the rewards and punishments result of above-mentioned determining first object object, details are not described herein again.

The embodiment can increase stage by stage stablizing, by object corresponding with player when the difference in preceding switch game Performance, gives the rewards and punishments of the corresponding object difference ability of player as a result, making the ability of the corresponding object of player quickly stable Into the horizontal segment for meeting oneself ability, so that match is more fair, fierce, while the game body of player is improved It tests.

The program according in target game scene the first game camp and second game camp both sides' capabilities gap, and The performance of the corresponding object of each player of authorities' game and ability determine the rewards and punishments of player's post-games as a result, to realize different players Corresponding object obtains reasonable rewards and punishments as a result, the player of different abilities is enabled to distinguish faster after authorities' end of match It comes, but also the ability of player can be stabilized to faster in the horizontal segment of oneself ability, so that the competition of every innings of game is more Fierceness, the sense of participation of player is stronger, and game experiencing is more preferable.

S202 to step S206 through the above steps, the first game camp of prediction and the second game camp carry out battle operation The first object probability of the first battle result is obtained, and obtains base values corresponding with first object probability, wherein basis Index is used to adjust any object in the first game camp and carries out the rewards and punishments result that battle operation obtains；According to first object object When the goal behavior data for carrying out generating when battle operation in preceding switch game, target contribution rate is determined, wherein first object Object is any object in the first game camp, and target contribution rate is that first object object is the tribute that the second battle result is made The ratio offered, the second battle result are that the first game camp and the second game camp carry out battle operation in working as preceding switch game Obtained result；Based on base values and target contribution rate, the rewards and punishments result of first object object is determined, wherein at second pair In the case that result of fighting is triumph result, first object object is rewarded according to rewards and punishments result, is in the second battle result In the case where failure result, first object object is punished according to rewards and punishments result.Since the target according to battle result is general Rate, the determining base values with destination probability, and determine contribution rate of the target object in one innings of game, then according to basis Index and contribution rate obtain rewards and punishments as a result, to carry out distinguishing rewards and punishments to target object, avoid to target object rewards and punishments not Rationally, unfair, it solves and the technical issues of differentiation is not added in the rewards and punishments in class game is being fought to target object, reached to mesh The technical effect that rewards and punishments of the mark object in battle class game distinguish.

As an alternative embodiment, the first game camp of prediction and the second game camp carry out in step S202 Before battle operation obtains the first object probability of the first battle result, this method further include: obtain the first game camp and going through The first historical behavior data of battle operation generation are carried out in history game, the second game camp carries out battle behaviour in historical game play Make the second historical behavior data generated and the first game camp and the second game camp carries out battle operation in historical game play The history of generation fights result；Structure is fought to first by the first historical behavior data, the second historical behavior data and history Object module is trained；Step S202, the first game camp of prediction and the second game camp carry out battle operation and obtain first The first object probability of battle result includes: to obtain the first game camp to carry out fighting what operation generated in working as preceding switch game First current behavior data and the second game camp are when the second current behavior for carrying out battle operation generation in preceding switch game Data；First current behavior data and the second current behavior data are input to trained first object model, obtain first Fight the first object probability of result.

In this embodiment it is possible to, for predicting the probability of the first battle result, can be winning rate by training pattern, It may be failure rate, which can be obtained by (1- winning rate).Predict the first game camp and the second game camp into Before row battle operation obtains the first object probability of the first battle result, obtains the first game camp and carried out in historical game play First historical behavior data of battle operation generation, the second game camp carry out battle operation generates second in historical game play Historical behavior data and the first game camp and the second game camp carry out the history pair that battle operation generates in historical game play War as a result, the first historical behavior data can be the history shot and long term behavior representation data of each object in the first game camp, Second historical behavior data can be the history shot and long term behavior representation data of each object in the second game camp, history battle It as a result can be history result of the match.Wherein, shot and long term behavior representation data be used to indicate the corresponding object of player long-term and Performance situation in gaming in a short time, for example, game is basketball game, behavior representation data just includes the corresponding object of player In more Basketball Match average, average backboard, average secondary attack, averagely grab etc. as statistical indicator, wherein it is long The fixture that phase representation data is covered is long, can be one month, that is, object corresponding with player all ratios in one month The statistical indicator of match, the period that acts and efforts for expediency representation data may be covered can be the corresponding object of player in three days The statistical indicator of game.

After obtaining the first historical behavior data, the second historical behavior data and history battle result, by the first history The input of behavioral data and the second historical behavior data as first object model, using history battle result as first object mould The output of type, to be trained to first object model, for example, the case where the first game camp has won the second game camp Under, result 1 is exported, it is on the contrary then export result 0.Wherein, first object model can be deep neural network model.

For example, the first object model of the embodiment is winning rate prediction model, and the basketball that battle class game is 3V3 is swum Play, in total there are six object, the corresponding object of current player including performance point to be calculated, that is, being above-mentioned first object pair As two teammates of first object object, three opponents of first object object, their feature vector can all be by above-mentioned Shot and long term behavior representation data composition, each object will correspond to a vector, then using will be as two objects of teammate Vector sum it up and be averaged, obtain vector a, will sum it up and be averaged as the three of opponent objects, obtain vector b, then Vector sum vector a, the vector b of first object object are stitched together, vector c is obtained, for example, if the first object object Original vector length is 20 dimensions, then the length of vector a, vector b are also all 20 dimensions, then the length of vector c is 3*20=60 Dimension, as soon as then the vector that length is 60 dimensions is input in a deep neural network model, the output of deep neural network model It is the win-or-lose result of the first object object in competition game, when first object object is won in competition game, then model The output of fitting is exactly 1, it is on the contrary then be 0, then by training can be obtained by a winning rate prediction model.

By the first historical behavior data, the second historical behavior data and history fight structure to first object model into Row obtains the first of the first battle result after training, predicting that the first game camp and the second game camp fight operating When destination probability, available first game camp is when the first current behavior for carrying out battle operation generation in preceding switch game Data and the second game camp are when the second current behavior data for carrying out battle operation generation in preceding switch game, wherein the One current behavior data can be the performance data of first object object and teammate in going game match, the second current behavior Data can be performance data of the opponent in going game match of first object object, by the first current behavior data and the Input of the two current behavior data as first object model, is predicted by first object model, obtains a 0-1's Value, this value is just the size of the first object probability of the first battle result.In an alternate embodiment of the invention, when first object model When for winning rate prediction model, the first battle result is triumph as a result, then first object probability is that the first game camp is swum to second The winning rate in play camp.

As an alternative embodiment, step S204, carries out according to first object object when in preceding switch game The goal behavior data generated when battle operation determine that target contribution rate includes: multiple first obtained in the first game camp Multiple second objects in object and the second game camp, wherein multiple first objects include first object object；From except first In multiple first objects and multiple second objects except target object, the second target pair of first object quantity is randomly selected As and the second target object of first object quantity being matched as the teammate of first object object, by multiple first objects and more Third target object in a second object in addition to first object object and the second target object matches as first object object Opponent；By the goal behavior data of first object object, the goal behavior data of the second target object and third target object Goal behavior data be input in trained first object model, obtain the second destination probability of the first battle result；It will Second destination probability is determined as target contribution rate.

In the implementation, the behavior expression in preceding switch game is being worked as according to the corresponding object of each player, it is available The corresponding object of each player is when the behavioral data in preceding switch game, the goal behavior data namely statistical data, statistics Index, for example, for data such as score, secondary attack, backboards.The winning rate in the camp where the corresponding object of the player to do very well, generally Also all want high, thus the embodiment is in the target contribution rate for determining first object object, it can be by constantly calculating this Winning rate of the first object object in Different matching mode calculates.

Obtain multiple second objects in multiple first objects and the second game camp in the first game camp, wherein Multiple first objects include first object object, when the teammate of the first object object of preceding switch game and opponent have determined , the goal behavior data and result of the match of each object are also determining.

Calculate first object object target contribution rate when, from addition to first object object multiple first objects and In multiple second objects, the second target object and first object object that first object quantity is randomly selected form teammate, will Third target object in multiple first objects and multiple second objects in addition to first object object and the second target object is made For the opponent of first object object, in an alternate embodiment of the invention, first object quantity adds the 1 quantity phase with third target object Deng.After obtaining multiple second objects in multiple first objects and the second game camp in the first game camp, by the The goal behavior number of the goal behavior data of one target object, the goal behavior data of the second target object and third target object According to being input in trained first object model, so that it may the first battle result under the matching way determined at random Second destination probability, second destination probability can be for due to the second target object that first object quantity is randomly selected and the One target object forms the random probability of teammate and determination, which can be determined as to target contribution rate, real Show according to first object object when the goal behavior data for carrying out generating when battle operation in preceding switch game, has determined target The purpose of contribution rate, and then it is based on base values and target contribution rate, the rewards and punishments of first object object are determined as a result, at second pair In the case that result of fighting is triumph result, first object object is rewarded according to rewards and punishments result, is in the second battle result In the case where failure result, first object object is punished according to rewards and punishments result, has been reached to target object in battle class The effect that rewards and punishments in game distinguish.

As an alternative embodiment, after the second target object of first object quantity is randomly selected, it should Method further include: determine multiple objective results that the second target object of first object quantity is randomly selected；In every kind of target As a result under, the second destination probability is obtained, obtains the second destination probability of the second destination number, wherein the second destination number is more The quantity of a objective result；It includes: by the second target of the second destination number that second destination probability, which is determined as target contribution rate, Probability is averaged, and obtains target contribution rate.

In this embodiment, the second target object that first object quantity is randomly selected to form with first object object Teammate can haveKind of objective result, wherein X is used to indicate multiple first objects in addition to first object object and multiple The quantity of second object, Y is for indicating first object quantity, in an alternate embodiment of the invention, 2* (Y+1)=X+1.

Above-mentioned every kind of objective result is used to be expressed as the teammate's as a result, in every kind of objective result of first object object matching Under, according to by the goal behavior data and third target object of the goal behavior data of first object object, the second target object Goal behavior data be input in trained first object model, obtain the second destination probability of the first battle result, altogether There is the second destination probability of the second destination number, which is the quantity of multiple objective results, that is, being above-mentionedIn this way when the second destination probability is determined as target contribution rate, the second destination probability of the second destination number is made even It, is target contribution rate by obtained average target determine the probability.

For example, the corresponding object participation competition game of marked as 1-6 players in preceding switch game, is being shared, The teammate and opponent that deserve the corresponding object of No. 1 player in preceding switch game it has been determined that the statistical data of every player and Result of the match has also determined that.It is right from remaining five at random when calculating the target contribution rate of corresponding object of No. 1 player Two objects object composition teammate corresponding with No. 1 player is selected as in, remaining three objects are as opponent, their system It counts to put and be input in winning rate prediction model, so that it may the winning rate under this imaginary matching way is obtained, due to one sharedKind selects mode, so can be calculated for No. 1 playerKind winning rate, then by thisKind winning rate is averaged As the target contribution rate of the corresponding object of No. 1 player, the target contribution of the corresponding object of other players similarly can be obtained by Rate.

Through the above method, in available first game camp and the second game camp in addition to first object object The target contribution rate of other objects, to distinguish each object when the percentage contribution in preceding switch game.

As an alternative embodiment, in step S204, according to first object object when in preceding switch game into The goal behavior data generated when row battle operation, before determining target contribution rate, this method further include: obtain the first game battle array It seeks in the interbehavior data generated when battle operation, wherein goal behavior data include interbehavior data；Based on friendship Mutual behavioral data establishes the second object module, wherein the node in the second object module is used to indicate in the first game camp Object, the path in the second object module, which is used to indicate between object corresponding with the node on path, there is interbehavior； Step S204, according to first object object in the goal behavior data generated when carrying out battle in preceding switch game and operating, really The contribution rate that sets the goal includes: that third destination number and the second target road in first object path are obtained in the second object module 4th destination number of diameter, wherein swum by first node corresponding with first object object and with first in first object path The corresponding second node of object in play camp in addition to first object object, the second destination path pass through second node；By The ratio of three destination numbers and the 4th destination number is determined as target contribution rate.

In this embodiment, target game scene is battle class scene of game, can be related to the behavior interaction of team, can adopt First object object is determined in the target contribution rate in the preceding switch game with the second object module, which can be with For interaction graph model, the interbehavior between object is showed.

According to first object object in the goal behavior data that generate when carrying out battle operation in preceding switch game, really Before the contribution rate that sets the goal, the first game camp is obtained in the interbehavior data generate when battle operation, the interaction row It can be the data of pass behavior for data.The second object module is established based on interbehavior data, it can be by the first game battle array Object in battalion, can be right by two when there is interbehavior between two objects as the node in the second object module As corresponding node with line be connected, form a paths, the path be it is directive, direction for indicate behavior interaction direction, For example, being the transmission direction of ball, the second object module of bout can be constructed in this way.In the second target mould Type is established after completion, and the significance level of each node in this people's object module can be calculated.

According to first object object in the goal behavior data that generate when carrying out battle operation in preceding switch game, really When the contribution rate that sets the goal, in the second object module, the third destination number and the second destination path in first object path are obtained The 4th destination number, the first object path by first node corresponding with first object object and with the first game camp In the corresponding second node of object in addition to first object object, for example, first object object is right in the second object module The node answered is indicated that the object in the first game camp in addition to first object object is corresponding in the second object module by C point Node by PF point, PG point indicate, then first object path be PF point, PG point in all paths pass through C point path, third Destination number is PF point, PG point to the item number in the path on the side for passing through C point in all paths, can be indicated with num (d*)；Second Destination path passes through second node, can be PF, PG point to the item number on all path sides, can be indicated with num (d).Due to warp The second destination path of second node to be crossed, first node will be passed through mostly, it is determined that first object object is important, than Such as, in all paths of PF, PG point pair, C is all passed through in most of paths, that is, the flowing of PF, PG ball will pass through C, it is determined that The corresponding first object object (centre forward) of C is important.

After obtaining the 4th destination number of third destination number and the second destination path in first object path, by the The ratio of three destination numbers and the 4th destination number is determined as target contribution rate, for example, by r_C=num (d^*)/num (d) is determined For the target contribution rate of first object object.

The embodiment can also obtain the target contribution rate of other objects in the first game camp by the above method, thus Each object is distinguished when the percentage contribution in preceding switch game.

As an alternative embodiment, in step S204, according to first object object when in preceding switch game into The goal behavior data generated when row battle operation, before determining target contribution rate, this method further include: obtain the first game battle array Battalion and the second game camp carry out the historical behavior data that battle operation generates in historical game play；Historical behavior data are carried out Clustering processing obtains the sub- historical behavior data of multiple types；Obtain the corresponding history of sub- historical behavior data of multiple types Contribution rate, wherein contribution to the history of rate is that the sub- historical behavior data of each type are the ratio for the contribution that the second battle result is made Example；Third object module is trained by multiple types and contribution to the history of rate；Step S204 exists according to first object object When the goal behavior data for carrying out generating when battle operation in preceding switch game, determine that target contribution rate comprises determining that target line For the target type of data；Target type is inputted into trained third object module, obtains target contribution rate.

In this embodiment, there can be different occupations in target game scene, the evaluation criterion of different occupation is also different, For example, centre forward, which is mainly responsible for, robs backboard, and score would not be especially more, and shooting guard is mainly responsible for throwing in Basketball Match scene Three points, score will mostly etc..Thus, when specific determining rewards and punishments result, it should be gone according to the standard of different occupation It determines, just rewards and punishments can be made more fair and just in this way.

In order to allow rewards and punishments result to be more in line with the subjective assessment of our people, experienced player or planning can first be allowed to go It gives a mark to the performance of the corresponding object of player in bout, is equivalent to obtain expertise in this way, it is then sharp again With there is monitor model to go to learn this expertise, finally there can be monitor model to realize scoring using this.But in game Usually there is hundreds of thousands even millions of match, one one manually to give a mark, workload is too big, which uses Cluster gathers the corresponding object of similar player for one kind, then gives a mark again to this class, is in this way all a kind of player The behavior of corresponding object tends to similitude, and the score that they are obtained should also be similar, to be considerably reduced plan The workload of marking, while being also possible that label is more accurate.

In this embodiment, the first game camp and the second game camp is obtained to carry out battle operation in historical game play and produce Raw historical behavior data, the historical behavior data namely history race statistics data, then gather historical behavior data Class processing, obtains the sub- historical behavior data of multiple types, can be first to all historical behavior data separation occupations, for example, will Historical behavior data separation is that five occupations of C, SG, SF, PG, PF can make in the corresponding historical behavior data of each occupation It is the sub- historical behavior data of multiple types by the corresponding historical behavior data clusters of each occupation with Kmeans clustering method, For example, being 20 classes by each occupation cluster.In cluster result, each occupation is divided into a variety of different types, for example, C occupation in, have the C of parting, have defensive C, there are also backboard type C, to find the similar player couple of behavior well The object answered, while C has been also divided into 20 different grades, that is, the performance of the corresponding object of different type, ability are not With, it should obtain different rewards and punishments.The central point of each class can be used in the embodiment, and (all data is averaged in such Value) the Lai Daibiao category.

Clustering processing is being carried out to historical behavior data, after obtaining the sub- historical behavior data of multiple types, is being obtained more The corresponding contribution to the history of rate of sub- historical behavior data of a type, the contribution to the history of rate are the sub- historical behavior data of each type For the ratio for the contribution that the second battle result is made, the center point data of 20 types of each occupation can be given to planning, Subjective 20 classes in each occupation of planning meeting are ranked up, and are labeled as 1-20, after label completion, obtain each class Label, and determine its contribution to the history of rate.It is being trained to third object module by multiple types and contribution to the history of rate Afterwards, third object module is trained by multiple types and contribution to the history of rate, which can be nerve net Network model, can use the above-mentioned label of neural network model fitting and contribution to the history of rate is trained.In an alternate embodiment of the invention, Different occupation corresponds to different third object modules.

According to first object object in the goal behavior data that generate when carrying out battle operation in preceding switch game, really When determining the target contribution rate of first object object, the target type of the goal behavior data of first object object can be first determined, For example, after a competition game terminates, it, can be first according to the occupation of the corresponding object of player when calculating target contribution rate Type selects third object module corresponding with occupation type, then unites the target type of goal behavior data as match It counts and is input in trained third object module corresponding with occupation type, pass through the trained third object module Calculate the target contribution rate of first object object.

In this embodiment, first object object can also be determined by the Rating Model of NBA efficiency index formula Target contribution rate, that is, object for appreciation directly can be calculated using the statistical data that player competes by existing capacity index The contribution rate of the corresponding object of family, the Rating Model of NBA efficiency index formula can be such that

Contribution rate=[(goals for+secondary attack number+total backboard number+grabs number+nut cap number)-(shooting, which is sold, count-to bury a shot Number)-(penalty shot sell number-penalty shot hits)-make mistakes number] competition sessions of/sportsman.

It in this embodiment, can also be by determining first object object based on the Rating Model of the efficiency value of Gmsc Target contribution rate.It can be such that using the Rating Model of the efficiency value of Gmsc

Contribution rate=score+0.4 × bury a shot number -0.7 × shooting sell number -0.4 × (penalty shot sell number-penalty shot life Middle number)+0.7 × front court backboard number+0.3 × back court backboard number+grab number+0.7 × secondary attack number+0.7 × nut cap number -0.4 × criminal Advise number-fault number.

As an alternative embodiment, step S204, carries out according to first object object when in preceding switch game The goal behavior data generated when battle operation determine that target contribution rate includes: the multiple objects obtained in the first game camp Contribution rate, wherein the contribution rate of each object is the ratio that each object is the contribution that the second battle result is made；By more Target contribution rate is normalized in the contribution rate of a object, the target contribution rate that obtains that treated.

In this embodiment, in order to guarantee the corresponding object of all players in the first game camp and the second game camp The sum of contribution rate is 1, need to the contribution rate of the corresponding object of players all in the first game camp and the second game camp into Row normalization, obtains the contribution rate of multiple objects in the first game camp, by the contribution rate of target contribution rate and multiple objects The sum of quotient, be determined as the target contribution rate after target contribution rate is normalized.

In an alternate embodiment of the invention, pass through formulaContribution rate to be normalized.Wherein, r_iTable Show the contribution rate of the corresponding object of calculated i player, d_iIt is the contribution rate after the corresponding normalized of i player, It is the sum of the contribution rate of the corresponding object of all players in camp where the corresponding object of i player, to ensure that

It can be to the corresponding object of each player in base values by the different contribution rates of the corresponding object of each player On carry out difference rewards and punishments,Wherein S_iFor table Show the rewards and punishments of i-th of player as a result, base_ScoreAFor indicating the base values in the camp A, base_ScoreBFor indicating the camp B Base values, if the corresponding object of i-th of player belongs to the camp A, the base values base in the camp A_ScoreAOn multiplied by The contribution rate d of the corresponding object of i player_i, obtain the rewards and punishments result S of i-th of player_i；If the corresponding object of i-th of player Belong to the camp B, then in the base values base in the camp B_ScoreBOn multiplied by the corresponding object of i-th of player contribution rate d_i, obtain The rewards and punishments result S of i-th of player_i。

The embodiment is after competition game terminates, according to the row when the corresponding object of different players in preceding switch game For performance, the corresponding object of each player is obtained by training pattern and is working as the contribution rate in preceding switch game, then according to table Contribution rate carries out distinguishing rewards and punishments point to the corresponding object of every player on base values, can be to the object to do very well More rewards or few punishment are carried out, few reward or more punishment are carried out to the object of performance difference, to realize that different players are corresponding right As obtain reasonable rewards and punishments after preceding switch competition game terminates as a result, enable the corresponding object of player ability faster Ground is stabilized in the horizontal segment of the ability of oneself, so that the competition of every innings of game is more fierce, the sense of participation of player is stronger, trip Play experience is more preferable.

As an alternative embodiment, being based on base values and target contribution rate in step S206, determining the first mesh Mark object rewards and punishments result after, this method further include: in the case where goal behavior data fit goal condition, by with mesh The corresponding correction value of mark condition corrects rewards and punishments result.

In this embodiment, be based on base values and target contribution rate, determine first object object rewards and punishments result it Afterwards, rewards and punishments result can further be corrected.It may determine that whether goal behavior data meet goal condition, the target item Part is for determining the condition being modified to rewards and punishments result, can be the condition of consecutive victories continuous loss, or preset Optimum value sportsman encourages the condition of the assessment rules of (Most Valuable Player Award, referred to as MVP).In target line In the case where for data fit goal condition, rewards and punishments result is corrected by correction value corresponding with goal condition.

As an alternative embodiment, by correction value corresponding with goal behavior data correct rewards and punishments result it Before, this method further includes at least one of: so that first object object is continuously available the second battle result in goal behavior data Number be more than or equal to targeted number in the case where, determine goal behavior data fit goal condition；Refer in goal behavior data It shows in the case where there is goal behavior in preceding switch game, determines goal behavior data fit goal condition.

In this embodiment, the number for making first object object be continuously available the second battle result in goal behavior data is big In the case where being equal to targeted number, goal behavior data fit goal condition is determined.In an alternate embodiment of the invention, targeted number It is 3 times, at consecutive victories play number >=3 time of first object object, determines goal behavior data fit goal condition, then competing End pair passes through correction value corresponding with goal condition and corrects rewards and punishments result.Revised rewards and punishments result=S_i+ (N-2) * m, In, S_iIt is by the calculated rewards and punishments of front training pattern as a result, (N-2) * m is for indicating amendment corresponding with goal condition Value, for N for indicating user's consecutive victories play, m, can be with self-defining for indicating modified basic coefficients of winning successively.

In an alternate embodiment of the invention, at continuous loss play number >=3 of first object object, true goal behavior data fit Goal condition corrects rewards and punishments result by correction value corresponding with goal condition then in end of match pair.User's rewards and punishments result= S_i+ (M-2) * h, wherein S_iIt is by the calculated rewards and punishments of training pattern as a result, (M-2) * h is for indicating corresponding with goal condition Correction value, M indicate user suffer successive lost play, h for indicate continuous loss modified basis coefficient, can be with self-defining.

The embodiment also indicates in the case where there is goal behavior in preceding switch game in goal behavior data, determines Goal behavior data fit goal condition.For example, setting MVP assessment rules during the games, including capture Basketball Match In final hit ball, key ball (crucial score, crucial backboard, key grab, crucial nut cap, crucial interference, crucial secondary attack etc.), even Continuous ball (run, continuous backboard, continuously grab, continuous nut cap, stepwise derivation, continuously secondary attack etc.), and according to these targets Behavior is finely adjusted rewards and punishments result.It indicates in goal behavior data when there is the case where goal behavior in preceding switch game Under, correcting rewards and punishments result by correction value corresponding with goal condition is=S_i+mvp_score, wherein S_iIt is to pass through training pattern Calculated rewards and punishments are as a result, mvp_scoreIt is intended to indicate that correction value corresponding with goal condition, obtains rewards and punishments result for correcting Constant, the importance according to MVP can be drawn for game strategy, rule of thumb a given in advance numerical value, for example, at one In match, the additional 10 points of reward scores of player for obtaining MVP can be given, then mvp_score=10.

In this embodiment, the corresponding object of each player by trained first object model, the second object module, The Rating Model of the efficiency value of third object module, the Rating Model of NBA efficiency index formula and Gmsc obtains five rewards and punishments knots Fruit, five rewards and punishments results can be modified by consecutive victories continuous loss, MVP, then can be by five revised rewards and punishments results Average rewards and punishments are as a result, be determined as the final rewards and punishments result of object.

It should be noted that the rewards and punishments that the embodiment finally uses the result is that above-mentioned five rewards and punishments results average value, recognize Practical situation can be more in line with for average value.It, can also be therein any several according to effect selection when actual use A rewards and punishments result come calculate final average value as output rewards and punishments as a result, as long as the rewards and punishments result of final output is more in line with The true horizon of the corresponding object of player, is not specifically limited herein.

As an alternative embodiment, being based on base values and target contribution rate in step S206, determining the first mesh Mark object rewards and punishments result after, this method further include: when preceding switch game be in the deciding grade and level stage in the case where, by rewards and punishments knot The product of fruit and target factor is determined as the target rewards and punishments of first object object as a result, wherein, target factor is with first object pair The increase of the triumph number of fields of elephant or failure number of fields and increase.

In this embodiment, after new hand has created account, the deciding grade and level stage can be entered, the deciding grade and level stage is by ten Match, the corresponding object of the player of capacity variance is subjected to rough differentiation, enable the corresponding object of player ability as early as possible Ground is stabilized to the section of suitable player capacity, to allow player to the ability different from of corresponding object, prevents that there are new hands Lopsided situation influences new hand's game experiencing.

It defines the level the stage in new hand, player capacity branch carries out corresponding rewards and punishments according to the rule being set in advance.Current In the case that one innings of game is in the deciding grade and level stage, by the product of rewards and punishments result and target factor, it is determined as the mesh of first object object Mark rewards and punishments are as a result, the target factor increases with the increase of the triumph number of fields or failure number of fields of first object object.

In an alternate embodiment of the invention, n-th new hand's deciding grade and level Games-time (n≤10) is being calculated, if n-th failure in a contest, that Calculating by the end of n-th failure number of fields is m (0 < m≤10), then corresponding with m multiplied by one on the basis of rewards and punishments result FACTOR P, to obtain final rewards and punishments result.If n-th victory, calculating is by the end of n-th triumph number of fields M ' (0 < m '≤10), then multiplied by a FACTOR P corresponding with m ' on the basis of rewards and punishments result, to obtain final Rewards and punishments result.

For example, the default integral of the corresponding object of every player is all identical, for example, being 3000 points, that is, every object for appreciation The initial ability score of family is all identical.In preceding ten deciding grade and level match of new hand, waited calculating rewards and punishments timesharing, it can be in obtained rewards and punishments Multiplied by a target factor P on the basis of point, wherein this target factor P can pass through the corresponding coefficient table of 1 victory or defeat constant of table It obtains.For example, this match of the corresponding object of the player is the match of third field, this match is defeated, is then calculated Rewards and punishments are divided into -25 points, while by having had two matches altogether defeated to third field, corresponding table 1, and available target factor P is 2.4, thus the corresponding object of the player is divided into -25*2.4=-60 in the rewards and punishments of this competition.In an alternate embodiment of the invention, such as This match of the corresponding object of the fruit player is his the 5th, while this match is won, the rewards and punishments being then calculated It is divided into 30 points, then corresponding table 1, available target factor P is 2.8 by three matches of the 5th his win-win, because And the corresponding object of the player is divided into 30*2.8=84 in the rewards and punishments of this competition.The rewards and punishments of other plays point and above-mentioned calculating side Formula is similar.

The corresponding coefficient table of 1 victory or defeat constant of table

By the end of n-th victory/negative number of fields	1	2	3	4	5	6	7	8	9	10
											P	2	2.4	2.8	3.2	3.6	4	4.4	4.8	5.2	5.6

In the case where initial score is 3000 points, matched by ten deciding grade and level, for example, high-end PlayerlScore will be to 3800 ~4000, low end players score will be to 2600~2800, and intermediate player can be distributed between 2800~3800, due to score There is difference, thus also would not be by the king-sized player matches of capacity gap to together, to make when matching By new hand define the level match after, the ability of the corresponding object of player has got certain differentiation so that every player all into Enter into corresponding ability section.

As an alternative embodiment, step S202, obtains base values packet corresponding with first object probability It includes: in the case where the second battle result is triumph result, obtaining first foundation index corresponding with first object probability, In, first object probability is bigger, and first foundation index is smaller；Second battle result be failure result in the case where, obtain with Corresponding second base values of first object probability, wherein first object probability is bigger, and the second base values is bigger.

The embodiment base values is related with the second battle result, can be by first object probability calculation base values.When First object probability > 0.5, then in preceding switch game, the first game camp ability is stronger, if the second battle result is victory Benefit is as a result, first foundation index can be reduced accordingly, and first object probability is bigger, and first foundation index is with regard to smaller；If the Two battle results are failure, then first foundation index can increase；If Shuai≤0.5 first object Gai, this competition One game camp ability is weaker, should not be too many to the punishment of the first game camp again if the second battle result is failure, Then the second base values is just corresponding is reduced, and first object probability is smaller, and the second base values is fewer, if the first game battle array Battalion wins, then can reward more, the second base values can more greatly, to consider game when determining base values The gap of both sides' strength.

Wherein, M is positive or negative points coefficient.

When the second battle result is that the side A defeats the side B, basis plus-minus score base_scoreAFor M (1-R), when second pair When war result is that the side A is defeated by the side B, basis plus-minus score base_scoreAFor MR.If R > 0.5, this competition A can power compared with By force, it is contemplated that can win, increasing point is just corresponding to be reduced, and R is bigger, and increasing point is fewer；If the side A is defeated, deduction is more；If R < 0.5, Then this competition A can power it is weaker, it is contemplated that can be defeated, deduction is just corresponding to be reduced, and R is smaller, and deduction is fewer, if A team wins, Increase and divides more.

As an alternative embodiment, step S202, obtains base values packet corresponding with first object probability It includes: obtaining basic score corresponding with first object probability, wherein basic score is any in the first game camp for adjusting Object carries out the rewards and punishments score that battle operation obtains；Step S206 is based on base values and target contribution rate, determines first object The rewards and punishments result of object includes: to determine the rewards and punishments score of first object object based on basic score and target contribution rate, wherein In the case where the second battle result is triumph result, increase rewards and punishments score to the raw score of first object object point, the In the case that two battle results are failure result, rewards and punishments score is reduced to the raw score of first object object point.

In this embodiment, base values can be specific numerical value, for example, for basic score base_score, rewards and punishments result It can be rewards and punishments score S, any object in the first game camp is adjusted by basic score and carries out the rewards and punishments that battle operation obtains Basic score and target contribution rate can must be beaten the rewards and punishments score of first object object by score.First object object it is original Score is that first object object accumulates the score got in historical game play, is triumph result in the second battle result In the case of, increase rewards and punishments score to the raw score of first object object point, the case where the second battle result is failure result Under, rewards and punishments score is reduced to the raw score of first object object point.

In the related art, the capacity variance of player is determined by artificial rule, to target object in battle class trip Differentiation is not added in rewards and punishments in play, which can provide according to the different manifestations of player in one innings of game and meet player's performance Or the rewards and punishments of ability, to distinguish every player in the contribution rate of local exchange, enable different abilities the corresponding object of player more It distinguishes fastly, makes match fiercer, player experience is more preferable.It is different to distinguish performance compared to artificial rule for this method The corresponding object of player, provide the rewards and punishments for meeting player's performance as a result, making the ability of the corresponding object of player point can be faster The ability level stage for being stabilized to oneself in；Compared to ELO rating model, this method is instructed by the historical behavior data of player Practice deep learning model, can more accurately predict the winning rate of opposing teams, can also distinguish and determine the different player of performance to most The contribution rate of battle result eventually, it is more rationally, more fair, so that the competition of every innings of game is more fierce, the participation of player Feel stronger, game experiencing is more preferable.

Technical solution of the present invention is illustrated below with reference to preferred embodiment, specifically using base values as rewards and punishments Score, rewards and punishments result are that rewards and punishments score is illustrated.

In sports game, the online online competitive game of more people, Massively Multiplayer Online Role Playing Games, all exist Vying each other between player and cooperate, since the ability of the game operation of player, consciousness etc. is different, the corresponding object of player Performance is also just different, and ability scoring is just desirable to through the corresponding object of different players in corresponding playing method or bout Performance, to identify the capacity variance of the corresponding object of player, for example, in the match of basketball game, different players couple The ability for the object answered is different, then triumphantly in side the corresponding object of all players be also for the contribution of triumph it is different, Conversely, the losing side is similarly, based on different contribution rates, can allow the corresponding object of player obtain the reward for meeting their performances or Person's punishment.

The embodiment can by history a large amount of in game compete and player history shot and long term behavior representation data come Winning rate prediction model is established, is predicted by the winning rate that winning rate prediction model may be matched both sides, to obtain two side's strength The assessment of gap.The embodiment can be determined in game according to the winning rate of prediction to the basic score of both sides player plus-minus.? After determining basis score, can according to the performance of the corresponding object of player in gaming, using statistical data and behavior sequence, Building covers behavior scoring model.It is given using behavior scoring model according to the performance of the corresponding object of different players in gaming Meet the contribution rate d of player's performance or ability out_i, d_iFor indicating the contribution of the corresponding object of i-th of player in local exchange game Rate, the contribution rate namely positive or negative points ratio.After obtaining positive or negative points ratio and basic score, so that it may this ratio be calculated Match the rewards and punishments score of this player.

After the positive or negative points by behavior scoring model, positive or negative points can also be carried out according to consecutive victories even negative and MVP etc. Certain amendment obtains the positive or negative points of final each player, then adjusts the ability point of each player.

The embodiment uses the above method, can solve game capabilities scoring positive or negative points for triumph side/the losing side player It does not distinguish, leads to the unreasonable inequitable problem of positive or negative points, it can be according to the difference of player in one innings of game by this method Performance provides and meets player's performance or the scoring of ability, to distinguish tribute of the corresponding object of every player in local exchange game Degree is offered, the player of different abilities is distinguished faster, to make match fiercer, player experience is more preferable.

Fig. 3 is a kind of stream of the method for ability scoring based on player's multi-mode behavior expression according to an embodiment of the present invention Cheng Tu.As shown in figure 3, this method comprises: deciding grade and level the stage and increase stage by stage.Wherein, the deciding grade and level stage includes creation account and new hand Preceding ten deciding grade and level matches；Increase includes the shot and long term behavior portrait based on the corresponding object of player stage by stage, and training winning rate estimates mould Type estimates the winning rate of both sides by winning rate prediction model, according to the winning rate of both sides, can roughly determine the difference of both sides' strength Away from, according to the gap of strength formulate this innings match victory or defeat basic positive or negative points；After end of match, not according to authorities' game The performance of corresponding object with player obtains the corresponding object of each player in the contribution of this innings by behavior expression Rating Model Then rate carries out difference positive or negative points to the corresponding object of every player according to contribution rate, that is, doing very well on basic score Add point or decrease deduction, show the few bonus point or more deductions of difference.Simultaneously in view of the corresponding ability point of some consecutive victories continuous loss is repaired The MVP of positive rule and final hit key ball is regular, and we are modified and finely tune to ability point, and it is corresponding right to finally obtain player The ability rewards and punishments score of elephant.

It defines the level to match to the new hand of first stage below and be introduced.

After new hand has created account, next ten initially enter the deciding grade and level stage than grand sports meet, and the deciding grade and level stage is logical This ten matches are crossed, are first roughly distinguished the corresponding object of the player of capacity variance, so that player capacity point can be to the greatest extent It is stabilized to the score section of suitable player capacity fastly, allows the ability score different from of the corresponding object of player, prevents from existing new The lopsided situation of hand, influences new hand's game experiencing.

It defines the level the stage in new hand, player capacity branch is added and subtracted accordingly according to the rule being set in advance, Ke Yishe The default integral of fixed every player is all identical, for example, being 3000 points, that is, the initial ability score of every player is 3000 points.

The positive or negative points of new hand's deciding grade and level match are one coefficients of superposition on the basis of the every rewards and punishments score calculated originally P, this coefficient increase with victory number of fields/negative number of fields increase, the corresponding coefficient such as table 1 of specific victory or defeat constant.

For example, defining the level Games-time (n≤10) calculating n-th new hand, if n-th failure in a contest, calculates and cut It is only m (0 < m≤10) to n-th failure number of fields, then multiplied by the corresponding system of a m on the basis of final rewards and punishments score Number P, to obtain final deduction score.If n-th victory, similarly available final bonus point score.

It, can be in obtained rewards and punishments point when calculating rewards and punishments score again for example, in preceding ten deciding grade and level match of new hand On the basis of multiplied by a FACTOR P.The third field for example, this of the corresponding object of the player competes, this match is defeated, then counts Obtained rewards and punishments are divided into -25 points, while by having had two matches altogether defeated to third field, corresponding table 1, available coefficient P is 2.4, thus the corresponding object of the player is divided into -25*2.4=-60 in the rewards and punishments of this competition.If the player is corresponding This match of object is the 5th, while this match is won, and the rewards and punishments being then calculated are divided into 30 points, then by arriving 5th win-win, three matches, corresponding table 1, available FACTOR P is 2.8, thus the rewards and punishments of player's this competition It is divided into 30*2.8=84.Other calculations are similar, no longer illustrate one by one herein.

After being matched by ten deciding grade and level, for example, initial score is 3000 points, then high-end PlayerlScore will to 3800~ 4000, low end players score will be to 2600~2800, and intermediate player can be distributed between 2800~3800, since score goes out Difference is showed, so also would not be the corresponding object matching of the king-sized player of capacity gap to together when matching ?.

The increasing of stablizing of the embodiment is introduced stage by stage below.

After match of defining the level by new hand, certain differentiation, every object for appreciation have been got to the ability of the corresponding object of player The score of the corresponding object of family has entered in corresponding ability section, stage by stage can be according to the corresponding object of player stablizing increasing Every competition game shows and as a result, the contribution rate of object corresponding with player is determined, so that the ability of player point more accords with The real ability for closing player is horizontal.

The embodiment is obtained by basic score by winning rate prediction model is introduced below.

In this embodiment, the winning rate R of competition game both sides is predicted by winning rate prediction model, wherein R is for indicating The side A can then obtain a true result of the match, in this way according to model prediction to the winning rate of the side B after competition game terminates The actual result of both sides' winning rate and match, the basic score of available game both sides, for example, the basic score of the side A calculates such as Under:The basic score of the side B can similarly be calculated, wherein M is system Coefficient number is divided on given basis in advance, for adjusting the size of basic score.If R > 0.5, the ability of this competition side A compared with By force, it is contemplated that can win, increasing point is just corresponding to be reduced, and R is bigger, and increasing point is fewer, if the side A is defeated, deduction is more；If R < 0.5, Then this competition A can power it is weaker, it is contemplated that can be defeated, deduction is just corresponding to be reduced, and R is smaller, and deduction is fewer, if the side A wins, Increase and divides more.

Due in authorities' competition game, the performance situation of different players is different, and in triumph side, bonus point and the losing side deduction be not Should be identical, that is, do very well in authorities' game should bonus point it is more or deduction is few, show difference should bonus point it is few or subtract Divide more.

The embodiment is about player's behavior expression model, including three kinds of different models and two different indexs.

It is introduced below to based on winning rate prediction model.

The embodiment predicts mould by history result of the match and the history of player match representation data, one winning rate of training Type.

Fig. 4 is a kind of schematic diagram of winning rate prediction model according to an embodiment of the present invention.As shown in figure 4, with the basketball of 3V3 For, there are six objects, wherein the corresponding object of current player including contribution rate to be calculated, two teammates with the player Corresponding object, object corresponding with three opponents of the player, their feature vector is all history shot and long term representation data Composition, so each object will correspond to a vector, then sums it up and makes even using by the vector as the two of teammate objects , vector a is obtained, will sum it up and be averaged as the three of opponent objects, vector b is obtained, then by first object object Vector sum vector a, vector b are stitched together, and obtain vector c, for example, if the original vector length of the first object object is 20 It ties up, then the length of vector a, vector b are also all 20 dimensions, then the length of vector c is 3*20=60 dimension, then length is tieed up for 60 Vector be input in a deep neural network model, the output of deep neural network model is exactly that the first object object exists Win-or-lose result in competition game, when first object object is won in competition game, then the output of models fitting is exactly 1, instead Then be 0, then by training can be obtained by a winning rate prediction model.

In training, use the history representation data of opponent, teammate's (comprising oneself) as input, output is the knot of match Fruit, if the side A has won the side B, export result be 1, on the contrary it is then export result be 0.

In forecast period, the performance data that the authorities of opponent, teammate's (including oneself) are competed are as trained winning rate The input of prediction model, trained winning rate prediction model can be obtained by the value of a 0-1, this value can be used to the side of expression To the winning rate size of the side B.

The embodiment can obtain each according to the corresponding object of each player when the behavior expression in preceding switch game The statistical data of the corresponding object of player in gaming, for example, the data such as score, secondary attack, backboard, it is believed that one is done very well Player where the winning rate in camp be typically all to want high, thus when determining the contribution rate of each player, by constantly counting Winning rate of the corresponding object of this player in Different matching mode is calculated to calculate.

For example, one shared marked as 1-6, the corresponding object of totally six players participates in match, this innings of ratio in this innings of game The teammate and opponent of the corresponding object of No. 1 player of match have determined that, the statistical data and result of the match of every player is also It determines, when calculating the contribution rate of the corresponding object of No. 1 player, two objects can be selected from remaining five objects at random Object corresponding with No. 1 player forms teammate, opponent of remaining three objects as the corresponding object of No. 1 player, by them Statistical data be input in winning rate prediction model, so that it may obtain the winning rate under various imaginary matching ways, due to altogether HaveKind selects mode, so object corresponding for No. 1 player, can be calculatedKind winning rate, then by thisKind victory Rate takes the average contribution rate as No. 1 player, similarly can be obtained by the performance point of the corresponding object of other players.

The Rating Model based on figure is introduced below.

In this embodiment, it is contemplated that the interaction problems in team project are the most suitable using graph model.Pass through one The behavior sequence of the corresponding object of player can establish graph model in the match of field, and the interaction between the corresponding object of player is (main If pass behavior) it shows.In an alternate embodiment of the invention, using three objects for the side that competes as the section in graph model Point, when there is pass behavior between the corresponding object of two players, so that it may connect between the corresponding object of two players Connect a line, and this edge be it is oriented, the direction on the side illustrates the transmission direction of ball, in this way can structure The player's interaction graph model for building bout.It, can be according to the center (flow of stream after the completion of player's interaction figure model foundation Centrality) this index goes the significance level for calculating three nodes in the graph model.

Fig. 5 is a kind of schematic diagram of Rating Model based on figure according to an embodiment of the present invention.As shown in figure 5, each object for appreciation The contribution rate of the corresponding object of family is r_C=num (d^*)/num (d), for calculating the contribution rate of centre forward player, wherein num (d^*) for indicating item number of PF, PG point to the side for passing through C point in all paths, num (d) is for indicating PF, PG point to all The item number on path side.If C is all passed through in most of paths in all paths of PF, PG point pair it can be seen from the index, that Prove that C (centre forward) object is important, because the power flowing of PF, PG ball will pass through C.Similarly other players of available our team Contribution rate.

The Rating Model based on cluster label of the embodiment is introduced below.

There can be different occupations in game, the evaluation criterion of different occupation is different, for example, centre forward, which is mainly responsible for, robs basket Plate, score would not be especially more, and shooting guard is mainly responsible for three points of throwing, and score will mostly etc..Thus specifically scoring When should go to score according to the standard of different occupation, in this way can be more fair and just.

In order to allow scoring to be more in line with subjective assessment, which can allow experienced player or planning to go to one The performance marking of player, is equivalent to obtain expertise in this way, then recycling has monitor model to learn this in match Expertise finally can have monitor model to realize scoring using this.But the ratio for having hundreds of thousands even millions of in game Match wants one one to go to give a mark, and such workload is too big, therefore in the way of clustering, and similar player is corresponding right It as gathering for one kind, then gives a mark again to this class, the behavior of the corresponding object of of a sort player itself tends to similar Property, the score of acquisition should also be similar, be considerably reduced the workload of planning label in this way, while be also possible that mark Remember more accurate.

The embodiment distinguishes occupation, professional big of available C, SG, SF, PG, PF five to all history competition datas Measure the history race statistics data of player.Then in each occupation, the utilization race statistics data of the occupation use Kmeans Clustering method carries out clustering processing to history race statistics data, and each occupation cluster can be 20 classes.

In cluster result, each occupation is divided into a variety of different types, for example, having to obtain parting in C occupation C, there are defensive C, also backboard type C, to arrive the corresponding object of the similar player of behavior, while C occupation is divided into 20 different grades, that is, the performance, ability between inhomogeneity are different, it should obtain different scorings.

The embodiment uses central point (average values of all data in such) Lai Daibiao category of each class.Then will The center point data of 20 classes of each occupation gives planning, and planning can be according to their subjectivity to 20 classes in each occupation It is ranked up, and is labeled as 1-20.After label is completed, the label of each class is just obtained, nerve net then can be utilized Network is fitted this label marking, and Lai Xunlian obtains the neural network Rating Model of each occupation.

After bout terminates, timesharing is showed calculating, the occupation first used according to the player selects corresponding mind Through network Rating Model, then using the race statistics data of the player as the input of neural network Rating Model, so that it may To the table contribution rate of the player, the similarly contribution rate of the corresponding object of other players of available this competition.

The Rating Model based on NBA efficiency index formula is introduced below.

The embodiment can use existing capacity index, and player directly is calculated using the statistical data that player competes Contribution rate, NBA efficiency index formula can be used, specific formula for calculation is as follows:

The Rating Model of the efficiency value based on Gmsc is introduced below.

The embodiment is as follows using the specific calculation formula of efficiency value of Gmsc:

The contribution rate that above-mentioned five kinds of Rating Models obtain also is not the final contribution rate of final object, in order to guarantee to win The sum of performance contribution rate of the corresponding object of all players of side/the losing side is 1, is needed to triumph side/the losing side player performance Divide and be normalized, calculation formula is as follows:r_iIndicate the contribution of the corresponding object of calculated i player Rate, d_iIt is the contribution rate after the corresponding normalized of i player,It is to own in camp where the corresponding object of i player The sum of the contribution rate of the corresponding object of player, to ensure that

It can be to the corresponding object of each player in the enterprising of basic score by the different contribution rates of each player Row difference positive or negative points,Wherein S_iFor indicating the The rewards and punishments of i player are as a result, base_ScoreAFor indicating the base values in the camp A, base_ScoreBFor indicating the basis in the camp B Index, if the corresponding object of i-th of player belongs to the camp A, the base values base in the camp A_ScoreAOn multiplied by i-th The contribution rate d of the corresponding object of player_i, obtain the rewards and punishments result S of i-th of player_i；If the corresponding object of i-th of player belongs to The camp B, then in the base values base in the camp B_ScoreBOn multiplied by the corresponding object of i-th of player contribution rate d_i, obtain i-th The rewards and punishments result S of a player_i。

Being modified to consecutive victories continuous loss situation to score by rule for the embodiment is described below.

At corresponding object consecutive victories play number >=3 of player, corresponding amendment is done to rewards and punishments score in end of match.With Family rewards and punishments point=S_i+ (N-2) * m, wherein S_iIt is by the calculated rewards and punishments of front model point, N is for indicating that consecutive victories play, m are used In indicating consecutive victories modified basis point, self-defining can be carried out.

When corresponding object continuous loss play number >=3 of player, corresponding amendment is done to rewards and punishments score in end of match.With Family rewards and punishments point=S_i+ (M-2) * h, wherein S_iIt is by the calculated rewards and punishments of front model point, M is for indicating user's continuous loss field Secondary, h can carry out self-defining for indicating continuous loss modified basis point.

Being finely adjusted by MVP rule to score for the embodiment is described below.

The embodiment is setting MVP assessment rules during the games, captures final hit ball, key ball in Basketball Match (crucial score, crucial backboard, key grab, crucial nut cap, crucial interference, crucial secondary attack etc.), continuous ball (are run, even Continuous backboard is continuously grabbed, continuous nut cap, stepwise derivation, continuously secondary attack etc.), and according to the behavior at these MVP bloom moment to object for appreciation Family's ability scoring is finely adjusted.When this player obtains MVP, the rewards and punishments for correcting this player are divided into=S_i+mvp_score, In, S_iIt is by the calculated rewards and punishments of training pattern as a result, mvp_scoreIt is for correcting the normal of the reward score of acquisition MVP player Number, can carry out self-defining.

By above-mentioned steps, the corresponding object of each player by trained winning rate prediction model, based on the scoring of figure Model, based on cluster label Rating Model, NBA efficiency index formula Rating Model and Gmsc efficiency value Rating Model Five rewards and punishments scores are obtained, which can be modified by consecutive victories continuous loss, MVP, then can correct five The average rewards and punishments score of rewards and punishments result afterwards is determined as the final rewards and punishments score of object.If troop where player's object is defeated , calculated rewards and punishments score is just subtracted, if instead having won just plus calculated rewards and punishments score.

The embodiment proposes the scoring solution of the game capabilities based on player's multi-mode behavior expression.The program is first Shot and long term behavior portrait based on player, the winning rate of both sides is estimated by training deep neural network model, according to both sides' Winning rate can predict both sides' capabilities gap, and then the basic score of this innings of match victory or defeat is formulated according to the gap of strength.It is competing After end, behavior expression Rating Model can be passed through according to the performance when the corresponding object of players different in preceding switch game Each player is obtained to the contribution rate of the battle result of this innings of game, is then existed according to contribution rate to the corresponding object of every player Distinguishing positive or negative points are carried out on basic score, that is, the object to do very well can add point or decrease deduction, show difference Object can lack bonus point or more deductions.The embodiment be additionally contemplates that simultaneously some consecutive victories suffer successive lost corresponding ability divide modification rule with And the MVP rule of final hit key ball is modified and finely tunes to rewards and punishments score, finally obtains the rewards and punishments score for meeting player's performance.

The embodiment is according to the performance of competitive game both sides capabilities gap and the corresponding object of each player of authorities' game And ability is the scoring that the corresponding object of player carries out post-games, is tied to realize that the corresponding object of different players is competed in authorities Reasonable rewards and punishments score is obtained after beam, so that the ability of the corresponding object of player point can quickly be stabilized to the energy for meeting oneself In the score section of power level, so that the competition of every innings of game is more fierce, the sense of participation of player is stronger, and game experiencing is more preferable.

This method can be distinguished compared to artificial rule and show different players, provide the positive or negative points for meeting player's performance, make The ability point for obtaining player can be stabilized to faster in the horizontal segment for meeting player capacity；Compared to ELO model, this method passes through depth Learning model and player's shot and long term portrait are spent, can more accurately predict that the winning rate of opposing teams, same this method can be treated with a certain discrimination Different players is showed, different performance points is provided, it is more rationally, more fair.

It should be noted that step shown in the flowchart of the accompanying drawings can be in such as a group of computer-executable instructions It is executed in computer system, although also, logical order is shown in flow charts, and it in some cases, can be with not The sequence being same as herein executes shown or described step.

The embodiment of the invention also provides a kind of game data processing units.It should be noted that the game of the embodiment Data processing equipment can be used for executing the game data processing method of the embodiment of the present invention.

Fig. 6 is a kind of schematic diagram of game data processing unit according to an embodiment of the present invention.As shown in fig. 6, the game Data processing equipment 600 includes: processing unit 10, the first determination unit 20 and the second determination unit 30.

Processing unit 10 obtains the first battle for predicting that the first game camp and the second game camp carry out battle operation As a result first object probability, and obtain base values corresponding with first object probability, wherein base values is for adjusting Any object carries out the rewards and punishments result that battle operation obtains in first game camp.

First determination unit 20, for according to first object object when carried out in preceding switch game battle operation when generation Goal behavior data, determine target contribution rate, wherein first object object be the first game camp in any object, mesh Mark contribution rate is the ratio that first object object is the contribution that the second battle result is made, and the second battle result is the first game battle array Battalion carries out the result that battle operation obtains with the second game camp in working as preceding switch game.

Second determination unit 30 determines the rewards and punishments knot of first object object for being based on base values and target contribution rate Fruit.

In an alternate embodiment of the invention, the device further include: first acquisition unit, for predicting the first game camp and the Two game camps carry out battle operation obtain the first object probability of the first battle result before, obtain the first game camp going through The first historical behavior data of battle operation generation are carried out in history game, the second game camp carries out battle behaviour in historical game play Make the second historical behavior data generated and the first game camp and the second game camp carries out battle operation in historical game play The history of generation fights result；First training unit, for passing through the first historical behavior data, the second historical behavior data and going through History battle structure is trained first object model；Processing unit 10 includes: the first acquisition module, for obtaining the first game Camp is when the first current behavior data and the second game camp for carrying out battle operation generation in preceding switch game are when previous The second current behavior data that battle operation generates are carried out in office's game；First input module is used for the first current behavior number It is input to trained first object model according to the second current behavior data, the first object for obtaining the first battle result is general Rate.

In an alternate embodiment of the invention, the first determination unit 20 includes: the second acquisition module, for obtaining the first game camp In multiple first objects and the second game camp in multiple second objects, wherein multiple first objects include first object Object；Selecting module, for randomly choosing from multiple first objects and multiple second objects in addition to first object object Second target object of first object quantity out, and the second target object of first object quantity is matched as first object object Teammate, by the third mesh in multiple first objects and multiple second objects in addition to first object object and the second target object Object is marked, is matched as the opponent of first object object；Second input module, for by the goal behavior number of first object object Trained first mesh is input to according to the goal behavior data of, goal behavior data of the second target object and third target object It marks in model, obtains the second destination probability of the first battle result；First determining module, for the second destination probability to be determined as Target contribution rate.

In an alternate embodiment of the invention, the first determination unit 20 further include: be randomly selected the second of first object quantity After target object, the second determining module, for the multiple of determining the second target object that first object quantity is randomly selected Objective result；Third obtains module, for obtaining the second destination probability, obtaining the second destination number under every kind of objective result The second destination probability, wherein the second destination number can be multiple objective results quantity；First determining module includes: true Stator modules obtain target contribution rate for the second destination probability of the second destination number to be averaged.

In an alternate embodiment of the invention, the device further include: second acquisition unit, for working as according to first object object The goal behavior data generated when battle operation are carried out in preceding switch game, before determining target contribution rate, obtain the first game Camp is in the interbehavior data generate when battle operation, wherein goal behavior data include interbehavior data；It establishes Unit, for establishing the second object module based on interbehavior data, wherein the node in the second object module is used to indicate the Object in one game camp, the path in the second object module are used to indicate between object corresponding with the node on path Interbehavior is showed；First determination unit 20 includes: the 4th acquisition module, for obtaining the first mesh in the second object module Mark the third destination number in path and the 4th destination number of the second destination path, wherein first object path is passed through and first The corresponding first node of target object and the second section corresponding with the object in the first game camp in addition to first object object Point, the second destination path pass through second node；Third determining module, for by the ratio of third destination number and the 4th destination number Value, is determined as target contribution rate.

In an alternate embodiment of the invention, the device further include: third acquiring unit, for working as according to first object object The goal behavior data generated when battle operation are carried out in preceding switch game, before determining target contribution rate, obtain the first game Camp and the second game camp carry out the historical behavior data that battle operation generates in historical game play；Cluster cell, for pair Historical behavior data carry out clustering processing, obtain the sub- historical behavior data of multiple types；4th acquiring unit is more for obtaining The corresponding contribution to the history of rate of sub- historical behavior data of a type, wherein contribution to the history of rate is the sub- historical behavior of each type Data are the ratio for the contribution that the second battle result is made；Second training unit, for passing through multiple types and contribution to the history of rate Third object module is trained；First determination unit 20 includes: the 4th determining module, for determining goal behavior data Target type；Third input module obtains target contribution rate for target type to be inputted trained third object module.

In an alternate embodiment of the invention, the first determination unit 20 includes: the 5th acquisition module, for obtaining the first game camp In multiple objects contribution rate, wherein the contribution rate of each object is that each object is the contribution that the second battle result is made Ratio；Processing module is normalized target contribution rate for the contribution rate by multiple objects, after obtaining processing Target contribution rate.

In an alternate embodiment of the invention, the device further include: amending unit, for being based on base values and target contribution Rate, after the rewards and punishments result for determining first object object, in the case where goal behavior data fit goal condition, by with mesh The corresponding correction value of mark condition corrects rewards and punishments result.

In an alternate embodiment of the invention, which further includes at least one of: third determination unit, for by with mesh Before marking behavioral data corresponding correction value amendment rewards and punishments result, first object object is set to be continuously available the in goal behavior data In the case that the number of two battle results is more than or equal to targeted number, goal behavior data fit goal condition is determined；4th really Order member, for indicating in the case where there is goal behavior in preceding switch game in goal behavior data, determines target line For data fit goal condition.

In an alternate embodiment of the invention, the device further include: the 4th determination unit, for being based on base values and target tribute Offer rate, after the rewards and punishments result for determining first object object, when preceding switch game be in the deciding grade and level stage in the case where, by rewards and punishments As a result with the product of target factor, it is determined as the target rewards and punishments of first object object as a result, wherein, target factor is with first object The increase of the triumph number of fields of object or failure number of fields and increase.

In an alternate embodiment of the invention, processing unit 10 includes: the 6th acquisition module, for being triumph in the second battle result As a result in the case where, first foundation index corresponding with first object probability is obtained, wherein first object probability is bigger, the One base values is smaller；7th obtains module, for obtaining and the first mesh in the case where the second battle result is failure result Mark corresponding second base values of probability, wherein first object probability is bigger, and the second base values is bigger.

In an alternate embodiment of the invention, processing unit 10 includes: the 8th acquisition module, for obtaining and first object probability phase Corresponding basis score, wherein basic score is used to adjust any object in the first game camp and carries out what battle operation obtained Rewards and punishments score；Second determination unit 30 includes: to determine the rewards and punishments point of first object object based on basic score and target contribution rate Number, wherein in the case where the second battle result is triumph result, increase rewards and punishments point to the raw score of first object object point Number reduces rewards and punishments score to the raw score of first object object point in the case where the second battle result is failure result.

The embodiments of the present invention also provide a kind of storage medium, computer program is stored in the storage medium, wherein The computer program is arranged to execute the step in any of the above-described embodiment of the method when operation.

In an alternate embodiment of the invention, in the present embodiment, above-mentioned storage medium can include but is not limited to: USB flash disk read-only is deposited Reservoir (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), the various media that can store computer program such as mobile hard disk, magnetic or disk.

The embodiments of the present invention also provide a kind of electronic device, including memory and processor, stored in the memory There is computer program, which is arranged to run computer program to execute the step in any of the above-described embodiment of the method Suddenly.

In an alternate embodiment of the invention, above-mentioned electronic device can also include transmission device and input-output equipment, wherein The transmission device is connected with above-mentioned processor, which connects with above-mentioned processor.

Obviously, those skilled in the art should be understood that each module of the above invention or each step can be with general Computing device realize that they can be concentrated on a single computing device, or be distributed in multiple computing devices and formed Network on, in an alternate embodiment of the invention, they can be realized with the program code that computing device can perform, it is thus possible to It is stored in storage device and is performed by computing device, and in some cases, it can be to be different from herein suitable Sequence executes shown or described step, and perhaps they are fabricated to each integrated circuit modules or will be in them Multiple modules or step are fabricated to single integrated circuit module to realize.In this way, the present invention is not limited to any specific hardware It is combined with software.

The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field For art personnel, the invention may be variously modified and varied.It is all within principle of the invention, it is made it is any modification, etc. With replacement, improvement etc., should all be included in the protection scope of the present invention.

Claims

1. a kind of game data processing method characterized by comprising

Predict that the first game camp and the second game camp carry out battle operation and obtain the first object probability of the first battle result, And obtain base values corresponding with the first object probability, wherein the base values is for adjusting first trip Any object carries out the rewards and punishments result that the battle operation obtains in play camp；

According to first object object when the goal behavior data for carrying out generating when the battle operation in preceding switch game, determine Target contribution rate, wherein the first object object is any object in first game camp, the target contribution rate It is the ratio for the contribution that the second battle result is made for the first object object, the second battle result is first trip Play camp and second game camp are in the result for working as and carrying out the battle operation in preceding switch game and obtaining；

Based on the base values and the target contribution rate, the rewards and punishments result of the first object object is determined.

2. the method according to claim 1, wherein the case where the second battle result is triumph result Under, the first object object is rewarded according to the rewards and punishments result, is failure result in the second battle result In the case of, the first object object is punished according to the rewards and punishments result.

3. the method according to claim 1, wherein

First battle is obtained determining that first game camp and the progress battle of second game camp operate As a result before the first object probability, the method also includes: obtain first game camp in historical game play into Described in first historical behavior data of the row battle operation generation, second game camp carry out in the historical game play Second historical behavior data of battle operation generation and first game camp and second game camp are in the history The history battle result that the battle operation generates is carried out in game；It is gone through by the first historical behavior data, described second History behavioral data and history battle structure are trained first object model；

Determine that first game camp and second game camp carry out the battle operation and obtain the first battle knot The first object probability of fruit includes: to obtain first game camp to carry out the battle in the preceding switch game described It operates the first current behavior data generated and second game camp and carries out the battle in the preceding switch game described Operate the second current behavior data generated；The first current behavior data and the second current behavior data are input to The trained first object model obtains the first object probability of the first battle result.

4. according to the method described in claim 3, it is characterized in that, according to the first object object described when preceding switch is swum The goal behavior data generated when the battle operation are carried out in play, determine that the target contribution rate includes:

Multiple second objects in multiple first objects and second game camp in first game camp are obtained, In, the multiple first object includes the first object object；

From the multiple first object and the multiple second object in addition to the first object object, it is randomly selected Second target object of first object quantity, and it is described that second target object of the first object quantity, which is matched, The teammate of one target object will remove the first object object and institute in the multiple first object and the multiple second object The third target object except the second target object is stated, is matched as the opponent of the first object object；

By the goal behavior data of the first object object, the goal behavior data of second target object and The goal behavior data of the third target object are input in the trained first object model, obtain described Second destination probability of one battle result；

Second destination probability is determined as the target contribution rate.

5. according to the method described in claim 4, it is characterized in that,

After second target object of the first object quantity is randomly selected, the method also includes: determine with Machine selects multiple objective results of second target object of the first object quantity；In every kind of objective result Under, second destination probability is obtained, obtains second destination probability of the second destination number, wherein second target Quantity is the quantity of the multiple objective result；

It includes: by second mesh of second destination number that second destination probability, which is determined as the target contribution rate, Mark probability is averaged, and obtains the target contribution rate.

6. the method according to claim 1, wherein

According to the first object object in the mesh that generates when carrying out the battle in preceding switch game and operating Behavioral data is marked, before determining the target contribution rate, the method also includes: it obtains first game camp and is carrying out institute State the interbehavior data generated when battle operation, wherein the goal behavior data include the interbehavior data；It is based on The interbehavior data establish the second object module, wherein the node in second object module is used to indicate described Object in one game camp, it is corresponding with the node on the path right that the path in second object module is used to indicate Occurs interbehavior as between；

According to the first object object described when the target for carrying out generating when the battle operation in preceding switch game Behavioral data determines that the target contribution rate includes: to obtain the third mesh in first object path in second object module Mark the 4th destination number of quantity and the second destination path, wherein the first object path is passed through and the first object pair As corresponding first node and with the object corresponding second in first game camp in addition to the first object object Node, second destination path pass through the second node；

By the ratio of the third destination number and the 4th destination number, it is determined as the target contribution rate.

7. the method according to claim 1, wherein

According to the first object object in the mesh that generates when carrying out the battle in preceding switch game and operating Behavioral data is marked, before determining the target contribution rate, the method also includes: obtain first game camp and described the Two game camps carry out the historical behavior data that the battle operation generates in historical game play；To the historical behavior data into Row clustering processing obtains the sub- historical behavior data of multiple types；Obtain the sub- historical behavior data of the multiple type Corresponding contribution to the history of rate, wherein the contribution to the history of rate is that the sub- historical behavior data of each type are described The ratio for the contribution that second battle result is made；By the multiple type and the contribution to the history of rate to third object module into Row training；

According to the first object object described when the target for carrying out generating when the battle operation in preceding switch game Behavioral data determines that the target contribution rate comprises determining that the target type of the goal behavior data；By the target type The trained third object module is inputted, the target contribution rate is obtained.

8. method as claimed in any of claims 1 to 7, which is characterized in that exist according to the first object object The goal behavior data generated when carrying out the battle in preceding switch game and operating, determine the target contribution rate Include:

Obtain the contribution rate of multiple objects in first game camp, wherein the contribution rate of each object is each The object is the ratio for the contribution that the second battle result is made；

The target contribution rate is normalized by the contribution rate of the multiple object, the mesh that obtains that treated Mark contribution rate.

9. method as claimed in any of claims 1 to 7, which is characterized in that be based on the base values and institute Target contribution rate is stated, after the rewards and punishments result for determining the first object object, the method also includes:

In the case where the goal behavior data fit goal condition, corrected by correction value corresponding with the goal condition The rewards and punishments result.

10. according to the method described in claim 9, it is characterized in that, passing through amendment corresponding with the goal behavior data Before value corrects the rewards and punishments result, the method also includes at least one of:

The goal behavior data make the first object object be continuously available it is described second battle result number be greater than etc. In the case where targeted number, goal condition described in the goal behavior data fit is determined；

It is indicated in the goal behavior data described in the case where there is goal behavior in preceding switch game, determines the mesh Mark behavioral data meets the goal condition.

11. method as claimed in any of claims 1 to 7, which is characterized in that be based on the base values and institute Target contribution rate is stated, after the rewards and punishments result for determining the first object object, the method also includes:

Described in the case that preceding switch game is in the deciding grade and level stage, the product of the rewards and punishments result and target factor determines For the first object object target rewards and punishments as a result, wherein, the target factor with the first object object triumph Number of fields or failure number of fields increase and increase.

12. method as claimed in any of claims 1 to 7, which is characterized in that obtain and the first object probability Corresponding base values includes:

In the case where the second battle result is triumph result, the first base corresponding with the first object probability is obtained Plinth index, wherein the first object probability is bigger, and the first foundation index is smaller；

In the case where the second battle result is failure result, the second base corresponding with the first object probability is obtained Plinth index, wherein the first object probability is bigger, and second base values is bigger.

13. method as claimed in any of claims 1 to 7, which is characterized in that

Obtaining the base values corresponding with the first object probability includes: that acquisition is opposite with the first object probability The basic score answered, wherein the basis score carries out the battle for adjusting any object in first game camp Operate obtained rewards and punishments score；

Based on the base values and the target contribution rate, determine that the rewards and punishments result of the first object object includes: to be based on The basis score and the target contribution rate, determine the rewards and punishments score of the first object object, wherein at described second pair In the case that result of fighting is triumph result, the raw score point of Xiang Suoshu first object object increases the rewards and punishments score, in institute In the case where the second battle result is stated as failure result, the raw score point of Xiang Suoshu first object object reduces the rewards and punishments point Number.

14. a kind of game data processing unit characterized by comprising

Processing unit obtains the first battle result for predicting that the first game camp and the second game camp carry out battle operation First object probability, and obtain base values corresponding with the first object probability, wherein the base values is for adjusting Any object carries out the rewards and punishments result that the battle operation obtains in whole first game camp；

First determination unit, for according to first object object when carrying out in preceding switch game generating when the battle operates Goal behavior data determine target contribution rate, wherein the first object object is any right in first game camp As, the target contribution rate is the ratio that the first object object is the contribution that the second battle result is made, described second pair War result is first game camp and second game camp described when carrying out the battle behaviour in preceding switch game Make obtained result；

Second determination unit determines the first object object for being based on the base values and the target contribution rate Rewards and punishments result, wherein in the case where the second battle result is triumph result, to the first object object according to described Rewards and punishments result is rewarded, it is described second battle result be failure result in the case where, to the first object object according to The rewards and punishments result is punished.

15. a kind of storage medium, which is characterized in that be stored with computer program in the storage medium, wherein the computer Program is arranged to execute method described in any one of claim 1 to 13 when operation.

16. a kind of electronic device, including memory and processor, which is characterized in that be stored with computer journey in the memory Sequence, the processor are arranged to run the computer program to execute described in any one of claim 1 to 13 Method.