CN112402982B - User cheating behavior detection method and system based on machine learning - Google Patents
User cheating behavior detection method and system based on machine learning Download PDFInfo
- Publication number
- CN112402982B CN112402982B CN202010090357.9A CN202010090357A CN112402982B CN 112402982 B CN112402982 B CN 112402982B CN 202010090357 A CN202010090357 A CN 202010090357A CN 112402982 B CN112402982 B CN 112402982B
- Authority
- CN
- China
- Prior art keywords
- combat
- target
- data
- game
- character
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/70—Game security or game management aspects
- A63F13/75—Enforcing rules, e.g. detecting foul play or generating lists of cheating players
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F2300/00—Features of games using an electronically generated display having two or more dimensions, e.g. on a television screen, showing representations related to the game
- A63F2300/50—Features of games using an electronically generated display having two or more dimensions, e.g. on a television screen, showing representations related to the game characterized by details of game servers
- A63F2300/55—Details of game data or player data management
- A63F2300/5586—Details of game data or player data management for enforcing rights or rules, e.g. to prevent foul play
Abstract
The embodiment of the application discloses a user cheating behavior detection method based on machine learning, which comprises the following steps: receiving a plurality of groups of combat data corresponding to a plurality of game plays reported by a target user through a target terminal, wherein each group of combat data corresponds to one game play, and each group of combat data comprises reported combat results and role data of a target role in the corresponding game play; based on the character data and the prediction model of the target character of each group of fighting data, obtaining the predicted fighting result of the target user in each game match; calculating target deviation values between a plurality of predicted combat results and a plurality of reported combat results corresponding to a plurality of game matches according to the predicted combat result of each game match and the reported combat result of each game match; judging whether the target offset value is within a preset offset range; and if the target offset value is not within the preset offset range, determining that the target user has cheating behaviors. The embodiment of the application can detect cheating behaviors in the online game.
Description
Technical Field
The embodiment of the application relates to the technical field of computers, in particular to a method and a system for detecting user cheating behaviors based on machine learning, computer equipment, a computer readable storage medium and a prediction model training method for detecting the user cheating behaviors.
Background
The network game is a multiplayer online game, such as an online battle game, which takes a network as a transmission medium, a game operator server and a user computer as processing terminals, and game client software as an information interaction window and aims to realize entertainment, leisure, communication and virtual achievement. With the development of network games, players have higher and higher requirements on the picture surfaces of the games, so that the fighting calculation amount of the games is very large. In order to reduce the calculation burden of the server, the user terminal is sometimes allowed to bear part of the combat calculation burden, namely, the combat result is reported to the server by the user terminal, and the server obtains corresponding reward according to the combat result reported by the user terminal.
However, the method of reporting the fighting results to the server by the user terminal has a security risk, and some users tamper the game fighting results through the user terminal by using a cheating program and report the tampered game fighting results to the server, so that the game rewards are obtained in the game through the cheating means, and the cheating action seriously damages the fairness. Therefore, how to detect cheating in the network game becomes one of the concerns of the network game industry.
Disclosure of Invention
In view of the above, an object of the embodiments of the present application is to provide a method, a system, a computer device, and a computer-readable storage medium for detecting user cheating behaviors based on machine learning, so as to solve the problem of how to detect cheating behaviors in a network game.
One aspect of the present application provides a method for detecting a user cheating behavior based on machine learning, including: receiving a plurality of groups of combat data corresponding to a plurality of game plays reported by a target user through a target terminal, wherein each group of combat data corresponds to one game play, and each group of combat data comprises reported combat results and role data of a target role in the corresponding game plays; based on the character data and the prediction model of the target character of each group of fighting data, obtaining the predicted fighting result of the target user in each game match; calculating target deviation values between a plurality of predicted combat results and a plurality of reported combat results corresponding to the plurality of game matches according to the predicted combat result of each game match and the reported combat result of each game match; judging whether the target offset value is within a preset offset range; and if the target deviation value is not within the preset deviation range, determining that the target user has cheating behaviors.
Optionally, the multiple groups of combat data include N groups of combat data, where N is a positive integer greater than or equal to 2; based on the character data and the prediction model of the target character of each group of fighting data, obtaining the predicted fighting result of the target user in each game match, wherein the predicted fighting result comprises the following steps: converting the character data of the target character in the ith group of fighting data into a target fighting capacity characteristic value, wherein the ith group of fighting data corresponds to the ith game match, and i is more than or equal to 1 and less than or equal to N; inputting the target fighting capacity characteristic value and the fighting capacity characteristic values of other characters in the ith game match into the prediction model to obtain a predicted fighting result of the target user in the ith game match.
Optionally, the prediction model is obtained by training based on a support vector machine algorithm.
Optionally, the character data of the target character in the ith group of battle data includes a plurality of character attributes of the target character; converting the role data of the target role in the ith group of combat data into a target combat power characteristic value, which comprises the following steps: and performing weighted summation operation on a plurality of character attributes of the target character in the ith group of combat data to obtain the characteristic value of the target combat power.
Optionally, the method further includes: and if the target offset value is not within the preset offset range, locking the target terminal to prevent the access operation of the target terminal.
An aspect of the present embodiment further provides a system for detecting a user cheating behavior based on machine learning, including: the receiving module is used for receiving a plurality of groups of fighting data corresponding to a plurality of game matches reported by a target user through a target terminal, each group of fighting data corresponds to one game match, and each group of fighting data comprises a reported fighting result in the corresponding game match and role data of a target role; the obtaining module is used for obtaining the predicted combat result of the target user in each game match based on the character data and the prediction model of the target character of each group of combat data; the calculating module is used for calculating target deviation values between a plurality of predicted combat results and a plurality of reported combat results corresponding to the game play of the plurality of times according to the predicted combat result of the game play of each time and the reported combat result of the game play of each time; the judging module is used for judging whether the target offset value is within a preset offset range; and the determining module is used for determining that the target user has cheating behaviors if the target offset value is not within the preset offset range.
An aspect of the embodiments of the present application further provides a computer device, where the computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor is configured to implement the steps of the method for detecting user cheating behavior based on machine learning when executing the computer program.
An aspect of the embodiments of the present application further provides a computer-readable storage medium, in which a computer program is stored, where the computer program is executable by at least one processor to cause the at least one processor to execute the steps of the above-mentioned method for detecting user cheating behavior based on machine learning.
An aspect of an embodiment of the present application further provides a prediction model training method for predicting a game combat result, including: acquiring a plurality of groups of sample combat data, wherein each group of sample combat data corresponds to one game match, and each group of sample combat data comprises actual combat results of each virtual formation in the corresponding game match and role data of each role in each virtual formation; according to the character data of each character in each group of sample combat data, acquiring the fighting capacity characteristic value of each character in each group of sample combat data; and training a pre-selected machine learning model to obtain a prediction model according to the fighting capacity characteristic value of each character in each group of sample fighting data and the actual fighting result of each virtual formation corresponding to each group of sample fighting data.
Optionally, the character data of each character in each group of sample combat data includes a plurality of character attributes of a plurality of dimensions, and the combat power feature value of each character in each group of sample combat data is obtained by weighted summation of the plurality of character attributes of the corresponding character in the group of sample combat data.
The user cheating behavior detection method, system, device and computer readable storage medium based on machine learning provided by the embodiment of the application obtain a plurality of predicted combat results of a plurality of game matches based on a plurality of groups of combat data of the plurality of game matches of a target user in a certain time period, and judge whether the target user has cheating behaviors in the time period according to the plurality of predicted combat results and the plurality of reported combat results of the plurality of game matches. In addition, compared with the method for judging whether the user cheats through abnormal combat, the method and the device have the advantages that the cheating detection is carried out through a plurality of corresponding behaviors in the game-to-game for a plurality of times in an integral detection mode in a certain time period, so that the detection precision is higher, and the computing resource for carrying out the cheating detection on abnormal combat data in real time in the game-to-game process every time is saved.
Drawings
Fig. 1 schematically illustrates an application environment diagram of a method for detecting user cheating behaviors based on machine learning according to an embodiment of the present application;
fig. 2 schematically shows a flowchart of a method for detecting user cheating behaviors based on machine learning according to an embodiment of the present application;
FIG. 3 is a detailed flowchart of step S202 in FIG. 2;
FIG. 4 is a detailed flowchart of step S300 in FIG. 3;
FIG. 5 is a diagram schematically illustrating a game screen in a certain game play;
fig. 6 schematically shows another flowchart of a method for detecting user cheating behavior based on machine learning according to an embodiment of the present application;
fig. 7 schematically shows a block diagram of a system for detecting user cheating behaviors based on machine learning according to a second embodiment of the present application;
fig. 8 schematically illustrates a hardware architecture diagram of a computer device suitable for implementing a method for detecting user cheating behaviors based on machine learning according to a third embodiment of the present application; and
fig. 9 schematically shows a flowchart of a predictive model training method for detecting user cheating behavior according to a fifth embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without making any creative effort belong to the protection scope of the present application.
It should be noted that the descriptions in this application referring to "first", "second", etc. are for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In addition, technical solutions between various embodiments may be combined with each other, but must be realized by a person skilled in the art, and when the technical solutions are contradictory or cannot be realized, such a combination should not be considered to exist, and is not within the protection scope of the present application.
Various embodiments will be provided below, and it will be appreciated that various embodiments provided below may be used to solve one or more of the technical problems described above.
Fig. 1 schematically illustrates an environment application diagram of a method for detecting user cheating behaviors based on machine learning according to an embodiment of the present application. In an exemplary embodiment, the gaming service platform 2 may connect a plurality of user terminals (such as 6A, 6B, 6C, 6D, 6E, 6F, \8230;) through the network 4.
The game service platform 2 may provide services through one or more networks 4. Network 4 may include various network devices such as routers, switches, multiplexers, hubs, modems, bridges, repeaters, firewalls, proxy devices, and/or the like. The network 4 may include physical links, such as coaxial cable links, twisted pair cable links, fiber optic links, combinations thereof, and/or the like. The network 4 may include wireless links such as cellular links, satellite links, wi-Fi links, and/or the like.
The game service platform 2 may be composed of a single or multiple computer devices (e.g., servers). The single or multiple computing devices may include virtualized compute instances. The virtualized computing instance may include a virtual machine, such as an emulation of a computer system, an operating system, a server, and so forth. The computing device may load the virtual machine based on a virtual image and/or other data that defines the particular software (e.g., operating system, dedicated application, server) used for emulation. As the demand for different types of processing services changes, different virtual machines may be loaded and/or terminated on one or more computing devices. A hypervisor may be implemented to manage the use of different virtual machines on the same computing device.
User terminals (such as 6A, 6B, 6C, 6D, 6E, 6F, \8230;) may be configured to access content and services of the gaming service platform 2. The user terminal (such as 6A, 6B, 6C, 6D, 6E, 6F, \8230;) may be any type of computing device, such as a mobile device, a tablet device, a laptop computer, a computing station, a smart device (e.g., smart glasses), a gaming device, a set-top box, a robot, a vehicle mounted terminal, a smart television, a television box, etc.
A user terminal (such as 6A, 6B, 6C, 6D, 6E, 6F, \ 8230;) may be associated with one or more users. A single user may access the gaming service platform 2 using one or more of the user terminals 6 (such as 6A, 6B, 6C, 6D, 6E, 6F, \8230;). User terminals (such as 6A, 6B, 6C, 6D, 6E, 6F, \ 8230;) may use different networks to access the game service platform 2.
In an exemplary embodiment, a user terminal (such as 6A, 6B, 6C, 6D, 6E, 6F, \ 8230;) may include a game client program 8. The game client program 8 outputs (e.g., displays, renders) the content to the user. The user terminal 6 may comprise a terminal interface 10, which terminal interface 10 may comprise an input element. For example, the input element may be configured to receive a user instruction. In some embodiments, a user terminal (such as 6A, 6B, 6C, 6D, 6E, 6F, \ 8230;) may generate and transmit game instructions based on information input by a user.
Example one
Fig. 2 schematically shows a flowchart of a method for detecting user cheating behavior based on machine learning according to an embodiment of the present application. It will be appreciated that the method embodiments may be implemented in the gaming service platform 2 (single or multiple computer devices) and that the flow charts of the method embodiments are not intended to limit the order in which the steps are performed.
As shown in fig. 2, the method for detecting user cheating behavior based on machine learning may include steps S200 to S208, wherein:
step S200, the game service platform 2 receives a plurality of groups of combat data corresponding to a plurality of game matches reported by a target user through a target terminal (for example, 6A), wherein each group of combat data corresponds to one game match, and each group of combat data comprises reported combat results in the corresponding game matches and role data of target roles.
The game service platform 2 may, to ease the burden, transfer part of the game computing task to the user terminal participating in the game play.
When the game service platform 2 hands over the fighting calculation task related to the target terminal 6A for execution, the target terminal 6A needs to perform operation on the fighting data of the target account of the target user in the game match, and reports the fighting data and the fighting result of the target account of the target user in the game match to the game service platform 2, and the fighting result reported to the game service platform 2 is also called as a reported fighting result.
The game service platform 2 receives a plurality of groups of fighting data reported by the target terminal 6A. In units of days, if the target user performs N game plays through the target terminal 6A, the target terminal 6 reports N sets of combat data generated in the N game plays to the game service platform 2. That is, the game service platform 2 receives N sets of combat data corresponding to N game plays from the target terminal on the day, and each set of combat data includes character data of a target character corresponding to a target user in the corresponding game play and a reported combat result (win-or-lose result) of the target user in the game play. It should be noted that the target character of the target user in each game play may be different. Taking the game of 'three kingdoms killing' as an example, in the first game match of the day, the target user may select 'yogy' as the virtual character in the game match, and then the 'yogy' in the first game match is the target character; in a second game play, the target user may select "prepare" as his virtual character in that game play and "prepare" as the target character in the second game play.
Of course, the game service platform 2 may also receive the combat data reported by other user terminals, such as the combat data reported by the user terminal 6B. Illustratively, the target terminal 6A and the user terminals 6B, 6C, 6D, 6E, 6F group participate in a game play, wherein the role selected by the three parties, the target terminal 6A and the user terminals 6B, 6C, is a first play, the role selected by the three parties, the user terminals 6D, 6E, 6F, is a second play, and the win-or-lose results are generated in the battle of the first and second plays. In the exemplary game play, the game service platform 2 not only receives a set of combat data generated by the target terminal 6A (i.e., the character data of the target character "yoga" selected by the target terminal 6A and the reported combat result), but also receives multiple sets of combat data reported by other user terminals, such as: a set of combat data reported by the user terminal 6B (i.e. the role data of the other role "cao" selected by the user terminal 6B and the reported combat result), a set of combat data reported by the user terminal 6C (i.e. the role data of the other role "cao" selected by the user terminal 6C and the reported combat result), 8230.
In step S202, the game service platform 2 may obtain the predicted combat result of the target user in each game play based on the character data and the prediction model of the target character of each set of combat data.
For example:
inputting a set of combat data generated in a first game play into a predictive model to obtain a predicted combat result of a target character of the target user in the first game play;
in the second game play, inputting a group of fighting data generated in the second game play into a prediction model to obtain a predicted fighting result of the target role of the target user in the second game play;
by analogy, the game service platform 2 can obtain the predicted combat result of the target user in each game play.
A group of combat data generated in each game match comprises reported combat results and role data of target roles. The role data includes a plurality of role attributes of a plurality of dimensions, such as a role attribute representing an attack power, a role attribute representing an attack action, and the like. In some embodiments, the character attributes may be some custom attributes preset in the game, such as: physical attack power, magic attack power, physical defense power, magic attack power, physical assault, magic assault, HP, TP auto-recovery, avoidance, increase in recovery, HP auto-recovery, TP increase, HP absorption, TP consumption reduction, travel speed, and normal attack interval. It will be understood that the game service platform 2 will use all the character data as the input of the prediction model, which will certainly consume excessive computing resources.
In order to reduce the calculation amount in the prediction process, the input parameters input into the prediction model are changed, namely, the characteristics of a plurality of character attributes of a plurality of dimensions originally input into the prediction model are extracted, and the fighting capacity characteristic value obtained by characteristic extraction is used as the input of the prediction model. Assume that the multiple sets of combat data reported by the target terminal 6A include N sets of combat data, where N is a positive integer greater than or equal to 2. As shown in fig. 3, step S202 includes the steps of: step S300, converting the character data of the target character in the ith group of fighting data into a target fighting capacity characteristic value, wherein the ith group of fighting data corresponds to the ith game match, and i is more than or equal to 1 and less than or equal to N; step S302, inputting the target fighting capacity characteristic value and the fighting capacity characteristic values of other characters in the ith game match into the prediction model to obtain the predicted fighting result of the target user in the ith game match. Steps S300 and S302 may be looped until N predicted combat results corresponding to N game plays are obtained.
It is understood that the character data of the target character of the ith group of battle data includes a plurality of character attributes of the target character, as shown in fig. 4, and step S300 may be implemented by the following steps, step S300': performing weighted summation operation on a plurality of character attributes of the target character in the ith group of combat data to obtain the characteristic value of the target combat power, as follows:
source_factor＝∑factor_score j *w j
wherein, sigma factor _ score j J-th role attribute, w, representing the target role j The weighting coefficient represents the jth role attribute of the target role, the source _ factor represents the fighting capacity characteristic value of the target role in the ith game match corresponding to the ith group of fighting data, and the weighting coefficient of each role data is preset manually or in other manners.
For example: as shown in fig. 5, the group of the target terminal 6A and the user terminals 6B, 6C, 6D, 6E, 6F participates in one game play, in which the role selected by the three parties of the target terminal 6A and the user terminals 6B, 6C is the first play, the role selected by the three parties of the user terminals 6D, 6E, 6F is the second play, and the win-or-lose results are generated in the battle of the first play and the second play. The game service platform 2 receives a set of combat data generated by the target terminal 6A (i.e., character data of the target character "weekly yog" selected by the target terminal 6A and reported combat results), and receives a plurality of sets of combat data reported by other user terminals, such as: a set of combat data reported by the user terminal 6B (i.e. the role data of the other role "cao" selected by the user terminal 6B and the reported combat result), a set of combat data reported by the user terminal 6C (i.e. the role data of the other role "cao" selected by the user terminal 6C and the reported combat result), 8230. The game service platform 2 converts the character data of the target character 'Zhouyi' provided by the target terminal 6A into a target fighting capacity characteristic value 800, converts the character data of other characters 'Cao cao' selected by the user terminal 6B into a fighting capacity characteristic value 1001, and the like to finally obtain the fighting capacity characteristic values of all characters in the game play as follows: the target character 'week yoga' has a target fighting capacity characteristic value of 800, the character 'Cao' has a fighting capacity characteristic value of 1001, and the character 'yellow cover' has a fighting capacity characteristic value of 500; and (3) second formation: the feature value of the fighting power of the character "Zhuge Liang" is 500, the feature value of the fighting power of the character "Dianwei" is 600, and the feature value of the fighting power of the character "Semayi" is 1100. The game service platform 2 generates a parameter vector (800, 1001, 500, 500, 600, 1100) from the plurality of fighting power feature values, and then inputs the parameter vector into the prediction model so that the prediction model outputs a predicted fighting result (predicted fighting win-win) of the first or second battle.
In an exemplary embodiment, the predictive model is trained based on a Support Vector Machine (SVM) algorithm. A plurality of combat data are calculated through a linear kernel function of an SVM algorithm to obtain a hyperplane for online real-time prediction, and the accuracy of combat win-loss predicted by the method is as high as 90%. The SVM algorithm is a generalized linear classifier (generalized linear classifier) that performs binary classification on data in a supervised learning (supervised learning) manner, and a decision boundary of the SVM algorithm is a maximum-margin hyperplane (maximum-margin hyperplane) that solves for a learning sample. The SVM algorithm calculates an empirical risk (empirical risk) using a hinge loss function (change loss) and adds a regularization term to a solution system to optimize a structural risk (structural risk), and is a classifier with sparsity and robustness. The SVM algorithm can be classified non-linearly by a kernel method.
In step S204, the game service platform 2 may calculate target deviation values between a plurality of predicted combat results and a plurality of reported combat results corresponding to the game plays of the plurality of times according to the predicted combat result and the reported combat result of each game play.
Continuing to use the day as a unit, if the target user performs game match for 10 times through the target terminal 6A, the reported fighting results of the target terminal 6A are as follows in sequence: win, lose, win, lose. The game service platform 2 obtains the predicted combat results of the target user as follows in sequence: lose, win, lose, win, lose. The first calculation method is as follows: the game service platform 2 may count the game match times for which the predicted combat result and the reported combat result are inconsistent, and take a ratio between the game match times for which the predicted combat result and the reported combat result are inconsistent and the total game match times as the target offset value. Specifically, in an exemplary 10 game plays, if the number of game plays for which the predicted combat result is inconsistent with the reported combat result is 6, the target offset value is 0.6. And a second calculation method: the game service platform 2 may count the winning proportions in all the predicted fighting results and the winning proportions in all the reported fighting results, and take the difference between the winning proportions in all the predicted fighting results and the differences between all the reported fighting results as the target deviation value. Specifically, in the exemplary 10 game plays, the winning ratio of all predicted fighting results is 0.2, the winning ratio of all reported fighting results is 0.6, and the target offset value is 0.4. The two calculation methods are only exemplary and are not intended to limit the application of the present application to other calculation methods.
In step S206, the game service platform 2 determines whether the target offset value is within a preset offset range.
The preset offset range can be a priori value, and the value can also be adjusted according to the prediction accuracy of the preset model.
Step S208, if the target offset value is not within the preset offset range, the game service platform 2 determines that the target user has a cheating behavior.
And if the target deviation value exceeds the preset deviation range, indicating that the target user is suspected to cheat the reward. Therefore, the game service platform 2 needs to perform corresponding subsequent operations, such as generating a short message or mail content and triggering a short message system or a mailbox system to notify a background operator, so that the latter operator can perform important monitoring on the target user.
In an exemplary embodiment, to prevent continuous cheating, as shown in fig. 6, the method for detecting cheating behavior of a user based on machine learning further includes step S210: and if the target offset value is not within the preset offset range, locking the target terminal 6A to prevent the access operation of the target terminal 6A.
The user cheating behavior detection method based on machine learning provided by the embodiment of the application obtains a plurality of predicted combat results of a plurality of game matches based on a plurality of groups of combat data of the plurality of game matches of a target user in a certain time period, and judges whether the target user has cheating behaviors in the time period according to the plurality of predicted combat results and the plurality of reported combat results of the plurality of game matches.
It should be noted that, compared with the method for judging whether the user cheats through abnormal combat, the method and the device for detecting cheating through a plurality of corresponding behaviors in game-to-game for a plurality of times in an integral detection mode in a certain time period have higher detection precision, and save computing resources for detecting cheating on abnormal combat data in real time in the game-to-game process each time.
Example two
Fig. 7 is a block diagram schematically illustrating a machine learning-based user cheating-behavior detection system according to a second embodiment of the present application, which can be divided into one or more program modules, the one or more program modules being stored in a storage medium and executed by one or more processors to complete the second embodiment of the present application. The program modules referred to in the embodiments of the present application refer to a series of computer program instruction segments that can perform specific functions, and the following description will specifically describe the functions of the program modules in the embodiments.
As shown in fig. 7, the machine learning-based user cheating behavior detection system 700 may include a receiving module 710, an obtaining module 720, a calculating module 730, a determining module 740, and a determining module 750, wherein:
a receiving module 710, configured to receive multiple sets of combat data corresponding to multiple game plays reported by a target user through a target terminal, where each set of combat data corresponds to one game play, and each set of combat data includes a reported combat result and role data of a target role in the corresponding game play;
an obtaining module 720, configured to obtain a predicted combat result of the target user in each game match based on the character data and the prediction model of the target character of each set of combat data;
a calculating module 730, configured to calculate, according to the predicted combat result of each game match and the reported combat result of each game match, a target offset value between a plurality of predicted combat results and a plurality of reported combat results corresponding to the plurality of game matches;
a determining module 740, configured to determine whether the target offset value is within a preset offset range; and
a determining module 750, configured to determine that the target user has a cheating behavior if the target offset value is not within the preset offset range.
In an exemplary embodiment, the plurality of sets of combat data includes N sets of combat data, where N is a positive integer greater than or equal to 2. The obtaining module 720 is further configured to: converting the character data of the target character in the ith group of combat data into a target combat power characteristic value, wherein the ith group of combat data corresponds to the ith game match, and i is more than or equal to 1 and less than or equal to N; and inputting the target fighting capacity characteristic value and the fighting capacity characteristic values of other characters in the ith game play into the prediction model to obtain a predicted fighting result of the target user in the ith game play.
In an exemplary embodiment, the predictive model is trained based on a support vector machine algorithm.
In an exemplary embodiment, the character data of the target character of the ith set of combat data includes a plurality of character attributes of the target character. Converting the role data of the target role in the ith group of combat data into a target combat power characteristic value, which comprises the following steps: and performing weighted summation operation on a plurality of character attributes of the target character in the ith group of combat data to obtain the characteristic value of the target combat power.
In an exemplary embodiment, the machine learning based user cheating behavior detection system 700 may further comprise a locking module for: and if the target offset value is not within the preset offset range, locking the target terminal to prevent the access operation of the target terminal.
EXAMPLE III
Fig. 8 schematically shows a hardware architecture diagram of a computer device suitable for implementing a machine learning-based user cheating behavior detection method according to a third embodiment of the present application. The computer device 2 may be a game service platform or a node device in a game service platform. In the present embodiment, the computer device 2 is a device capable of automatically performing numerical calculation and/or information processing in accordance with a command set or stored in advance. For example, the computer device 2 may be a rack server, a blade server, a tower server or a cabinet server (including an independent server or a server cluster composed of a plurality of servers), and the like. As shown in fig. 8, the computer device 2 includes at least, but is not limited to: memory 810, processor 820, and network interface 830 may be communicatively linked to each other by a system bus. Wherein:
the memory 810 includes at least one type of computer-readable storage medium including flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the storage 810 may be an internal storage module of the computer device 2, such as a hard disk or a memory of the computer device 2. In other embodiments, the memory 810 may also be an external storage device of the computer device 2, such as a plug-in hard disk provided on the computer device 2, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like. Of course, memory 810 may also include both internal and external memory modules of computer device 2. In this embodiment, the memory 810 is generally used for storing an operating system installed on the computer device 2 and various types of application software, such as program codes of a user cheating behavior detection method based on machine learning, and the like. In addition, the memory 810 may also be used to temporarily store various types of data that have been output or are to be output.
Processor 820 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data Processing chip in some embodiments. The processor 820 is generally used for controlling the overall operation of the computer device 2, such as performing control and processing related to data interaction or communication with the computer device 2. In this embodiment, the processor 820 is used to execute program codes stored in the memory 810 or process data.
Network interface 830 may include a wireless network interface or a wired network interface, and network interface 830 is typically used to establish communication links between computer device 2 and other computer devices. For example, the network interface 830 is used to connect the computer device 2 with an external terminal via a network, establish a data transmission channel and a communication link between the computer device 2 and the external terminal, and the like. The network may be an Intranet (Internet), the Internet (Internet), a Global System of Mobile communication (GSM), wideband Code Division Multiple Access (WCDMA), a 4G network, a 5G network, bluetooth (Bluetooth), wi-Fi, or other wireless or wired network.
It should be noted that fig. 8 only shows a computer device having components 810-830, but it should be understood that not all of the shown components are required to be implemented, and that more or fewer components may be implemented instead.
In this embodiment, the machine learning based user cheating behavior detection system stored in the memory 810 can be further divided into one or more program modules and executed by one or more processors (in this embodiment, the processor 820) to complete the present application.
Example four
The present embodiments also provide a computer-readable storage medium having stored thereon a computer program, which when executed by a processor, implements the steps of the machine learning-based user cheating behavior detection method of the embodiments.
In this embodiment, the computer-readable storage medium includes a flash memory, a hard disk, a multimedia card, a card type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Programmable Read Only Memory (PROM), a magnetic memory, a magnetic disk, an optical disk, and the like. In some embodiments, the computer readable storage medium may be an internal storage unit of the computer device, such as a hard disk or a memory of the computer device. In other embodiments, the computer readable storage medium may be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like provided on the computer device. Of course, the computer-readable storage medium may also include both internal and external storage units of the computer device. In this embodiment, the computer-readable storage medium is generally used to store an operating system and various types of application software installed in a computer device, for example, the program code of the machine learning-based user cheating behavior detection method in the embodiment, and the like. In addition, the computer-readable storage medium may also be used to temporarily store various types of data that have been output or are to be output.
EXAMPLE five
Fig. 9 schematically shows a flowchart of a predictive model training method for detecting user cheating behavior according to a fifth embodiment of the present application. It will be appreciated that the method embodiments may be performed in a single computer device or in a distributed computer device, and that the flow charts of the method embodiments are not intended to limit the order in which the steps are performed.
As shown in fig. 9, the method for detecting user cheating behavior based on machine learning may include steps S900 to S904, wherein:
and S900, acquiring a plurality of groups of sample combat data, wherein each group of sample combat data corresponds to one game match, and each group of sample combat data comprises the actual combat result of each virtual formation in the corresponding game match and the role data of each role in each virtual formation.
That is, a plurality of sets of combat data of a plurality of non-cheating game plays are collected, and virtual battle forms, attributes, actual combat results, and the like of each game play are set as reference data.
Step S902, according to the character data of each character in each group of sample combat data, obtaining the fighting capacity characteristic value of each character in each group of sample combat data.
For example, a character fighting capacity calculation function is defined, character data of each character is converted into a fighting capacity characteristic value, and the fighting capacity characteristic value is changed into a machine language; and constructing key two-dimensional characteristic data by reducing the dimensionality of the multi-dimensional character data, preprocessing the key two-dimensional characteristic data, and carrying out descaler and normalization on the character data of each character.
Step S904, training a pre-selected machine learning model to obtain a prediction model according to the fighting power characteristic value of each character in each group of sample fighting data and the actual fighting result of each virtual formation corresponding to each group of sample fighting data.
The machine learning model may be a support vector machine or other model, such as a deep neural network model. When the machine learning model is a support vector machine, the processed multiple groups of sample combat data can be trained through a linear kernel function, and a prediction model for predicting a combat result is constructed.
In an exemplary embodiment, the character data of each character in each set of sample combat data includes a plurality of character attributes of a plurality of dimensions, and the fighting capacity characteristic value of each character in each set of sample combat data is obtained by weighted summation of the plurality of character attributes of the corresponding character in the set of sample combat data.
The prediction model training method provided by the embodiment changes input parameters input into the machine learning model, namely, performs feature extraction on a plurality of character attributes of a plurality of dimensions originally input into the machine learning model, and uses a fighting capacity feature value obtained by feature extraction as the input of the machine learning model, thereby effectively reducing the calculation operand in the training process.
It should be obvious to those skilled in the art that the modules or steps of the embodiments of the present application described above can be implemented by a general-purpose computing device, they can be centralized on a single computing device or distributed on a network composed of a plurality of computing devices, alternatively, they can be implemented by program code executable by the computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a sequence different from that shown or described, or they can be separately manufactured as individual integrated circuit modules, or a plurality of modules or steps in them can be manufactured as a single integrated circuit module. Thus, embodiments of the present application are not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are included in the scope of the present application.
Claims (10)
1. A user cheating behavior detection method based on machine learning is characterized by comprising the following steps:
receiving a plurality of groups of combat data corresponding to a plurality of game plays reported by a target user through a target terminal, wherein each group of combat data corresponds to one game play, and each group of combat data comprises reported combat results and role data of a target role in the corresponding game play;
based on the character data and the prediction model of the target character of each group of fighting data, obtaining the predicted fighting result of the target user in each game match;
calculating target deviation values between a plurality of predicted combat results and a plurality of reported combat results corresponding to the plurality of game matches according to the predicted combat result of each game match and the reported combat result of each game match;
judging whether the target offset value is within a preset offset range; and
and if the target offset value is not within the preset offset range, determining that the target user has cheating behaviors.
2. The method for detecting user cheating behaviors based on machine learning according to claim 1, wherein the multiple groups of fighting data comprise N groups of fighting data, wherein N is a positive integer greater than or equal to 2;
based on the character data and the prediction model of the target character of each group of fighting data, obtaining the predicted fighting result of the target user in each game match, wherein the predicted fighting result comprises the following steps:
converting the character data of the target character in the ith group of combat data into a target combat power characteristic value, wherein the ith group of combat data corresponds to the ith game match, and i is more than or equal to 1 and less than or equal to N;
and inputting the target fighting capacity characteristic value and the fighting capacity characteristic values of other characters in the ith game play into the prediction model to obtain a predicted fighting result of the target user in the ith game play.
3. The method of claim 2, wherein the predictive model is trained based on a support vector machine algorithm.
4. The machine-learning-based user cheating behavior detection method according to claim 2, wherein character data of a target character of said ith group of battle data comprises a plurality of character attributes of said target character;
converting the role data of the target role in the ith group of combat data into a target combat power characteristic value, wherein the step comprises the following steps:
and performing weighted summation operation on a plurality of character attributes of the target character in the ith group of combat data to obtain the characteristic value of the target combat power.
5. The method for detecting user cheating behaviors based on machine learning according to any one of claims 1 to 4, further comprising:
and if the target offset value is not within the preset offset range, locking the target terminal to prevent the access operation of the target terminal.
6. A machine learning based user cheating behavior detection system, the system comprising:
the receiving module is used for receiving a plurality of groups of fighting data corresponding to a plurality of game matches reported by a target user through a target terminal, each group of fighting data corresponds to one game match, and each group of fighting data comprises a reported fighting result in the corresponding game match and role data of a target role;
the obtaining module is used for obtaining the predicted combat result of the target user in each game match based on the character data and the prediction model of the target character of each group of combat data;
the calculating module is used for calculating target deviation values between a plurality of predicted combat results and a plurality of reported combat results corresponding to the game games for a plurality of times according to the predicted combat result of each game and the reported combat result of each game;
the judging module is used for judging whether the target offset value is within a preset offset range; and
and the determining module is used for determining that the target user has cheating behaviors if the target offset value is not within the preset offset range.
7. A computer arrangement comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any of claims 1 to 5 when executing the computer program.
8. A computer-readable storage medium, having stored thereon a computer program, the computer program being executable by at least one processor to cause the at least one processor to perform the steps of the method according to any one of claims 1 to 5.
9. A predictive model training method for predicting a game combat outcome, comprising:
acquiring a plurality of groups of sample combat data, wherein each group of sample combat data corresponds to one game match, and each group of sample combat data comprises actual combat results of each virtual formation in the corresponding game match and role data of each role in each virtual formation;
according to the character data of each character in each group of sample combat data, acquiring the fighting capacity characteristic value of each character in each group of sample combat data; and
training a pre-selected machine learning model to obtain a prediction model according to the fighting capacity characteristic value of each character in each group of sample fighting data and the actual fighting result of each virtual formation corresponding to each group of sample fighting data;
the predictive model is to: based on the role data of the target role of each group of combat data, obtaining the predicted combat result of the target user in each game match; and a target offset value between the predicted combat result in each game match and the reported combat result of each game match is used for determining whether the target user has cheating behaviors.
10. The predictive model training method for predicting game combat results of claim 9, wherein the character data of each character in each set of sample combat data comprises a plurality of character attributes of a plurality of dimensions, and the combat power characteristic value of each character in each set of sample combat data is obtained by weighted summation of the plurality of character attributes of the corresponding character in the set of sample combat data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010090357.9A CN112402982B (en) | 2020-02-13 | 2020-02-13 | User cheating behavior detection method and system based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010090357.9A CN112402982B (en) | 2020-02-13 | 2020-02-13 | User cheating behavior detection method and system based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112402982A CN112402982A (en) | 2021-02-26 |
| CN112402982B true CN112402982B (en) | 2022-12-23 |
Family
ID=74844081
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010090357.9A Active CN112402982B (en) | 2020-02-13 | 2020-02-13 | User cheating behavior detection method and system based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112402982B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113181637B (en) * | 2021-05-10 | 2024-04-16 | 上海幻电信息科技有限公司 | Game playback method and system |
| CN113413605B (en) * | 2021-07-08 | 2022-08-26 | 腾讯数码（深圳）有限公司 | Method, device, equipment and storage medium for detecting cheating account of game |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109603159A (en) * | 2018-12-18 | 2019-04-12 | 网易（杭州）网络有限公司 | Match the method, apparatus and system of game player |
| CN110119547A (en) * | 2019-04-28 | 2019-08-13 | 腾讯科技（深圳）有限公司 | A kind of prediction group defeats negative method, apparatus and control equipment |
| CN110458295A (en) * | 2019-08-19 | 2019-11-15 | 腾讯科技（深圳）有限公司 | Chess and card outpost generation method, training method and device based on artificial intelligence |
| CN110772797A (en) * | 2019-10-29 | 2020-02-11 | 腾讯科技（深圳）有限公司 | Data processing method, device, server and storage medium |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3454357B2 (en) * | 2001-07-05 | 2003-10-06 | 株式会社コナミコンピュータエンタテインメントスタジオ | Network game server device, network game progress control method, and network game progress control program |
| US10537801B2 (en) * | 2013-07-11 | 2020-01-21 | International Business Machines Corporation | System and method for decision making in strategic environments |
| US10713965B2 (en) * | 2017-12-28 | 2020-07-14 | Mcgraw-Hill Global Education Holdings, Llc | Negative learning behavior alert system |
| CN109453514A (en) * | 2018-10-25 | 2019-03-12 | 珠海市君天电子科技有限公司 | A kind of game cheat detection method and device |
| CN109718556B (en) * | 2019-01-30 | 2021-09-28 | 腾讯科技（深圳）有限公司 | Game data processing method and device and server |
| CN109966743A (en) * | 2019-03-06 | 2019-07-05 | 网易（杭州）网络有限公司 | A kind of prediction technique, model generating method and the device of game winning rate |
-
2020
- 2020-02-13 CN CN202010090357.9A patent/CN112402982B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109603159A (en) * | 2018-12-18 | 2019-04-12 | 网易（杭州）网络有限公司 | Match the method, apparatus and system of game player |
| CN110119547A (en) * | 2019-04-28 | 2019-08-13 | 腾讯科技（深圳）有限公司 | A kind of prediction group defeats negative method, apparatus and control equipment |
| CN110458295A (en) * | 2019-08-19 | 2019-11-15 | 腾讯科技（深圳）有限公司 | Chess and card outpost generation method, training method and device based on artificial intelligence |
| CN110772797A (en) * | 2019-10-29 | 2020-02-11 | 腾讯科技（深圳）有限公司 | Data processing method, device, server and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112402982A (en) | 2021-02-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109621422B (en) | Electronic chess and card decision model training method and device and strategy generation method and device | |
| CN109513215B (en) | Object matching method, model training method and server | |
| CN112700252B (en) | Information security detection method and device, electronic equipment and storage medium | |
| CN113599803A (en) | Data processing method and device based on edge calculation and readable storage medium | |
| US20130339473A1 (en) | Real time analytics via stream processing | |
| CN111330279B (en) | Strategy decision model training method and device for game AI | |
| CN112402982B (en) | User cheating behavior detection method and system based on machine learning | |
| CN110175299B (en) | Recommendation information determining method and server | |
| CN108319974B (en) | Data processing method, data processing device, storage medium and electronic device | |
| CN114021737B (en) | Reinforced learning method, system, terminal and storage medium based on game | |
| CN111957047A (en) | Checkpoint configuration data adjusting method, computer equipment and storage medium | |
| CN112418259A (en) | Method for configuring real-time rules based on user behaviors in live broadcast process, computer equipment and readable storage medium | |
| CN115481441A (en) | Difference privacy protection method and device for federal learning | |
| CN110796591A (en) | GPU card using method and related equipment | |
| US20220343180A1 (en) | Learning device, learning method, and learning program | |
| CN113230650B (en) | Data processing method and device and computer readable storage medium | |
| CN113011886B (en) | Method and device for determining account type and electronic equipment | |
| CN111770352B (en) | Security detection method and device, electronic equipment and storage medium | |
| CN113230665B (en) | Resource allocation method and device | |
| CN114463117A (en) | User behavior prediction method, system and device | |
| CN113887708A (en) | Multi-agent learning method based on mean field, storage medium and electronic device | |
| CN113318448A (en) | Game resource display method and device, equipment and model training method | |
| CN117618918B (en) | Virtual scene processing method and device, electronic equipment and storage medium | |
| CN112190950A (en) | Method and device for detecting abnormal player account | |
| KR20200072391A (en) | Method and apparatus for predicting game indicator information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |