CN110189151A - A kind of account detection method and relevant device - Google Patents
A kind of account detection method and relevant device Download PDFInfo
- Publication number
- CN110189151A CN110189151A CN201910507060.5A CN201910507060A CN110189151A CN 110189151 A CN110189151 A CN 110189151A CN 201910507060 A CN201910507060 A CN 201910507060A CN 110189151 A CN110189151 A CN 110189151A
- Authority
- CN
- China
- Prior art keywords
- account
- processed
- detection data
- behavioural characteristic
- behavior
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
- G06Q30/0185—Product, service or business identity fraud
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Accounting & Taxation (AREA)
- Probability & Statistics with Applications (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Strategic Management (AREA)
- Finance (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The present invention provides a kind of account detection method and relevant device, after the behavioral data for obtaining account to be processed, behavioral data based on account to be processed, obtain a plurality of types of behavioural characteristics of account to be processed, and a plurality of types of behavioural characteristics based on account to be processed, obtain the feature vector of account to be processed, feature vector based on account to be processed and the account number classification detection data cluster constructed in advance, determine the account type of account to be processed, each type of behavioural characteristic shows a kind of behavior generated based on account to be processed in the behavioural characteristic of many of type, so that the feature vector of account to be processed many-sided can consider the behavior of account to be processed, so as to improve the accuracy of account type detection, and it can be based on the feature vector for embodying the behavioural characteristic after changing when the behavioural characteristic of abnormal account changes Account number classification detection data cluster is rebuild, abnormal account rule is modified without artificial, improves the flexibility of account detection.
Description
Technical field
The invention belongs to data analysis technique fields, more specifically, more particularly to a kind of account detection method and correlation set
It is standby.
Background technique
User can become the member of the multimedia platform of each network company by register account number mode at present, more to watch
Multimedia messages on media platform, and each network company is to promote itself multimedia platform, can pass through
Activity makes user become member as sold the mode of activation code, wherein becoming the cost of member far below straight by activation code mode
Connecing other becomes the cost of member's mode.
Illegal users some in the case can pass through some account or a large amount of activation of multiple accounts purchase during activity
Code, is then sold to other users for the activation code of purchase again to make profit.In order to hit this illegal act, can be gone by analysis
Determine whether account is abnormal account for data mode, the account can if it is be forbidden to obtain activation code, wherein by point
Analysis behavioral data mode determines whether account is that the mode of abnormal account has: by analyze behavioral data obtain account behavior it is special
Levy vector, such as by analyze behavioral data obtain account purchase activation code behavioural characteristic vector, such as by behavioural characteristic to
Amount may indicate that account in the activation number of codes that section at the same time is bought and buy between the different time sections of activation code
Interval etc., if behavioural characteristic vector meets preset abnormal account rule, it is determined that the account is abnormal account.
Wherein preset abnormal account rule is obtained by carrying out analysis for several times to activation code buying behavior, if activation code
Buying behavior changes, and needs to modify to preset abnormal account rule, thus illustrates existing account detection method
Flexibility is lower, and the behavioural characteristic of the data cover for detecting account exception is very little, can reduce the accuracy of detection.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of account detection method and relevant devices, for improving account
The accuracy and flexibility of detection.Technical solution is as follows:
The present invention provides a kind of account detection method, which comprises
Obtain the behavioral data of account to be processed；
Based on the behavioral data of the account to be processed, a plurality of types of behavioural characteristics of the account to be processed are obtained,
And a plurality of types of behavioural characteristics based on the account to be processed, the feature vector of the account to be processed is obtained, wherein institute
It states each type of behavioural characteristic in a plurality of types of behavioural characteristics and shows a kind of behavior generated based on the account to be processed；
Feature vector based on the account to be processed and the account number classification detection data cluster constructed in advance, determine it is described to
The account type for handling account, wherein the account number classification detection data cluster is the feature vector clusters for having account based on history
It obtains.
Preferably, a plurality of types of behavioural characteristics include: registration behavioural characteristic, activation behavioural characteristic, login behavior
At least one of feature and viewing behavioural characteristic；
The registration behavioural characteristic activates behavioural characteristic, logs in any one in behavioural characteristic and viewing behavioural characteristic
Behavioural characteristic includes: the associated each behavioral parameters of the affiliated behavior of behavior feature and the corresponding each row of each behavioral parameters
For subparameter.
Preferably, the feature vector based on the account to be processed and the account number classification detection data constructed in advance
Cluster determines that the account type of the account to be processed includes:
Obtain the feature vector of the reference sample of each account number classification detection data cluster；
The feature vector of feature vector and each reference sample based on the account to be processed determines described wait locate
Manage account number classification detection data cluster belonging to account；
If account number classification detection data cluster belonging to the account to be processed is abnormal clusters, it is determined that the account to be processed
Number account type be Exception Type；If account number classification detection data cluster belonging to the account to be processed is normal clusters,
The account type for determining the account to be processed is normal type.
Preferably, the preparatory building process of the account number classification detection data cluster includes:
Obtain the behavioral data that each history in account set has account；
Have the behavioral data of account based on each history, obtains a plurality of types of behaviors spy that each history has account
Sign, and a plurality of types of behavioural characteristics based on the existing account of each history, obtain the feature vector that each history has account；
The similarity for having the feature vector of account based on each history is clustered, and the account number classification testing number is obtained
According to cluster；
It obtains the history that account type in the account number classification detection data cluster is Exception Type and has account in the account
Accounting in number classification detection data cluster；
If the history that account type is Exception Type has accounting of the account in the account number classification detection data cluster
Greater than preset ratio, determine that the account number classification detection data cluster is abnormal clusters.
Preferably, after the account type for determining the account to be processed, the method also includes: based on described wait locate
The similarity that each history in the feature vector and the account set of account has the feature vector of account is managed, to described wait locate
Reason account and each history have account and are clustered, and the account number classification detection data cluster are retrieved, to the account
Classification detection data cluster is updated.
The present invention also provides a kind of account detection device, described device includes:
Data acquiring unit, for obtaining the behavioral data of account to be processed；
Vector obtains unit, for the behavioral data based on the account to be processed, obtains the more of the account to be processed
The behavioural characteristic of seed type, and a plurality of types of behavioural characteristics based on the account to be processed, obtain the account to be processed
Feature vector, wherein each type of behavioural characteristic shows based on the account to be processed in a plurality of types of behavioural characteristics
Number generate a kind of behavior；
Determination unit, the account number classification detection data constructed for the feature vector based on the account to be processed and in advance
Cluster determines the account type of the account to be processed, wherein the account number classification detection data cluster is that have account based on history
Feature vector clusters obtain.
Preferably, a plurality of types of behavioural characteristics include: registration behavioural characteristic, activation behavioural characteristic, login behavior
At least one of feature and viewing behavioural characteristic；
The registration behavioural characteristic activates behavioural characteristic, logs in any one in behavioural characteristic and viewing behavioural characteristic
Behavioural characteristic includes: the associated each behavioral parameters of the affiliated behavior of behavior feature and the corresponding each row of each behavioral parameters
For subparameter.
Preferably, the vector obtains unit, is also used to obtain the reference sample of each account number classification detection data cluster
This feature vector；
The determination unit, specifically for based on the account to be processed feature vector and each reference sample
Feature vector determines account number classification detection data cluster belonging to the account to be processed, if belonging to the account to be processed
Account number classification detection data cluster is abnormal clusters, it is determined that the account type of the account to be processed is Exception Type；If described
Account number classification detection data cluster belonging to account to be processed is normal clusters, it is determined that the account type of the account to be processed is positive
Normal type.
Preferably, described device further include: cluster cell, for feature vector based on the account to be processed and described
Each history has the similarity of the feature vector of account in account set, has account to the account to be processed and each history
It number is clustered, the account number classification detection data cluster is retrieved, to be updated to the account number classification detection data cluster.
The present invention also provides a kind of storage medium, one or more computer program generations are stored in the storage medium
Code, one or more of computer program codes realize above-mentioned account detection method when being run.
The present invention also provides a kind of server, the server includes memory and processor, is stored in the memory
There are one or more computer program codes, one or more of computer program codes are realized when being executed by the processor
Above-mentioned account detection method.
From above-mentioned technical proposal it is found that after the behavioral data for obtaining account to be processed, the behavior based on account to be processed
Data obtain a plurality of types of behavioural characteristics of account to be processed, and a plurality of types of behavioural characteristics based on account to be processed,
The feature vector of account to be processed is obtained, feature vector based on account to be processed and the account number classification detection data constructed in advance
Cluster determines the account type of account to be processed, and each type of behavioural characteristic shows base in the behavioural characteristic of many of type
In account to be processed generate a kind of behavior, therefore can the behavioural characteristic based on account to be processed from many aspects to account to be processed
Number behavior be described and analyze, enable the many-sided row for considering account to be processed of the feature vector of account to be processed
For, so as to improve the accuracy of account type detection, and can be based on preparatory when determining the account type of account to be processed
The account number classification detection data cluster of building determines, can be based on embodying variation when the behavioural characteristic of abnormal account changes
The feature vector of behavioural characteristic afterwards rebuilds account number classification detection data cluster, modifies abnormal account rule without artificial
Then, the flexibility of account detection is improved.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is the present invention
Some embodiments for those of ordinary skill in the art without creative efforts, can also basis
These attached drawings obtain other attached drawings.
Fig. 1 is a kind of flow chart of account detection method provided in an embodiment of the present invention；
Fig. 2 is the flow chart of another account detection method provided in an embodiment of the present invention；
Fig. 3 is a kind of structural schematic diagram of account detection device provided in an embodiment of the present invention；
Fig. 4 is the structural schematic diagram of another account detection device provided in an embodiment of the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
Referring to Fig. 1, it illustrates a kind of flow charts of account detection method provided in an embodiment of the present invention, for passing through
The behavioral data of account to be processed determines the account type of account to be processed, specifically, account detection method shown in Fig. 1 can wrap
Include following steps:
S101: the behavioral data of account to be processed is obtained.Wherein the behavioral data of account to be processed is for obtaining wait locate
The data basis of the account type foundation of account is managed, behavioral data shows that the behavior that account to be processed occurs within a certain period of time is special
Sign, such as the behavioral data of account to be processed is the data generated when being occurred based on behavior performed by account to be processed, behavior
Etc., as account sometime, a certain place be directed to the behavior of certain an object, and account to be processed can be in any software
Used in account, such as logging in the account of some video software, then the account for being used to log in some video software
Behavioral data can be the data that video software generation is logged in by the account, and the data etc. of video are browsed by the account.
In the present embodiment, the behavioral data of account to be processed can include but is not limited to the user behaviors log of account to be processed
In data, and account to be processed is on different devices using can lead to the behavior number of account to be processed because equipment is arranged
According to format may be different, therefore need after the behavioral data for obtaining account to be processed the behavioral data to account to be processed
It is pre-processed, it, in this way can be in order to subsequent true at least to guarantee that the format of behavioral data of different accounts to be processed is consistent
The account type of fixed account to be processed.Such as after getting behavioral data in the user behaviors log of account to be processed, to behavior number
According to being pre-processed.
Pretreated process can be but not limited to: the format of the behavioral data of account to be processed is converted to default lattice
Formula, and the abnormal behaviour data in the behavioral data of account to be processed are deleted, so-called abnormal behaviour data include but is not limited to not
Meet behavior data corresponding types value require data, by taking the time as an example, the format of time have certain year in such a month, and on such a day,
The formats such as Year/Month/Day, date, year. month. day, year-month-day are obtaining such behavioral data thus, are needing to carry out
Uniform format is extracted from the data with unified format so that can be ensured of when subsequent characteristics are extracted, and is improved feature and is mentioned
The accuracy taken, such as the preset format of time is year. month. day, then after obtaining time such behavioral data, needs
The time for not having preset format is converted to the time of the preset format, such as 20180901 and 2018 on Septembers, 1, format
It is converted to the preset format of 2017.09.01.And for behavioral data such for the time, the time is usually to correspond to
Some numerical value, and the numerical value cannot be negative value, so if it is negative value that the value of time this behavioral data, which is empty or value,
Then think that behavior data are abnormal behaviour data, needs to delete behavior data.
S102: the behavioral data based on account to be processed obtains a plurality of types of behavioural characteristics of account to be processed, and base
In a plurality of types of behavioural characteristics of account to be processed, the feature vector of account to be processed, the behavior of many of type are obtained
Each type of behavioural characteristic shows a kind of behavior generated based on account to be processed in feature, it is possible thereby to from different types of
Account to be processed is described in behavioural characteristic.
In order to make different types of behavioural characteristic carry out accurate description to account to be processed, inventor is to behavior spy
Sign is repeatedly analyzed, such as it is analysis in the case of the account obtained by using activation code that inventor, which analyzes account to be processed,
Using the behavioral data of a period of time before and after activation code (such as a few days ago with latter two days), finds abnormal account and be different from normal account
Number behavioural characteristic, the account that notes abnormalities is in place of being different from normal account: abnormal account can be in time, sky in activation
Between, there is apparent aggregation characteristic in the dimensions such as device parameter, such as abnormal account can some period carry out activation or
The device parameter used when activating account: ip (Internet Protocol Address, internet protocol address), useragent
(user agent, abbreviation ua), device type are closely similar or even identical, therefore can be using activation behavioural characteristic as a kind of behavior
Feature can mention to analyze account type for such behavioural characteristic from dimensions such as time, space, device parameters
Take feature.Furthermore inventors have found that the row of these types can also be extracted from registering, logging in and the angle of subsequent service condition
It is characterized the middle feature that can be showed and significantly build up phenomenon.
Based on this, in the present embodiment, a plurality of types of behavioural characteristics include: registration behavioural characteristic, activation behavioural characteristic,
At least one of behavioural characteristic and viewing behavioural characteristic are logged in, wherein registration behavioural characteristic shows that account to be processed was registering
Behavioural characteristic in journey, activation behavioural characteristic show that account to be processed in the behavioural characteristic of activation member's permission process, logs in row
It is characterized and shows account to be processed in the behavioural characteristic of login account process, viewing behavioural characteristic shows account to be processed in viewing
Behavioural characteristic in the process.
And register any one in behavioural characteristic, activation behavioural characteristic, login behavioural characteristic and viewing behavioural characteristic
Behavioural characteristic includes: the associated each behavioral parameters of the affiliated behavior of behavior feature and the corresponding each row of each behavioral parameters
For subparameter, wherein the associated each behavioral parameters of the affiliated behavior of behavioural characteristic are for showing that the affiliated behavior of behavior feature is examined
The aspect of amount, to embody corresponding behavior by each behavioral parameters, the similarly corresponding each behavior of each behavioral parameters
Subparameter is used to show the aspect that a style of writing parameter is considered, to state corresponding behavioral parameters by behavior subparameter,
Corresponding behavior is further embodied from the different aspect of a behavioral parameters.
Such as registration behavioural characteristic includes that at least one behavioral parameters can be but not limited to: registion time, registered place,
Register ip, registration ua, registration mailbox and registration cell-phone number, thus, it is possible to from registion time, registered place, registration ip, registration ua,
Registration mailbox embodies the registration behavior of account to be processed with registration cell-phone number；Activation behavioural characteristic similarly includes at least one
Behavioral parameters can be but not limited to: the ua used when the ip that uses when activation, activation, when city where when activation and activation
Between；Logging in behavioural characteristic includes that at least one behavioral parameters can be but not limited to: logging in ip, logs in ua, debarkation point, login
When the equipment that uses and login time；Viewing behavioural characteristic includes that at least one behavioral parameters can be but not limited to: viewing ip,
Viewing ua, viewing time, viewing place and viewing video, to embody the activated row of account to be processed by above-mentioned behavioral parameters
For, log in behavior and viewing behavior.
For any behavioral parameters under each behavioural characteristic, it can also be extracted from least one behavior subparameter
Feature, i.e., any behavioral parameters include at least one behavior subparameter, below to registration behavioural characteristic, log in behavioural characteristic, swash
Each behavioral parameters in living behavioural characteristic and viewing behavioural characteristic are illustrated respectively, to each behavioral parameters never Tongfang
Embody the behavior of account to be processed in face:
Behavior subparameter in registration behavioural characteristic in any one behavioral parameters is as follows:
The behavior subparameter that registion time includes can be but not limited to: the account number number of same time registration and on the same day together
The account number number of one place registration, the behavior subparameter that registered place includes can be but not limited to: the account number of same place registration
Number, registration ip include to behavior subparameter can be but not limited to: the account number number of same ip registration and on the same day same ip is infused
The account number number of volume, the registration ua behavior subparameter that includes can be but not limited to: the account number number of same ua registration and on the same day together
The account number number of one ua registration, the behavior subparameter that registration mailbox includes can be but not limited to: the account of same mailbox suffix registration
Number and on the same day the account number number of same mailbox suffix registration, the behavior subparameter that registration cell-phone number includes can be but unlimited
In: the account number number that same cell-phone number ownership place is registered and the account number number that same cell-phone number ownership place is registered on the same day.
Behavior subparameter in logging in behavioural characteristic in any one behavioral parameters is as follows:
Logging in the behavior subparameter that ip includes can be but not limited to: log under login times, the ip number logged in, same ip
Account number number and the account number number that logs in of same ip and same ua, the behavior subparameter that login time includes can be but not limited to:
Whether the account number number and landing time that the Annual distribution that logs in, same time log in are abnormal, and the equipment that when login uses includes
Behavior subparameter can be but not limited to: whether the number of devices that logs in, beaching accommodation are that commonly used equipment and same equipment log in
Account number number, logging in the behavior subparameter that ua includes can be but not limited to: the account number number logged under the ua number and same ua that log in,
The behavior subparameter that debarkation point includes can be but not limited to: the city number that logs in, registration city and log in city whether one
Cause the account number number logged in same city.
Behavior subparameter in activation behavioural characteristic in any one behavioral parameters is as follows:
The behavior subparameter that the ip used when activation includes can be but not limited to: the account number number of same ip activation and same
The ip and account number number of same ua registration, the behavior subparameter that the ua that when activation uses includes can be but not limited to: same ua swashs
Account number number living, the behavior subparameter that city where when activation includes can be but not limited to: activation city and being logged in city and be
No consistent, activation city is with the account number number for registering the whether consistent and same place activation in city, behavior that activationary time includes
Parameter can be but not limited to: the account number number of same time activation.
Behavior subparameter in viewing behavior in any one behavioral parameters is as follows:
The behavior subparameter that viewing ip includes can be but not limited to: viewing under the ip number of viewing and on the same day same ip
Account number quantity, the behavior subparameter that viewing ua includes can be but not limited to: the ua number of viewing, viewing under same ua on the same day
The account number number of viewing under account number number and on the same day same ip and same ua, the behavior subparameter that the viewing time includes can be but not
Be limited to: the account number number of the Annual distribution of viewing and same time viewing, the behavior subparameter that viewing place includes can be but not
It is limited to: the distribution of the place of viewing, the ground points of viewing and the account number number of same city viewing on the same day, the row that viewing video includes
It can be but not limited to for subparameter: the video counts of viewing.
It is available to consider to obtain registration behavior spy from different aspect based on the explanation of above-mentioned a plurality of types of behavioural characteristics
Sign, activation behavioural characteristic, login behavioural characteristic and viewing behavioural characteristic, a kind of behavioral data based on account to be processed obtain
The mode of behavioural characteristic is: the value of the corresponding each behavior subparameter of each behavioral parameters of any of the above-described behavior is extracted, by
Each behavior subparameter under the same behavior value composition embody the behavior behavioural characteristic, with register behavioural characteristic as
Example, forms the registration behavioural characteristic for the value of each behavior subparameter under registration behavior.
After obtaining a variety of behavioural characteristics, a feature vector is obtained by these features, a kind of feature vector that obtains
Mode is: if the numeralization form that the value of behavior subparameter uses indicates, not needing to handle it, if behavior
The value of subparameter using nonumericization form indicate, then need by the value from nonumericization form expression be converted to
Numeralization form indicates, then these are formed to the feature of account to be processed with each behavior subparameter that the form of quantizing indicates
Vector, such as the feature vector of account to be processed is [A1,A2,B1,B2], A and B are a kind of behavioural characteristic, [A respectively1,A2] it is A
The corresponding vector of behavioural characteristic, A1It is the value of a behavioral parameters, if similarly the feature vector of account to be processed is [A11,A12,
A21,A22,B1,B2], A11It is a behavioral parameters A1The value of a corresponding behavior subparameter.
In the present embodiment, value is converted to from the expression of nonumericization form is in such a way that numeralization form indicates:
Coding form is used to indicate value in the form of quantizing, by taking the registered place of account to be processed as an example, it is assumed that account to be processed
Registered place is Shanghai, it is known that all registered places arrangement are as follows: Beijing, Shanghai, Guangzhou, Hebei etc., then this line of the registered place
Numeralization form for the value " Shanghai " of parameter indicates are as follows: 0100 ....
Furthermore after the feature vector for obtaining account to be processed, the feature vector of account to be processed can also be returned
One change processing (such as linear normalization, standard deviation normalize), so that the feature vector of different accounts to be processed is in same
Normalized process the present embodiment is no longer described in detail in the order of magnitude.
And feature vector is by registration behavioural characteristic, activation behavioural characteristic, logs in behavioural characteristic and viewing behavioural characteristic
A variety of behavioral parameters and a variety of behavior subparameters composition, allow feature vector to describe account to be processed from multi-angle
Behavior, improves the accuracy of description for single angle, and no matter the behavior of the unspecified angle in multi-angle becomes
Change, the present embodiment can also carry out the behavior of accurate description account to be processed from other angles, accordingly even when the behavior of some angle
It changes, still is able to analyze account to be processed from other angles.
In addition, except it is above-mentioned obtain the mode of feature vector in addition to, the present embodiment can also obtain feature by other means
Vector such as constructs a machine mould by the behavioral data of history account, the row of account to be processed is extracted by machine mould
It is characterized, then is made of the feature vector of account to be processed the behavioural characteristic of account to be processed, or by account to be processed and its
His account (history account and/or other accounts to be processed) constructs one and shows behavior relation between account itself behavior and account
Characteristic pattern, and then obtain from this feature figure the behavioural characteristic of account to be processed, it is then special by the behavior of account to be processed again
Sign forms the feature vector of account to be processed.
Herein it should be noted is that: in order to reduce the data volume for needing to analyze in account detection process, above-mentioned row
Being characterized can be the behavioral data based on a period of time and obtains, and such as activation behavioural characteristic, can be based on activation
The behavioral data that the front and back of account to be processed obtains for a period of time obtains, as front and back can be but not limited to a few days ago for a period of time
With latter two days, such as by activate activation code obtain for account mode, can based on activation activation code a few days ago and after
Two days behavioral datas obtain activation behavioural characteristic.For activation behavioural characteristic for there are a kind of situations: there is no activation to
The behavioral data before account is handled, is also considered as a kind of behavioural characteristic in this case.It is special for other kinds of behavior
It will not be described for the corresponding extraction time the present embodiment of sign.
S103: feature vector based on account to be processed and the account number classification detection data cluster constructed in advance are determined wait locate
The account type of account is managed, wherein account number classification detection data cluster is that the feature vector clusters for being had account based on history are obtained.
In the present embodiment, determining a kind of feasible pattern of the account type of account to be processed is: obtaining each account class
The feature vector of the reference sample of other detection data cluster, the feature of feature vector and each reference sample based on account to be processed
Vector determines account number classification detection data cluster belonging to the feature vector of account to be processed, if the feature of account to be processed to
Account number classification detection data cluster belonging to amount is abnormal clusters, it is determined that the account type for handling account is Exception Type, if to
Handling account number classification detection data cluster belonging to the feature vector of account is normal clusters, it is determined that the account type of account to be processed
For normal type, it is possible thereby to account number classification detection data cluster belonging to the feature vector based on account to be processed, determination is wait locate
Manage the account type of account.
Wherein the reference sample of each account number classification detection data cluster is that belong to can in the account number classification detection data cluster
The existing account of the type of the account number classification detection data cluster is embodied, for example, if the account number classification detection data cluster is one different
Regular data cluster, then can will belong to account type in the account number classification detection data cluster is that the existing account of Exception Type is considered as
The reference sample of the account number classification detection data cluster.And belong to phase in all existing accounts of the account number classification detection data cluster
As have account (the close existing account of similarity) can collect to the center of the account number classification detection data cluster, it is possible thereby to
The a certain range of existing account in center for being located at the account number classification detection data cluster is considered as account number classification detection data cluster
Reference sample, such as can will as the existing account of the central point of the account number classification detection data cluster be used as refer to sample
This.
The feature vector of the corresponding feature vector based on account to be processed and each reference sample, determines account to be processed
Number feature vector belonging to a kind of feasible pattern of account number classification detection data cluster be: the feature vector based on account to be processed
With the feature vector of each reference sample, calculate account to be processed to each reference sample distance, and select distance meet
The affiliated account number classification detection data cluster of reference sample of preset condition (as distance is minimum) is as account belonging to the account to be processed
Number classification detection data cluster.
In the present embodiment, the preparatory building process of account number classification detection data cluster is as follows:
1) behavioral data that each history in account set has account is obtained, wherein account set includes that account type is
Exception Type and account type are the existing account of normal type, and the existing account of both types is clustered as sample
Existing account to obtain account number classification detection data cluster, and then based on both types determines each account number classification testing number
According to the type of cluster.Have the explanation of the behavioral data of account for each history and obtain process and please refers to account to be processed
Behavioral data explanation, and will not be described here in detail；
2) have the behavioral data of account based on each history, obtain a plurality of types of behaviors that each history has account
Feature, and based on each history have account a plurality of types of behavioural characteristics, obtain each history have account feature to
Amount, a plurality of types of behavioural characteristics that wherein history has account, which can show that, has a kind of row that account generates based on the history
For the mode for obtaining the feature vector that each history has account please refers to the above-mentioned feature vector for obtaining account to be processed
Process no longer illustrates this present embodiment；
3) similarity for being had the feature vector of account based on each history is clustered, and account number classification detection data is obtained
Cluster, wherein the similarity that each history has the feature vector of account can pass through cosine similarity algorithm, Euclidean distance algorithm etc.
It obtains, and account is had to each history based on clustering algorithm and is clustered, such as based on K mean cluster algorithm or K arest neighbors point
Class algorithm has account to each history and clusters, and realizes the division for having account to each history, obtains account set pair
The multiple account number classification detection data clusters answered, and each history can be had account by cluster and be divided to corresponding account
In classification detection data cluster；
4) it obtains the history that account type in account number classification detection data cluster is Exception Type and has account in account number classification
Accounting in detection data cluster, if the history that account type is Exception Type has account in account number classification detection data cluster
Accounting be greater than preset ratio, determine account number classification detection data cluster be abnormal clusters, can basis for the setting of preset ratio
Depending on practical application, to this present embodiment without limiting.
Certain the present embodiment can also be the existing of normal type based on account type in the account number classification detection data cluster
Account, determines the type of account number classification detection data cluster, such as obtains account type in the account number classification detection data cluster and be positive
Accounting of the existing account of normal type in the account number classification detection data cluster, if account type is the existing account of normal type
Accounting number in the account number classification detection data cluster is greater than or equal to default accounting, then the class of the account number classification detection data cluster
Type is normal data cluster, and otherwise the type of the account number classification detection data cluster is abnormal data cluster, for presetting the setting of accounting
It can be depending on practical application, to this present embodiment without limiting.
From above-mentioned technical proposal it is found that after the behavioral data for obtaining account to be processed, the behavior based on account to be processed
Data obtain a plurality of types of behavioural characteristics of account to be processed, and a plurality of types of behavioural characteristics based on account to be processed,
The feature vector of account to be processed is obtained, feature vector based on account to be processed and the account number classification detection data constructed in advance
Cluster determines the account type of account to be processed, and each type of behavioural characteristic shows base in the behavioural characteristic of many of type
In account to be processed generate a kind of behavior, therefore can the behavioural characteristic based on account to be processed from many aspects to account to be processed
Number behavior be described and analyze, enable the many-sided row for considering account to be processed of the feature vector of account to be processed
For, so as to improve the accuracy of account type detection, and can be based on preparatory when determining the account type of account to be processed
The account number classification detection data cluster of building determines, can be based on embodying variation when the behavioural characteristic of abnormal account changes
The feature vector of behavioural characteristic afterwards rebuilds account number classification detection data cluster, modifies abnormal account rule without artificial
Then, the flexibility of account detection is improved.
Referring to Fig. 2, it is illustrated the embodiment of the invention provides the flow chart of another account detection method, for more
New at least two account number classifications detection data cluster, the account detection method may comprise steps of:
S201: the behavioral data of account to be processed is obtained.
S202: the behavioral data based on account to be processed obtains a plurality of types of behavioural characteristics of account to be processed, and base
In a plurality of types of behavioural characteristics of account to be processed, the feature vector of account to be processed is obtained, wherein account to be processed is more
Each type of behavioural characteristic shows a kind of behavior generated based on account to be processed in the behavioural characteristic of seed type.
S203: feature vector based on account to be processed and the account number classification detection data cluster constructed in advance are determined wait locate
The account type of account is managed, wherein account number classification detection data cluster is that the feature vector clusters for being had account based on history are obtained.
Above-mentioned steps S201 to S203 is identical as the implementation procedure of above-mentioned steps S101 to S103 and principle, no longer superfluous herein
It states.
S204: each history has the feature vector of account in feature vector and account set based on account to be processed
Similarity has account to account to be processed and each history and clusters, account number classification detection data cluster retrieved, with right
Account number classification detection data cluster is updated.
Such as account to be processed can be added in the account set for being used to construct account number classification detection data cluster, it is based on
The similarity that each history has the feature vector of account in the feature vector and account set of account to be processed is clustered, with
Again have account to each history in account set and account to be processed carries out subseries again, obtain account number classification testing number
According to cluster, or by an account in account to be processed replacement account set, the account replaced can have with account to be processed
There is same type, then the similarity of the feature vector again based on all accounts in updated account set is clustered,
Retrieve account number classification detection data cluster.
After retrieving account number classification detection data cluster through the above way, based on the account number classification inspection retrieved
Measured data cluster determines that the process of the account type of account to be processed please refers to the explanation in above-described embodiment.Why based on again
Obtained account number classification detection data cluster carry out detection be because are as follows: the behavioural characteristic that history in account set has account is one
A little historical behavior features, the historical behavior feature may be using changing for a period of time, and the behavior of account to be processed
Feature may be changed behavioural characteristic, and it is special new behavior can be added by way of introducing account to be processed in this way
Sign, obtains account number classification detection data cluster based on the new corresponding feature vector of behavioural characteristic, so that account class
Other detection data cluster may indicate that the rule of new behavioural characteristic, avoid the accuracy for reducing account detection.And when behavior is special
After sign changes, it is only necessary to the affiliated account of changed behavioural characteristic be added in account set and update account number classification
Detection data cluster improves the flexibility of account detection for existing artificial modification exception account rule.
For the various method embodiments described above, for simple description, therefore, it is stated as a series of action combinations, but
Be those skilled in the art should understand that, the present invention is not limited by the sequence of acts described because according to the present invention, certain
A little steps can be performed in other orders or simultaneously.Secondly, those skilled in the art should also know that, it is retouched in specification
The embodiment stated belongs to preferred embodiment, and related actions and modules are not necessarily necessary for the present invention.
Corresponding with above method embodiment, the embodiment of the present invention also provides a kind of account detection device, structure such as Fig. 3
It is shown, it may include: that data acquiring unit 10, vector obtain unit 20 and determination unit 30.
Data acquiring unit 10, for obtaining the behavioral data of account to be processed, for the behavioral data of account to be processed
Explanation please refer to above method embodiment, this present embodiment is no longer illustrated.
Vector obtains unit 20, for the behavioral data based on account to be processed, obtains the multiple types of account to be processed
Behavioural characteristic, and a plurality of types of behavioural characteristics based on account to be processed obtain the feature vector of account to be processed, wherein
Each type of behavioural characteristic shows a kind of behavior generated based on account to be processed in a plurality of types of behavioural characteristics, thus may be used
Account to be processed to be described from different types of behavioural characteristic, especially to the behavior generated based on account to be processed into
Row description.
In the present embodiment, a plurality of types of behavioural characteristics include: registration behavioural characteristic, activation behavioural characteristic, log in row
At least one of be characterized with viewing behavioural characteristic.It registers behavioural characteristic, activation behavioural characteristic, log in behavioural characteristic and viewing
Any one behavioural characteristic in behavioural characteristic includes: the associated each behavioral parameters of the affiliated behavior of behavior feature and each
The corresponding each behavior subparameter of behavioral parameters, with associated each behavioral parameters of the affiliated behavior of Behavior-based control feature and each
The corresponding each behavior subparameter of behavioral parameters can be stated from the various aspects of the behavior of embodiment, concrete behavior parameter and behavior
Which subparameter has please refer to above method embodiment.
Behavioral data of the one of which based on account to be processed, the mode for obtaining behavioural characteristic is: extracting any of the above-described row
For the corresponding each behavior subparameter of each behavioral parameters value, by taking for each behavior subparameter under the same behavior
Value composition embodies the behavioural characteristic of the behavior, for registering behavioural characteristic, by each behavior subparameter under registration behavior
Value forms the registration behavioural characteristic.
After obtaining a variety of behavioural characteristics, a feature vector is obtained by these features, a kind of feature vector that obtains
Mode is: if the numeralization form that the value of behavior subparameter uses indicates, not needing to handle it, if behavior
The value of subparameter using nonumericization form indicate, then need by the value from nonumericization form expression be converted to
Numeralization form indicates, then these are formed to the feature of account to be processed with each behavior subparameter that the form of quantizing indicates
Vector.
In the present embodiment, value is converted to from the expression of nonumericization form is in such a way that numeralization form indicates:
Coding form is used to indicate value in the form of quantizing, by taking the registered place of account to be processed as an example, it is assumed that account to be processed
Registered place is Shanghai, it is known that all registered places arrangement are as follows: Beijing, Shanghai, Guangzhou, Hebei etc., then this line of the registered place
Numeralization form for the value " Shanghai " of parameter indicates are as follows: 0100 ....
Furthermore after the feature vector for obtaining account to be processed, the feature vector of account to be processed can also be returned
One change processing (such as linear normalization, standard deviation normalize), so that the feature vector of different accounts to be processed is in same
Normalized process the present embodiment is no longer described in detail in the order of magnitude.
Determination unit 30, the account number classification detection data constructed for the feature vector based on account to be processed and in advance
Cluster determines the account type of account to be processed, wherein account number classification detection data cluster be had based on history the feature of account to
Amount cluster obtains.
In the present embodiment, determining a kind of feasible pattern of the account type of account to be processed is: vector obtains unit 20
Obtain the feature vector of the reference sample of each account number classification detection data cluster, feature of the determination unit 30 based on account to be processed
The feature vector of each reference sample of vector sum determines account number classification detection data belonging to the feature vector of account to be processed
Cluster, if account number classification detection data cluster belonging to the feature vector of account to be processed is abnormal clusters, it is determined that handle account
Account type is Exception Type, if account number classification detection data cluster belonging to the feature vector of account to be processed is normal clusters,
The account type for then determining account to be processed is normal type, it is possible thereby to account belonging to the feature vector based on account to be processed
Number classification detection data cluster, determines the account type of account to be processed.
Wherein the reference sample of each account number classification detection data cluster is that belong to can in the account number classification detection data cluster
The existing account of the type of the account number classification detection data cluster is embodied, for example, if the account number classification detection data cluster is one different
Regular data cluster, then can will belong to account type in the account number classification detection data cluster is that the existing account of Exception Type is considered as
The reference sample of the account number classification detection data cluster.And belong to phase in all existing accounts of the account number classification detection data cluster
As have account (the close existing account of similarity) can collect to the center of the account number classification detection data cluster, it is possible thereby to
The a certain range of existing account in center for being located at the account number classification detection data cluster is considered as account number classification detection data cluster
Reference sample, such as can will as the existing account of the central point of the account number classification detection data cluster be used as refer to sample
This.
The feature vector of feature vector and each reference sample of the corresponding determination unit 30 based on account to be processed, really
A kind of feasible pattern of account number classification detection data cluster belonging to the feature vector of fixed account to be processed is: based on account to be processed
Feature vector and each reference sample feature vector, calculate account to be processed to the distance of each reference sample, and choose
Distance meets the affiliated account number classification detection data cluster of reference sample of preset condition (as distance is minimum) as the account to be processed out
Account number classification detection data cluster belonging to number.
From above-mentioned technical proposal it is found that after the behavioral data for obtaining account to be processed, the behavior based on account to be processed
Data obtain a plurality of types of behavioural characteristics of account to be processed, and a plurality of types of behavioural characteristics based on account to be processed,
The feature vector of account to be processed is obtained, feature vector based on account to be processed and the account number classification detection data constructed in advance
Cluster determines the account type of account to be processed, and each type of behavioural characteristic shows base in the behavioural characteristic of many of type
In account to be processed generate a kind of behavior, therefore can the behavioural characteristic based on account to be processed from many aspects to account to be processed
Number behavior be described and analyze, enable the many-sided row for considering account to be processed of the feature vector of account to be processed
For, so as to improve the accuracy of account type detection, and can be based on preparatory when determining the account type of account to be processed
The account number classification detection data cluster of building determines, can be based on embodying variation when the behavioural characteristic of abnormal account changes
The feature vector of behavioural characteristic afterwards rebuilds account number classification detection data cluster, modifies abnormal account rule without artificial
Then, the flexibility of account detection is improved.
Referring to Fig. 4, it illustrates the structures of another account detection device provided in an embodiment of the present invention, in Fig. 3 base
Can also include: cluster cell 40 on plinth, for constructing account number classification detection data cluster in advance, preparatory building process is as follows:
1) behavioral data that each history in account set has account is obtained, wherein account set includes that account type is
Exception Type and account type are the existing account of normal type, and the existing account of both types is clustered as sample
To obtain account number classification detection data cluster；
2) have the behavioral data of account based on each history, obtain a plurality of types of behaviors that each history has account
Feature, and based on each history have account a plurality of types of behavioural characteristics, obtain each history have account feature to
Amount, a plurality of types of behavioural characteristics that wherein history has account, which can show that, has a kind of row that account generates based on the history
For the mode for obtaining the feature vector that each history has account please refers to the above-mentioned feature vector for obtaining account to be processed
Process no longer illustrates this present embodiment；
3) similarity for being had the feature vector of account based on each history is clustered, and account number classification detection data is obtained
Cluster；
4) it obtains the history that account type in account number classification detection data cluster is Exception Type and has account in account number classification
Accounting in detection data cluster, if the history that account type is Exception Type has account in account number classification detection data cluster
Accounting be greater than preset ratio, determine account number classification detection data cluster be abnormal clusters, can basis for the setting of preset ratio
Depending on practical application, to this present embodiment without limiting.
The process of above-mentioned preparatory building account number classification detection data cluster, please refers to embodiment of the method, not to this present embodiment
It illustrates again.
After the account type for determining account to be processed, cluster cell 40 be also used to the feature based on account to be processed to
Each history has the similarity of the feature vector of account in amount and account set, has account to account to be processed and each history
It number is clustered, retrieves account number classification detection data cluster, to be updated to account number classification detection data cluster, so as to
The detection of account type is carried out based on the account number classification detection data cluster retrieved.
Why carrying out detection based on the account number classification detection data cluster retrieved is because of are as follows: the history in account set
The behavioural characteristic of existing account is some historical behavior features, which may be become using a period of time
Change, and the behavioural characteristic of account to be processed may be changed behavioural characteristic, it in this way can be by introducing account to be processed
Mode new behavioural characteristic is added, allow the account number classification detection data cluster based on the new corresponding feature of behavioural characteristic
Vector obtains, so that account number classification detection data cluster may indicate that the rule of new behavioural characteristic, avoids reducing account detection
Accuracy.And after behavioural characteristic changes, it is only necessary to which account is added in the affiliated account of changed behavioural characteristic
Account number classification detection data cluster is updated in number set, for existing artificial modification exception account rule, is improved
The flexibility of account detection.
The embodiment of the present invention also provides a kind of storage medium, and one or more computer program generations are stored in storage medium
Code, one or more computer program codes realize above-mentioned account detection method when being run.
The embodiment of the present invention also provides a kind of server, and server includes memory and processor, is stored in memory
One or more computer program codes, one or more computer program codes realize above-mentioned account inspection when being executed by processor
Survey method.
It should be noted that all the embodiments in this specification are described in a progressive manner, each embodiment weight
Point explanation is the difference from other embodiments, and the same or similar parts between the embodiments can be referred to each other.
For device class embodiment, since it is basically similar to the method embodiment, so being described relatively simple, related place ginseng
See the part explanation of embodiment of the method.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning
Covering non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes that
A little elements, but also including other elements that are not explicitly listed, or further include for this process, method, article or
The intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence "including a ...", is not arranged
Except there is also other identical elements in the process, method, article or apparatus that includes the element.
The foregoing description of the disclosed embodiments can be realized those skilled in the art or using the present invention.To this
A variety of modifications of a little embodiments will be apparent for a person skilled in the art, and the general principles defined herein can
Without departing from the spirit or scope of the present invention, to realize in other embodiments.Therefore, the present invention will not be limited
It is formed on the embodiments shown herein, and is to fit to consistent with the principles and novel features disclosed in this article widest
Range.
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered
It is considered as protection scope of the present invention.
Claims (10)
1. a kind of account detection method, which is characterized in that the described method includes:
Obtain the behavioral data of account to be processed；
Based on the behavioral data of the account to be processed, a plurality of types of behavioural characteristics of the account to be processed, and base are obtained
In a plurality of types of behavioural characteristics of the account to be processed, the feature vector of the account to be processed is obtained, wherein described more
Each type of behavioural characteristic shows a kind of behavior generated based on the account to be processed in the behavioural characteristic of seed type；
Feature vector based on the account to be processed and the account number classification detection data cluster constructed in advance, determine described to be processed
The account type of account, wherein the account number classification detection data cluster is that the feature vector clusters for being had account based on history are obtained
It arrives.
2. the method according to claim 1, wherein a plurality of types of behavioural characteristics include: registration behavior
Feature, logs at least one of behavioural characteristic and viewing behavioural characteristic at activation behavioural characteristic；
Any one behavior in the registration behavioural characteristic, activation behavioural characteristic, login behavioural characteristic and viewing behavioural characteristic
Feature includes: the associated each behavioral parameters of the affiliated behavior of behavior feature and corresponding each behavior of each behavioral parameters
Parameter.
3. the method according to claim 1, wherein the feature vector based on the account to be processed and pre-
The account number classification detection data cluster first constructed determines that the account type of the account to be processed includes:
Obtain the feature vector of the reference sample of each account number classification detection data cluster；
The feature vector of feature vector and each reference sample based on the account to be processed, determines the account to be processed
Account number classification detection data cluster belonging to number；
If account number classification detection data cluster belonging to the account to be processed is abnormal clusters, it is determined that the account to be processed
Account type is Exception Type；If account number classification detection data cluster belonging to the account to be processed is normal clusters, it is determined that
The account type of the account to be processed is normal type.
4. the method according to claim 1, wherein the preparatory building process of the account number classification detection data cluster
Include:
Obtain the behavioral data that each history in account set has account；
Have the behavioral data of account based on each history, obtain a plurality of types of behavioural characteristics that each history has account,
And have a plurality of types of behavioural characteristics of account based on each history, obtain the feature vector that each history has account；
The similarity for having the feature vector of account based on each history is clustered, and the account number classification detection data is obtained
Cluster；
It obtains the history that account type in the account number classification detection data cluster is Exception Type and has account in the account class
Accounting in other detection data cluster；
If the history that account type is Exception Type has accounting of the account in the account number classification detection data cluster and is greater than
Preset ratio determines that the account number classification detection data cluster is abnormal clusters.
5. according to the method described in claim 4, it is characterized in that, after the account type for determining the account to be processed,
The method also includes: each history has account in feature vector and the account set based on the account to be processed
The similarity of feature vector has account to the account to be processed and each history and clusters, retrieves the account
Classification detection data cluster, to be updated to the account number classification detection data cluster.
6. a kind of account detection device, which is characterized in that described device includes:
Data acquiring unit, for obtaining the behavioral data of account to be processed；
Vector obtains unit, for the behavioral data based on the account to be processed, obtains the multiple types of the account to be processed
The behavioural characteristic of type, and a plurality of types of behavioural characteristics based on the account to be processed, obtain the spy of the account to be processed
Vector is levied, wherein each type of behavioural characteristic shows to produce based on the account to be processed in a plurality of types of behavioural characteristics
A kind of raw behavior；
Determination unit, the account number classification detection data cluster constructed for the feature vector based on the account to be processed and in advance,
The account type of the account to be processed is determined, wherein the account number classification detection data cluster is the spy for having account based on history
Sign vector clusters obtain.
7. device according to claim 6, which is characterized in that a plurality of types of behavioural characteristics include: registration behavior
Feature, logs at least one of behavioural characteristic and viewing behavioural characteristic at activation behavioural characteristic；
Any one behavior in the registration behavioural characteristic, activation behavioural characteristic, login behavioural characteristic and viewing behavioural characteristic
Feature includes: the associated each behavioral parameters of the affiliated behavior of behavior feature and corresponding each behavior of each behavioral parameters
Parameter.
8. device according to claim 6, which is characterized in that the vector obtains unit, is also used to obtain each described
The feature vector of the reference sample of account number classification detection data cluster；
The determination unit, specifically for the feature of feature vector and each reference sample based on the account to be processed
Vector determines account number classification detection data cluster belonging to the account to be processed, if account belonging to the account to be processed
Classification detection data cluster is abnormal clusters, it is determined that the account type of the account to be processed is Exception Type；If described wait locate
Managing account number classification detection data cluster belonging to account is normal clusters, it is determined that the account type of the account to be processed is normal class
Type.
9. device according to claim 6, which is characterized in that described device further include: cluster cell, for based on described
Each history has the similarity of the feature vector of account in the feature vector of account to be processed and the account set, to described
Account to be processed and each history have account and are clustered, and the account number classification detection data cluster are retrieved, to described
Account number classification detection data cluster is updated.
10. a kind of storage medium, which is characterized in that one or more computer program codes are stored in the storage medium,
The account detection as described in claim 1 to 5 any one is realized when one or more of computer program codes are run
Method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910507060.5A CN110189151A (en) | 2019-06-12 | 2019-06-12 | A kind of account detection method and relevant device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910507060.5A CN110189151A (en) | 2019-06-12 | 2019-06-12 | A kind of account detection method and relevant device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110189151A true CN110189151A (en) | 2019-08-30 |
Family
ID=67721483
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910507060.5A Pending CN110189151A (en) | 2019-06-12 | 2019-06-12 | A kind of account detection method and relevant device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110189151A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110569509A (en) * | 2019-09-12 | 2019-12-13 | 广州荔支网络技术有限公司 | risk group identification method and device |
| CN110674479A (en) * | 2019-09-29 | 2020-01-10 | 武汉极意网络科技有限公司 | Abnormal behavior data real-time processing method, device, equipment and storage medium |
| CN111831825A (en) * | 2020-07-23 | 2020-10-27 | 咪咕文化科技有限公司 | Account detection method, account detection device, network equipment and storage medium |
| CN113011886A (en) * | 2021-02-19 | 2021-06-22 | 腾讯科技（深圳）有限公司 | Method and device for determining account type and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1236173A2 (en) * | 1999-10-27 | 2002-09-04 | Biowulf Technologies, LLC | Methods and devices for identifying patterns in biological systems |
| US20160253750A1 (en) * | 2014-12-22 | 2016-09-01 | Palantir Technologies Inc. | Systems and user interfaces for dynamic and interactive investigation of bad actor behavior based on automatic clustering of related data in various data structures |
| CN106886518A (en) * | 2015-12-15 | 2017-06-23 | 国家计算机网络与信息安全管理中心 | A kind of method of microblog account classification |
| CN104283737B (en) * | 2014-09-30 | 2018-01-12 | 杭州华为数字技术有限公司 | The treating method and apparatus of data flow |
| CN109819282A (en) * | 2017-11-22 | 2019-05-28 | 腾讯科技（深圳）有限公司 | A kind of video user classification recognition methods, device and medium |
-
2019
- 2019-06-12 CN CN201910507060.5A patent/CN110189151A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1236173A2 (en) * | 1999-10-27 | 2002-09-04 | Biowulf Technologies, LLC | Methods and devices for identifying patterns in biological systems |
| CN104283737B (en) * | 2014-09-30 | 2018-01-12 | 杭州华为数字技术有限公司 | The treating method and apparatus of data flow |
| US20160253750A1 (en) * | 2014-12-22 | 2016-09-01 | Palantir Technologies Inc. | Systems and user interfaces for dynamic and interactive investigation of bad actor behavior based on automatic clustering of related data in various data structures |
| CN106886518A (en) * | 2015-12-15 | 2017-06-23 | 国家计算机网络与信息安全管理中心 | A kind of method of microblog account classification |
| CN109819282A (en) * | 2017-11-22 | 2019-05-28 | 腾讯科技（深圳）有限公司 | A kind of video user classification recognition methods, device and medium |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110569509A (en) * | 2019-09-12 | 2019-12-13 | 广州荔支网络技术有限公司 | risk group identification method and device |
| CN110674479A (en) * | 2019-09-29 | 2020-01-10 | 武汉极意网络科技有限公司 | Abnormal behavior data real-time processing method, device, equipment and storage medium |
| CN111831825A (en) * | 2020-07-23 | 2020-10-27 | 咪咕文化科技有限公司 | Account detection method, account detection device, network equipment and storage medium |
| CN111831825B (en) * | 2020-07-23 | 2024-03-15 | 咪咕文化科技有限公司 | Account detection method, device, network equipment and storage medium |
| CN113011886A (en) * | 2021-02-19 | 2021-06-22 | 腾讯科技（深圳）有限公司 | Method and device for determining account type and electronic equipment |
| CN113011886B (en) * | 2021-02-19 | 2023-07-14 | 腾讯科技（深圳）有限公司 | Method and device for determining account type and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110189151A (en) | A kind of account detection method and relevant device | |
| CN110225036B (en) | Account detection method, device, server and storage medium | |
| US9235846B2 (en) | Systems, methods, and apparatuses for populating a table having null values using a predictive query interface | |
| CN109906451A (en) | Use the similarity searching of polyphone | |
| CN107766568A (en) | Effective query processing is carried out using the histogram in columnar database | |
| CN101535944A (en) | Scalable user clustering based on set similarity | |
| Artico et al. | How rare are power-law networks really? | |
| JP2014106655A (en) | Setting automatic generation method of time sequence database, setting automatic generation system, and monitoring server | |
| CN110471821A (en) | Abnormal alteration detection method, server and computer readable storage medium | |
| CN104112084B (en) | Execution-based license discovery and optimization method and device | |
| WO2019061664A1 (en) | Electronic device, user's internet surfing data-based product recommendation method, and storage medium | |
| CN107480268A (en) | Data query method and device | |
| US20110184956A1 (en) | Accessing digitally published content using re-indexing of search results | |
| CN107481058A (en) | A kind of Products Show method and Products Show device | |
| CN111414410A (en) | Data processing method, device, equipment and storage medium | |
| CN109766370A (en) | Data processing method, data service system and equipment | |
| CN113935788B (en) | Model evaluation method, device, equipment and computer readable storage medium | |
| CN108429632B (en) | Service monitoring method and device | |
| CN114579398A (en) | Log storage method, device, equipment and storage medium | |
| CN110334055B (en) | Method for acquiring material calculation data | |
| CN112487307A (en) | Time-aware POI recommendation method | |
| US10558647B1 (en) | High performance data aggregations | |
| CN117522419B (en) | Resource allocation method applied to customer relationship management system | |
| CN115564332B (en) | Government risk analysis method and system based on big data | |
| Liu et al. | Distributed data aggregation at scale for large community of users |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190830 |