CN110152306A - Script user identification method and system - Google Patents
Script user identification method and system Download PDFInfo
- Publication number
- CN110152306A CN110152306A CN201910658519.1A CN201910658519A CN110152306A CN 110152306 A CN110152306 A CN 110152306A CN 201910658519 A CN201910658519 A CN 201910658519A CN 110152306 A CN110152306 A CN 110152306A
- Authority
- CN
- China
- Prior art keywords
- user
- game
- collection
- machine learning
- label
- 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.)
- Granted
Links
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
- A63F13/30—Interconnection arrangements between game servers and game devices; Interconnection arrangements between game devices; Interconnection arrangements between game servers
- A63F13/35—Details of game servers
-
- 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
- A63F13/45—Controlling the progress of the video game
- A63F13/46—Computing the game score
-
- 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
- A63F13/79—Game security or game management aspects involving player-related data, e.g. identities, accounts, preferences or play histories
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Business, Economics & Management (AREA)
- Computer Security & Cryptography (AREA)
- General Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The present invention discloses a kind of script user identification method, comprising: each game user is marked respectively by rules unit to obtain the first label result；Each game user is marked respectively by unsupervised machine learning unit to obtain the second label result；According to the first label result and the second label as a result, each game user is marked respectively by Supervised machine learning unit to obtain third label result；It is marked according to the third as a result, obtaining Supervised machine learning model by the training of Supervised machine learning module；Abnormality degree marking is carried out to each game user respectively by the Supervised machine learning model, score value is more than that the game user of first threshold is labeled as script user.Script user's recognition capability of system can be improved in the present invention, and adaptivity is good, and script user is allowed to be difficult to the identification by Rule Summary avoidance system.
Description
Technical field
The present invention relates to script users to identify field, and in particular to a kind of script user identification method and system.
Background technique
There are a large amount of script user in game, these scripts user is partially tryed to gain in game by some operations
Resource, influence other normal game user experiences, extra expenses caused to game operation and server.Existing script user
Recognition methods mainly passes through setting rule (such as setting threshold value), if some user passes through single IP repeat logon difference account
Number behavior violate the rule (or threshold value beyond setting) of setting, then the user is identified as script user, but such
Identification method is easy to allow script user by attempting discovery threshold value repeatedly to the identification of avoidance system.
Summary of the invention
In order to overcome the deficiencies of the prior art, the one side of the application is to provide a kind of script user identification method, passes through
Script user is identified in such a way that unsupervised machine learning and Supervised machine learning combine, adaptivity is good,
Script user's recognition capability of system can be improved, and script user is allowed to be difficult to the knowledge by Rule Summary avoidance system
Not.This method is realized by following technological means:
Script user identification method, comprising:
Each game user is marked to obtain the first label as a result, the first label result packet respectively by rules unit
Include the first normal users collection and the first abnormal user collection；
Each game user is marked to obtain the second label as a result, described second respectively by unsupervised machine learning unit
Marking result includes the second normal users collection, the second abnormal user collection and uncertain user collection；
According to the first label result and the second label as a result, being used respectively each game by Supervised machine learning unit
Family is marked to obtain third label as a result, third label result includes third normal users collection and third abnormal user
Collection；
It is marked according to the third as a result, being obtained by the Supervised machine learning module training in Supervised machine learning unit
Supervised machine learning model；
Abnormality degree marking is carried out to each game user respectively by the Supervised machine learning model, score value is more than the first threshold
The game user of value is labeled as script user.
Further, described that each game user is marked respectively by rules unit to obtain the first label result packet
It includes:
Login/log-on data of each game user is obtained respectively, includes corresponding game user in each login/log-on data
ID；
Judge whether login/log-on data of each game user violates preset rules respectively, if so, abnormal labeled as first
Otherwise user is labeled as the first normal users；
Wherein, all first normal users constitute the first normal users collection, and all first abnormal users constitute the first abnormal user
Collection.
Further, described that all game users are marked to obtain the second label by unsupervised machine learning unit
Result includes:
The first behavioral data of each game user is obtained respectively, and first behavioral data is that game user is different in game
Behavior topological data in scene of game includes the ID of corresponding game user in each first behavioral data；
According to the first behavioral data, the marking of suspicion degree is carried out to each game user using highly dense subgraph mining algorithm, score value is super
The game user of second threshold is crossed labeled as the second abnormal user, score value is being labeled as second just lower than the game user of third threshold value
Common family, game user of the score value between second threshold and third threshold value are labeled as uncertain user；
Wherein, all second normal users constitute the second normal users collection, and all second abnormal users constitute the second abnormal user
Collection, all uncertain users constitute uncertain user's collection.
Further, described to be marked according to the first label result and second as a result, passing through Supervised machine learning list
Member is marked to obtain third label result to all game users
According to the first label result and the second label as a result, the first normal users collection and the second normal users collection will be belonged to simultaneously
Game user is labeled as third normal users, will belong to the game user of the first abnormal user collection and the second abnormal user collection simultaneously
Labeled as third abnormal user, by belong to the first normal users collection and belong to the second abnormal user collection game user label the
Existing uncertain user is added in the user for belonging to the first abnormal user collection and belong to the second normal users collection by three abnormal users
Collection；
Wherein, all third normal users constitute third normal users collection, and all third abnormal users constitute third abnormal user
Collection.
Further, described to be marked according to the third as a result, by having intendant in Supervised machine learning unit
The training of device study module obtains Supervised machine learning model and includes:
The second behavioral data of each game user is obtained respectively, and second behavioral data is row of the game user in game
Data are characterized, each second behavioral data includes the ID of corresponding game user；
The ID that each game user is concentrated according to third abnormal user obtains third abnormal user concentrates each game user
Two behavioral datas, third abnormal user concentrate the second behavioral data of all game users to constitute exceptional sample together；
The ID that each game user is concentrated according to third normal users obtains third normal users concentrate each game user
Two behavioral datas, third normal users concentrate the second behavioral data of all game users to constitute normal sample together；
By the exceptional sample and normal sample together as the Supervised machine learning module in Supervised machine learning unit
Training sample, training obtain Supervised machine learning model.
Further, it is described by the exceptional sample and normal sample together as having in Supervised machine learning unit
The training sample of supervision machine study module, training the step for obtaining Supervised machine learning model include:
The exceptional sample and normal sample for extracting preset ratio out respectively, by the exceptional sample of extraction and normal sample together as having
The training sample of Supervised machine learning module in supervision machine unit, training obtain initial model；
Remaining exceptional sample and remaining normal sample verify the initial model together as verifying sample, test
Card is qualified, and the initial model is the Supervised machine learning model.
Further, described that each game user progress abnormality degree is beaten respectively by the Supervised machine learning model
Point, score value is more than that the game user of first threshold includes: labeled as script user
Respectively using the second behavioral data of each game user as the input parameter of the Supervised machine learning model, obtain
The abnormality degree marking score value of each game user；
Whether the abnormality degree marking score value for judging each game user respectively is more than first threshold, is more than by abnormality degree marking score value
The game user of first threshold is labeled as script user.
Further, the method also includes:
The game user that directly the first abnormal user is concentrated is labeled as script user.
The application's further aspect is that provide a kind of script user identifying system, comprising:
Rules unit: for being marked to obtain the first label to each game user respectively as a result, the first label result
Including the first normal users collection and the first abnormal user collection；
Unsupervised machine learning unit: for being marked to obtain the second label to each game user respectively as a result, described
Two label results include the second normal users collection, the second abnormal user collection and uncertain user collection；
Supervised machine learning unit: for being marked according to the first label result and second as a result, respectively to each game
User is marked to obtain third label as a result, third label result includes third normal users collection, third abnormal user
Collection and uncertain user collection；
The Supervised machine learning unit includes supervision machine study module, and the Supervised machine learning module is according to institute
It states third label result training and obtains Supervised machine learning model, the Supervised machine learning model is respectively to each game
User carries out abnormality degree marking, and score value is more than that the game user of first threshold is labeled as script user.
The mode that the present invention is combined by using unsupervised machine learning and Supervised machine learning to script user into
Row identification, adaptivity is good, and script user's recognition capability of system can be improved, and script user is allowed to be difficult to advise by summarizing
It restrains and the identification of avoidance system.
Detailed description of the invention
Fig. 1 is a kind of script user identification method flow chart shown according to an exemplary embodiment.
Fig. 2 is a kind of script user identifying system structural block diagram shown according to an exemplary embodiment.
Specific embodiment
It is with reference to the accompanying drawing and specific real in order to make those skilled in the art more fully understand technical solution of the present invention
Applying example, the present invention is described in further detail.
Embodiment
As shown in Figure 1, the present embodiment provides a kind of script user identification methods, comprising:
S1: each game user is marked to obtain the first label as a result, the first label knot respectively by rules unit
Fruit includes the first normal users collection and the first abnormal user collection；
S2: respectively each game user is marked to obtain the second label as a result, described by unsupervised machine learning unit
Second label result includes the second normal users collection, the second abnormal user collection and uncertain user collection；
S3: according to the first label result and the second label as a result, by Supervised machine learning unit respectively to each trip
Play user is marked to obtain third label as a result, third label result includes that third normal users collection and third are used extremely
Family collection；
S4: marking according to the third as a result, passing through the Supervised machine learning module training in Supervised machine learning unit
Obtain Supervised machine learning model；
S5: abnormality degree marking is carried out to each game user respectively by the Supervised machine learning model, score value is more than the
The game user of one threshold value is labeled as script user.
What needs to be explained here is that rules unit is used for the game user of real-time blocking Height Anomalies, first marked
Abnormal user collection can be directly labeled as script user, can be directly placed into blacklist to the script user marked, due to rule
Although then the mark mode accuracy rate of unit is high, there may be certain spill tag situations, so needing to combine subsequent nothing
Supervision machine unit and Supervised machine learning unit, in fact, rules unit and unsupervised machine learning unit are first
Game user is marked respectively in the way of oneself, Supervised machine learning unit marks result and nothing in rules unit
It is integrated to obtain the low third label knot of the high spill tag rate simultaneously of accuracy rate on the basis of supervision machine unit label result
Fruit marks according to third as a result, Supervised machine learning model is obtained by Supervised machine learning module training again, due to foot
This user is difficult to learn the marking rule of Supervised machine learning model, therefore is also difficult the identification of avoidance system, so that this reality
The script user identification method adaptivity for applying example offer is good, high to the recognition capability of script user.
Preferably, it is described by rules unit respectively each game user is marked to obtain the first label as a result,
That is step S1, comprising:
S11: obtaining login/log-on data of each game user respectively, includes that corresponding game is used in each login/log-on data
The ID at family；
S12: judging whether login/log-on data of each game user violates preset rules respectively, if so, being labeled as first
Otherwise abnormal user is labeled as the first normal users；
Wherein, all first normal users constitute the first normal users collection, and all first abnormal users constitute the first abnormal user
Collection.
Here, the marking convention of rules unit, which may is that, gets database first (all user data is in data
In library) in all users login/registion time, login/registration IP address, login/registration device id, according to getting
These data statistics go out IP login user quantity, IP registration number of users, equipment login user quantity, facility registration number of users
It measures, rule threshold is provided in rules unit, if the IP, device number (or device id) where the login of game user are certain
(IP login user quantity, IP register number of users, equipment login user quantity, facility registration user to statistical data in time
Quantity) it is more than rule threshold, then the game user is labeled as script user by rules unit, wherein rule threshold is to mention above
And a series of statistical indicators preset value, such as the continuous login user quantity of IP, the continuous login user quantity of equipment (continuously refer to
Every time log in interval time it is smaller, such as 10 seconds), IP go over one week registration number of users, equipment past one week registration number of users
Amount etc., every value of statistical indicant of normal game user should be respectively smaller than corresponding rule threshold.
Preferably, described be marked all game users to obtain the second label by unsupervised machine learning unit
As a result, i.e. step S2, comprising:
S21: obtaining the first behavioral data of each game user respectively, and first behavioral data is game user in game
Behavior topological data in different scene of game includes the ID of corresponding game user in each first behavioral data；
S22: according to the first behavioral data, carrying out the marking of suspicion degree to each game user using highly dense subgraph mining algorithm, point
Value is more than that the game user of second threshold is labeled as the second abnormal user, score value lower than third threshold value game user labeled as the
Two normal users, game user of the score value between second threshold and third threshold value are labeled as uncertain user；
Wherein, all second normal users constitute the second normal users collection, and all second abnormal users constitute the second abnormal user
Collection, all uncertain users constitute uncertain user's collection.
Here, the first behavioral data be behavior topological data of the game user in game, the first behavioral data it is specific
Data mode can be understood as game user and participate in different game contents in game, obtain the record of different stage properties, a use
Family corresponds to a plurality of data, and quantity is not fixed, and using highly dense subgraph mining algorithm, (higher-dimension isomery figure high density subgraph excavates class and calculates
Method, such as ZOOM, DCUBE etc.) behavior that game user each in game participation different scenes obtain different stage properties is abstracted into
It gives a mark for a higher-dimension isomery figure and to the suspicion degree of each game user.
Preferably, described mark according to the first label result and second as a result, passing through Supervised machine learning list
Member is marked to obtain third label as a result, i.e. step S3 to all game users, comprising:
According to the first label result and the second label as a result, the first normal users collection and the second normal users collection will be belonged to simultaneously
Game user is labeled as third normal users, will belong to the game user of the first abnormal user collection and the second abnormal user collection simultaneously
Labeled as third abnormal user, by belong to the first normal users collection and belong to the second abnormal user collection game user label the
Existing uncertain user is added in the user for belonging to the first abnormal user collection and belong to the second normal users collection by three abnormal users
Collection；
Wherein, all third normal users constitute third normal users collection, and all third abnormal users constitute third abnormal user
Collection.
For good understanding, as shown in the table, √ indicates normal users, × indicate abnormal user ,-indicate uncertain user.
Wherein, it is the first normal users and the first abnormal user that rules unit is corresponding；Corresponding unsupervised machine learning unit is
Two normal users, the second abnormal user and uncertain user；The corresponding third normal users of Supervised machine learning unit and
Three abnormal users, it is remaining for uncertain user.
| Game user | Rules unit marks result | Unsupervised machine learning unit marks result | Supervised machine learning unit marks result | Owning user collection |
| A | √ | — | — | Uncertain user's collection |
| B | √ | × | × | Third abnormal user collection |
| C | √ | √ | √ | Third normal users collection |
| D | √ | √ | √ | Third normal users collection |
| E | × | √ | — | Uncertain user's collection |
| F | × | × | × | Third abnormal user collection |
| G | × | × | × | Third abnormal user collection |
| H | × | — | — | Uncertain user's collection |
Preferably, described mark according to the third as a result, by having intendant in Supervised machine learning unit
The training of device study module obtains Supervised machine learning model, i.e. step S4, comprising:
S41: obtaining the second behavioral data of each game user respectively, and second behavioral data is game user in game
Behavioural characteristic data, each second behavioral data includes the ID of corresponding game user；
S42: concentrating the ID of each game user according to third abnormal user, obtains third abnormal user and concentrates each game user
The second behavioral data, third abnormal user concentrates the second behavioral data of all game users to constitute exceptional sample together；
S43: concentrating the ID of each game user according to third normal users, obtains third normal users and concentrates each game user
The second behavioral data, third normal users concentrate the second behavioral data of all game users to constitute normal sample together；
S44: by the exceptional sample and normal sample together as the Supervised machine learning in Supervised machine learning unit
The training sample of module, training obtain Supervised machine learning model.
Here, the second behavioral data refers to behavioural characteristic data of the game user in game, is after aggregation process
A series of obtained indexs (such as gold coin obtains quantity etc.) based on user's dimension, the corresponding data of a user.There is prison
Machine learning module is superintended and directed using machine learning algorithm training module, such as gradient boosted tree scheduling algorithm.
Preferably, it is described by the exceptional sample and normal sample together as having in Supervised machine learning unit
The training sample of supervision machine study module, training obtain Supervised machine learning model, i.e. step S44, comprising:
S441: extracting the exceptional sample and normal sample of preset ratio out respectively, together by the exceptional sample of extraction and normal sample
As the training sample of the Supervised machine learning module in Supervised machine learning unit, training obtains initial model；
S442: remaining exceptional sample and remaining normal sample test the initial model together as verifying sample
Card, verifying is qualified, and the initial model is the Supervised machine learning model.
Preferably, described respectively beat each game user progress abnormality degree by the Supervised machine learning model
Point, score value is more than that the game user of first threshold is labeled as script user, i.e. step S5, comprising:
S51: respectively using the second behavioral data of each game user as the input parameter of the Supervised machine learning model,
Obtain the abnormality degree marking score value of each game user；
S52: whether the abnormality degree marking score value for judging each game user respectively is more than first threshold, and abnormality degree is given a mark score value
Game user more than first threshold is labeled as script user.
Preferably, the present embodiment further include:
The game user that directly the first abnormal user is concentrated is labeled as script user.
Embodiment 2
As shown in Fig. 2, the present embodiment provides a kind of script user identifying systems, comprising:
Rules unit: for being marked to obtain the first label to each game user respectively as a result, the first label result
Including the first normal users collection and the first abnormal user collection；
Unsupervised machine learning unit: for being marked to obtain the second label to each game user respectively as a result, described
Two label results include the second normal users collection, the second abnormal user collection and uncertain user collection；
Supervised machine learning unit: for being marked according to the first label result and second as a result, respectively to each game
User is marked to obtain third label as a result, third label result includes third normal users collection, third abnormal user
Collection and uncertain user collection；
The Supervised machine learning unit includes supervision machine study module, and the Supervised machine learning module is according to institute
It states third label result training and obtains Supervised machine learning model, the Supervised machine learning model is respectively to each game
User carries out abnormality degree marking, and score value is more than that the game user of first threshold is labeled as script user.
Concrete operations mode about script user's identifying system has been described in detail in embodiment of the method, this
Place is not set forth in detail.
The above is only the preferred embodiment of the present invention, it is noted that above-mentioned preferred embodiment is not construed as pair
Limitation of the invention, protection scope of the present invention should be defined by the scope defined by the claims..For the art
For those of ordinary skill, without departing from the spirit and scope of the present invention, several improvements and modifications can also be made, these change
It also should be regarded as protection scope of the present invention into retouching.
Claims (9)
1. script user identification method characterized by comprising
Each game user is marked to obtain the first label as a result, the first label result packet respectively by rules unit
Include the first normal users collection and the first abnormal user collection；
Each game user is marked to obtain the second label as a result, described second respectively by unsupervised machine learning unit
Marking result includes the second normal users collection, the second abnormal user collection and uncertain user collection；
According to the first label result and the second label as a result, being used respectively each game by Supervised machine learning unit
Family is marked to obtain third label as a result, third label result includes third normal users collection and third abnormal user
Collection；
It is marked according to the third as a result, being obtained by the Supervised machine learning module training in Supervised machine learning unit
Supervised machine learning model；
Abnormality degree marking is carried out to each game user respectively by the Supervised machine learning model, score value is more than the first threshold
The game user of value is labeled as script user.
2. script user identification method according to claim 1, which is characterized in that it is described by rules unit respectively to every
A game user is marked to obtain the first label result
Login/log-on data of each game user is obtained respectively, includes corresponding game user in each login/log-on data
ID；
Judge whether login/log-on data of each game user violates preset rules respectively, if so, abnormal labeled as first
Otherwise user is labeled as the first normal users；
Wherein, all first normal users constitute the first normal users collection, and all first abnormal users constitute the first abnormal user
Collection.
3. script user identification method according to claim 2, which is characterized in that described to pass through unsupervised machine learning list
Member is marked to obtain the second label result to all game users
The first behavioral data of each game user is obtained respectively, and first behavioral data is that game user is different in game
Behavior topological data in scene of game includes the ID of corresponding game user in each first behavioral data；
According to the first behavioral data, the marking of suspicion degree is carried out to each game user using highly dense subgraph mining algorithm, score value is super
The game user of second threshold is crossed labeled as the second abnormal user, score value is being labeled as second just lower than the game user of third threshold value
Common family, game user of the score value between second threshold and third threshold value are labeled as uncertain user；
Wherein, all second normal users constitute the second normal users collection, and all second abnormal users constitute the second abnormal user
Collection, all uncertain users constitute uncertain user's collection.
4. script user identification method according to claim 3, which is characterized in that described according to the first label result
With the second label as a result, all game users are marked by Supervised machine learning unit to obtain third label result packet
It includes:
According to the first label result and the second label as a result, the first normal users collection and the second normal users collection will be belonged to simultaneously
Game user is labeled as third normal users, will belong to the game user of the first abnormal user collection and the second abnormal user collection simultaneously
Labeled as third abnormal user, by belong to the first normal users collection and belong to the second abnormal user collection game user label the
Existing uncertain user is added in the user for belonging to the first abnormal user collection and belong to the second normal users collection by three abnormal users
Collection；
Wherein, all third normal users constitute third normal users collection, and all third abnormal users constitute third abnormal user
Collection.
5. script user identification method according to claim 4, which is characterized in that described marked according to the third is tied
Fruit obtains Supervised machine learning model packet by the Supervised machine learning module training in Supervised machine learning unit
It includes:
The second behavioral data of each game user is obtained respectively, and second behavioral data is row of the game user in game
Data are characterized, each second behavioral data includes the ID of corresponding game user；
The ID that each game user is concentrated according to third abnormal user obtains third abnormal user concentrates each game user
Two behavioral datas, third abnormal user concentrate the second behavioral data of all game users to constitute exceptional sample together；
The ID that each game user is concentrated according to third normal users obtains third normal users concentrate each game user
Two behavioral datas, third normal users concentrate the second behavioral data of all game users to constitute normal sample together；
By the exceptional sample and normal sample together as the Supervised machine learning module in Supervised machine learning unit
Training sample, training obtain Supervised machine learning model.
6. script user identification method according to claim 5, which is characterized in that described by the exceptional sample and normal
Sample trains together as the training sample of the Supervised machine learning module in Supervised machine learning unit and has obtained supervision
The step for machine learning model includes:
The exceptional sample and normal sample for extracting preset ratio out respectively, by the exceptional sample of extraction and normal sample together as having
The training sample of Supervised machine learning module in supervision machine unit, training obtain initial model；
Remaining exceptional sample and remaining normal sample verify the initial model together as verifying sample, test
Card is qualified, and the initial model is the Supervised machine learning model.
7. script user identification method according to claim 5, which is characterized in that described to have supervision machine by described
It practises model and abnormality degree marking is carried out to each game user respectively, score value is more than that the game user of first threshold is used labeled as script
Family includes:
Respectively using the second behavioral data of each game user as the input parameter of the Supervised machine learning model, obtain
The abnormality degree marking score value of each game user；
Whether the abnormality degree marking score value for judging each game user respectively is more than first threshold, is more than by abnormality degree marking score value
The game user of first threshold is labeled as script user.
8. script user identification method according to claim 1, which is characterized in that further include:
The game user that directly the first abnormal user is concentrated is labeled as script user.
9. script user's identifying system characterized by comprising
Rules unit: for being marked to obtain the first label to each game user respectively as a result, the first label result
Including the first normal users collection and the first abnormal user collection；
Unsupervised machine learning unit: for being marked to obtain the second label to each game user respectively as a result, described
Two label results include the second normal users collection, the second abnormal user collection and uncertain user collection；
Supervised machine learning unit: for being marked according to the first label result and second as a result, respectively to each game
User is marked to obtain third label as a result, third label result includes third normal users collection, third abnormal user
Collection and uncertain user collection；
The Supervised machine learning unit includes supervision machine study module, and the Supervised machine learning module is according to institute
It states third label result training and obtains Supervised machine learning model, the Supervised machine learning model is respectively to each game
User carries out abnormality degree marking, and score value is more than that the game user of first threshold is labeled as script user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910658519.1A CN110152306B (en) | 2019-07-22 | 2019-07-22 | Script user identification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910658519.1A CN110152306B (en) | 2019-07-22 | 2019-07-22 | Script user identification method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110152306A true CN110152306A (en) | 2019-08-23 |
| CN110152306B CN110152306B (en) | 2019-11-15 |
Family
ID=67637865
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910658519.1A Active CN110152306B (en) | 2019-07-22 | 2019-07-22 | Script user identification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110152306B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111400357A (en) * | 2020-02-21 | 2020-07-10 | 中国建设银行股份有限公司 | Method and device for identifying abnormal login |
| CN111773732A (en) * | 2020-09-04 | 2020-10-16 | 完美世界(北京)软件科技发展有限公司 | Target game user detection method, device and equipment |
| US11541316B2 (en) | 2020-12-29 | 2023-01-03 | Acer Incorporated | Electronic device and method for detecting abnormal device operation |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101059830A (en) * | 2007-06-01 | 2007-10-24 | 华南理工大学 | Recognition method for robot external capable of combining game characteristic |
| WO2014014232A1 (en) * | 2012-07-20 | 2014-01-23 | Choi Gyu-Bin | Apparatus and method for captcha of gaming method |
| CN103825780A (en) * | 2014-02-26 | 2014-05-28 | 珠海市君天电子科技有限公司 | Tag-on program identification method, service and system |
| CN107185243A (en) * | 2017-04-26 | 2017-09-22 | 厦门吉比特网络技术股份有限公司 | A kind of recognition methods of network game external store |
| CN108096837A (en) * | 2016-11-25 | 2018-06-01 | 盛趣信息技术（上海）有限公司 | Game robot dynamic identifying method |
-
2019
- 2019-07-22 CN CN201910658519.1A patent/CN110152306B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101059830A (en) * | 2007-06-01 | 2007-10-24 | 华南理工大学 | Recognition method for robot external capable of combining game characteristic |
| WO2014014232A1 (en) * | 2012-07-20 | 2014-01-23 | Choi Gyu-Bin | Apparatus and method for captcha of gaming method |
| CN103825780A (en) * | 2014-02-26 | 2014-05-28 | 珠海市君天电子科技有限公司 | Tag-on program identification method, service and system |
| CN108096837A (en) * | 2016-11-25 | 2018-06-01 | 盛趣信息技术（上海）有限公司 | Game robot dynamic identifying method |
| CN107185243A (en) * | 2017-04-26 | 2017-09-22 | 厦门吉比特网络技术股份有限公司 | A kind of recognition methods of network game external store |
Non-Patent Citations (2)
| Title |
|---|
| KUAN-TA CHEN 等: "Game Bot Detection Based on Avatar Trajectory", 《ICEC"08 PROCEEDINGS OF THE 7TH INTERNATIONAL CONFERENCE ON ENTERTAINMENT COMPUTING》 * |
| 门昌骞 等: "一种基于多学习器标记的半监督SVM学习方法", 《广西师范大学学报(自然科学版)》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111400357A (en) * | 2020-02-21 | 2020-07-10 | 中国建设银行股份有限公司 | Method and device for identifying abnormal login |
| CN111773732A (en) * | 2020-09-04 | 2020-10-16 | 完美世界(北京)软件科技发展有限公司 | Target game user detection method, device and equipment |
| CN111773732B (en) * | 2020-09-04 | 2021-01-08 | 完美世界(北京)软件科技发展有限公司 | Target game user detection method, device and equipment |
| US11541316B2 (en) | 2020-12-29 | 2023-01-03 | Acer Incorporated | Electronic device and method for detecting abnormal device operation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110152306B (en) | 2019-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110152306B (en) | Script user identification method and system | |
| US10193772B1 (en) | User behavior analyzer | |
| Oleson et al. | Programmatic gold: Targeted and scalable quality assurance in crowdsourcing | |
| CN105809190B (en) | A kind of SVM cascade classifier methods based on Feature Selection | |
| CN105468760B (en) | The method and apparatus that face picture is labeled | |
| CN108616491A (en) | A kind of recognition methods of malicious user and system | |
| CN109871954A (en) | Training sample generation method, method for detecting abnormality and device | |
| CN107506793A (en) | Clothes recognition methods and system based on weak mark image | |
| CN104318562A (en) | Method and device for confirming quality of internet images | |
| CN105915960A (en) | User type determination method and device | |
| CN103309857B (en) | A kind of taxonomy determines method and apparatus | |
| CN109743286A (en) | A kind of IP type mark method and apparatus based on figure convolutional neural networks | |
| CN104992482A (en) | Athletic competition data processing system and method thereof | |
| CN107133892A (en) | The real-time estimating method and system of a kind of network Piano lesson | |
| CN110090446A (en) | Method and apparatus for handling the report information in game | |
| CN102984269A (en) | Method and device for peer-to-peer flow identification | |
| CN107483267A (en) | A kind of EIGRP routing failures recognition methods | |
| Gelman et al. | Acting the Same Differently: A Cross-Course Comparison of User Behavior in MOOCs. | |
| CN107633470A (en) | A kind of artistic Computer Examination Evaluating System and method based on video identification | |
| CN111957047A (en) | Checkpoint configuration data adjusting method, computer equipment and storage medium | |
| CN114553596B (en) | Multi-dimensional security condition real-time display method and system suitable for network security | |
| CN112363923A (en) | Test method, device, computer equipment and medium based on questionnaire system | |
| CN109450876A (en) | A kind of DDos recognition methods and system based on various dimensions state-transition matrix feature | |
| CN106383738A (en) | Task processing method and distributed computing framework | |
| CN109472694A (en) | A kind of suspicious trading activity discovery system |
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 |