CN107872436A - A kind of account recognition methods, apparatus and system - Google Patents
A kind of account recognition methods, apparatus and system Download PDFInfo
- Publication number
- CN107872436A CN107872436A CN201610857050.0A CN201610857050A CN107872436A CN 107872436 A CN107872436 A CN 107872436A CN 201610857050 A CN201610857050 A CN 201610857050A CN 107872436 A CN107872436 A CN 107872436A
- Authority
- CN
- China
- Prior art keywords
- account
- user
- identified
- feature
- user 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.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/316—User authentication by observing the pattern of computer usage, e.g. typical user behaviour
Landscapes
- Engineering & Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Social Psychology (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Telephonic Communication Services (AREA)
Abstract
Disclosed herein is a kind of account recognition methods, apparatus and system, for identifying the relevance between account to be identified and user, wherein, multiple features can be extracted from user behavior information of the account to be identified in scheduled duration, the account recognition methods includes:According to the user behavior information to be judged of account to be identified, the feature to be judged of account to be identified is determined;By feature to be judged and the multiple feature respectively compared with, obtain multiple comparative results;According to the multiple comparative result, the relevance between the account to be identified and user is identified.
Description
Technical field
The present invention relates to network data processing technique, more particularly to a kind of account recognition methods, apparatus and system.
Background technology
With the fast development of Internet technology, the life of people increasingly be unable to do without internet.User can interconnect
Service of the net platform register account number to be provided using internet platform.At present, user is after internet platform register account number, registration
Correct ID card information (such as including information such as ID card No., name, identity card address, user pictures), that is, be considered as
The account real name.However, at present, the situation that account is resell or usurped is more serious.Especially, some people especially one
A little criminals can buy other people identity information, and it is illegal that the account that even real-name authentication has been crossed carries out network fraud etc.
Criminal offence.It is difficult to navigate to real person liable due to assuming another's name to commit a crime, after crime, causes the network crime very rampant.In addition,
After account is usurped by criminal, because internet platform can not monitor the stolen of account in time, it may cause to use
The property loss at family.
In the related art, the technology that different-place login checking be present is monitored to account, i.e., when user's login account
Address when changing, prompt to carry out the authentication of account user.
To sum up, although the technology that correlation technique is verified with different-place login, it is the absence of continuing, accurately identifies account
The method of relevance between (i.e. account) and account user.
The content of the invention
It is the general introduction of the theme to being described in detail herein below.It is to limit the protection model of claim that this general introduction, which is not,
Enclose.
The embodiment of the present application provides a kind of account recognition methods, apparatus and system, can continue, accurately identify account with making
Relevance between user, to improve internet security.
The embodiment of the present application provides a kind of account recognition methods, for identifying associating between account to be identified and user
Property, wherein, multiple features, the account identification can be extracted from user behavior information of the account to be identified in scheduled duration
Method includes:According to the user behavior information to be judged of account to be identified, the feature to be judged of account to be identified is determined;Will
The feature to be judged of account to be identified and the multiple feature respectively compared with, obtain multiple comparative results;According to multiple
Comparative result, identify the relevance between account to be identified and user.
Wherein, according to multiple comparative results, the relevance between account to be identified and user is identified, including:
By being calculated using user's identification model multiple comparative results, come identify account to be identified and user it
Between relevance;Or
By judging whether multiple comparative results meet predetermined condition, to identify the pass between account to be identified and user
Connection property.
Wherein, the account recognition methods also includes:One of in the following manner from account to be identified in scheduled duration
User behavior information in extract multiple features:
Multiple users are extracted from user behavior information of the account to be identified in scheduled duration and are accustomed to feature;
Customer relationship feature and at least one is extracted from user behavior information of the account to be identified in scheduled duration
User is accustomed to feature.
Wherein, user behavior information includes multiple user behavior features, and each user behavior feature is corresponding with one or more
Individual characteristic value;
User is extracted from user behavior information of the account to be identified in scheduled duration in the following manner and is accustomed to feature:
In user behavior information out of scheduled duration, one or more characteristic values corresponding to user behavior feature are determined;
For each user behavior feature, determine to meet the first preparatory condition in one or more of characteristic values respectively
Characteristic value, and the characteristic value of the first preparatory condition of the satisfaction is defined as user and is accustomed to feature;
Wherein, first preparatory condition includes at least one of:Access times are most;It is most long using duration;Use
Number and the weighted sum maximum using duration.
Wherein, customer relationship feature, bag are extracted in the user behavior information from account to be identified in scheduled duration
Include:
According to the user behavior information in scheduled duration, the first relation score value and the second relation score value are determined respectively;Wherein,
The first relation score value includes the relation score value between the account to be identified and each first linked character, and described first closes
Connection is characterized in the linked character being associated with the account to be identified;The second relation score value include relation account with it is associated
The first linked character between relation score value, the relation account refers in addition to the account to be identified, associated with any first
The related account of feature;
According to the first relation score value and the second relation score value, determine the account to be identified and each relation account it
Between the 3rd relation score value;
According to the 3rd relation score value, the customer relationship feature of the account to be identified is determined;
Wherein, the linked character refers to build the characteristic value of the user behavior feature of relation between different accounts.
Wherein, the customer relationship feature includes:Meet the relation account of the second preparatory condition, wherein, described second is pre-
If condition includes:The 3rd relation score value between relation account and the account to be identified is more than or equal to first threshold, or,
3rd relation score value belongs to top n in sequence from high to low, and N is positive integer.
Wherein, the feature to be judged by the account to be identified and the multiple feature respectively compared with, obtain
To multiple comparative results, including:
Include user behavior feature to be judged in the feature to be judged, and the multiple feature include with it is described
When user corresponding to the user behavior feature judged is accustomed to feature, the characteristic value of user behavior feature to be judged described in comparison
It is whether consistent with user custom feature, obtain comparative result;
Include customer relationship feature in the multiple feature, and the feature to be judged includes user to be judged and closed
When being feature, between the customer relationship feature that customer relationship feature to be judged and the multiple feature described in calculating include
Jie Kade distances, the Jie Kade distances and Second Threshold, obtain comparative result.
Wherein, it is described by being calculated using user's identification model the multiple comparative result, to identify described treat
Before identifying the relevance between account and user, the account recognition methods also includes:Obtained by following steps described
User's identification model:
Sample data is chosen, the sample data includes the data for the account that known user changes;Using institute
State sample data to be trained machine learning algorithm model, obtain the user's identification model.
Wherein, it is described that the multiple comparative result is calculated using user's identification model, it is described to be identified to identify
Relevance between account and user, including:
The multiple comparative result is calculated using user's identification model, obtains the user of the account to be identified
The probable value not changed;When the probable value is more than three threshold values, identify that the user of the account to be identified does not change
Become;When the probable value is less than or equal to three threshold values, identify that the user of the account to be identified changes.
Wherein, multiple features are extracted in the user behavior information from account to be identified in scheduled duration, including:
From the registration moment of the account to be identified, at each acquisition of information moment, user's row out of scheduled duration
To extract multiple features in information;
Wherein, the scheduled duration referred between the registration moment of the account to be identified and newest acquisition of information moment
Duration, or, the scheduled duration refers to the interval duration between the adjacent acquisition of information moment.
The embodiment of the present application also provides a kind of account identification device, for identifying the pass between account to be identified and user
Connection property, the account identification device includes:
First acquisition module, for extracting multiple spies from user behavior information of the account to be identified in scheduled duration
Sign;
Second acquisition module, for the user behavior information to be judged according to the account to be identified, it is determined that described treat
Identify the feature to be judged of account;
Comparison module, for the feature to be judged of the account to be identified to be compared respectively with the multiple feature
Compared with obtaining multiple comparative results;
Identification module, for according to the multiple comparative result, identifying the pass between the account to be identified and user
Connection property.
Wherein, the identification module, for described to be identified according to the multiple comparative result, identification in the following manner
Relevance between account and user:
By being calculated using user's identification model the multiple comparative result, come identify the account to be identified with
Relevance between user;Or
By judging whether the multiple comparative result meets predetermined condition, to identify the account to be identified and user
Between relevance.
Wherein, first acquisition module, for one of in the following manner from account to be identified in scheduled duration
Multiple features are extracted in user behavior information:
Multiple users are extracted from user behavior information of the account to be identified in scheduled duration and are accustomed to feature;
Customer relationship feature and at least one is extracted from user behavior information of the account to be identified in scheduled duration
User is accustomed to feature.
Wherein, the user behavior information includes multiple user behavior features, and each user behavior feature is corresponding with one
Or multiple characteristic values;
First acquisition module, for the user behavior letter from account to be identified in scheduled duration in the following manner
User is extracted in breath and is accustomed to feature:
In user behavior information out of scheduled duration, one or more characteristic values corresponding to user behavior feature are determined;
For each user behavior feature, determine to meet the first preparatory condition in one or more of characteristic values respectively
Characteristic value, and the characteristic value of the first preparatory condition of the satisfaction is defined as user and is accustomed to feature;
Wherein, first preparatory condition includes at least one of:Access times are most;It is most long using duration;Use
Number and the weighted sum maximum using duration.
Wherein, first acquisition module, for the user from account to be identified in scheduled duration in the following manner
Customer relationship feature is extracted in behavioural information:
According to the user behavior information in scheduled duration, the first relation score value and the second relation score value are determined respectively;Wherein,
The first relation score value includes the relation score value between the account to be identified and each first linked character, and described first closes
Connection is characterized in the linked character being associated with the account to be identified;The second relation score value include relation account with it is associated
The first linked character between relation score value, the relation account refers in addition to the account to be identified, associated with any first
The related account of feature;
According to the first relation score value and the second relation score value, determine the account to be identified and each relation account it
Between the 3rd relation score value;
According to the 3rd relation score value, the customer relationship feature of the account to be identified is determined;
Wherein, the linked character refers to build the characteristic value of the user behavior feature of relation between different accounts.
Wherein, the customer relationship feature includes:Meet the relation account of the second preparatory condition, second preparatory condition
Including:The 3rd relation score value between relation account and the account to be identified is more than or equal to first threshold, or, the 3rd closes
It is that score value belongs to top n in sequence from high to low, N is positive integer.
Wherein, the comparison module, in the following manner entering feature to be judged and the multiple feature respectively
Row compares, and obtains multiple comparative results:
Include user behavior feature to be judged in the feature to be judged, and the multiple feature include with it is described
When user corresponding to the user behavior feature judged is accustomed to feature, the characteristic value of user behavior feature to be judged described in comparison
It is whether consistent with user custom feature, obtain comparative result;
Include customer relationship feature in the multiple feature, and the feature to be judged includes user to be judged and closed
When being feature, between the customer relationship feature that customer relationship feature to be judged and the multiple feature described in calculating include
Jie Kade distances, the Jie Kade distances and Second Threshold, obtain comparative result.
Wherein, the account identification device also includes:Model building module, for obtaining the user by following steps
Identification model:Sample data is chosen, the sample data includes the data for the account that known user changes;Using institute
State sample data to be trained machine learning algorithm model, obtain the user's identification model.
Wherein, the identification module, for using user's identification model to the multiple comparative result in the following manner
Calculated, to identify the relevance between the account to be identified and user:
The multiple comparative result is calculated using user's identification model, obtains the user of the account to be identified
The probable value not changed;When the probable value is more than three threshold values, identify that the user of the account to be identified does not change
Become;When the probable value is less than or equal to three threshold values, identify that the user of the account to be identified changes.
Wherein, first acquisition module, for the user from account to be identified in scheduled duration in the following manner
Multiple features are extracted in behavioural information:
From the registration moment of the account to be identified, at each acquisition of information moment, user's row out of scheduled duration
To extract multiple features in information;
Wherein, the scheduled duration referred between the registration moment of the account to be identified and newest acquisition of information moment
Duration, or, the scheduled duration refers to the interval duration between the adjacent acquisition of information moment.
The embodiment of the present application also provides a kind of account identifying system, for identifying the pass between account to be identified and user
Connection property, the account identifying system includes:First device and second device;
The first device, for extracting multiple spies from user behavior information of the account to be identified in scheduled duration
Sign, according to the user behavior information to be judged of the account to be identified, the feature to be judged of the account to be identified is determined,
By the feature to be judged and the multiple feature respectively compared with, obtain multiple comparative results, and by the multiple ratio
Relatively result is sent to the second device;
The second device, the multiple comparative result sent for receiving the first device, according to the multiple
Comparative result, identify the relevance between the account to be identified and user.
Wherein, the second device is used for described to be identified according to the multiple comparative result, identification in the following manner
Relevance between account and user:
By being calculated using user's identification model the multiple comparative result, come identify the account to be identified with
Relevance between user;Or
By judging whether the multiple comparative result meets predetermined condition, to identify the account to be identified and user
Between relevance.
Wherein, the first device is additionally operable to obtain the subscriber identification module in the following manner:Choose sample data,
Wherein, the sample data includes the data for the account that known user changes;Using the sample data to machine
Learning algorithm model is trained, and obtains the user's identification model;
The user's identification model that the first device is additionally operable to obtain is sent to the second device.
The embodiment of the present application also provides a kind of data processing electronics for account identification, including:Memory and place
Manage device;The memory is used to store the program for being used for account identification, and the program for account identification is by the processing
When device reads execution, following operation is performed:According to the user behavior information to be judged of the account to be identified, it is determined that described treat
Identify the feature to be judged of account;By the feature to be judged of the account to be identified and multiple features respectively compared with,
Obtain multiple comparative results;According to the multiple comparative result, the relevance between the account to be identified and user is identified,
Wherein, the multiple feature is extracted from user behavior information of the account to be identified in scheduled duration.
In the embodiment of the present application, can be extracted from user behavior information of the account to be identified in scheduled duration multiple
Feature;According to the user behavior information to be judged of the account to be identified, the spy to be judged of the account to be identified is determined
Sign;By feature to be judged and the multiple feature respectively compared with, obtain multiple comparative results;According to the multiple comparison
As a result, the relevance between the account to be identified and user is identified.The embodiment of the present application is excavated dependent on big data, depth
Multiple characteristic dimensions of behavior on user's line are dissected, realize the relevance efficiently identified between account and account user, example
Whether the user such as account monitoring changes, to improve internet security.
After reading and understanding accompanying drawing and being described in detail, it can be appreciated that other aspects.
Brief description of the drawings
Fig. 1 is the flow chart for the account recognition methods that the embodiment of the present application one provides;
Fig. 2 is the exemplary relationship figure of account and linked character in the embodiment of the present application one;
Fig. 3 is the exemplary relationship figure between account in the embodiment of the present application one;
Fig. 4 is the schematic diagram for the account identification device that the embodiment of the present application three provides;
Fig. 5 is the schematic diagram for the account identifying system that the embodiment of the present application four provides.
Embodiment
The embodiment of the present application is described in detail below in conjunction with accompanying drawing, it will be appreciated that embodiments described below is only
For instruction and explanation of the application, it is not used to limit the application.
It should be noted that term " first " in the description and claims of this application and above-mentioned accompanying drawing, "
Two " etc. be for distinguishing similar object, without for describing specific order or precedence.
If it should be noted that not conflicting, the feature in the embodiment of the present application and embodiment can be combined with each other,
Within the protection domain of the application.In addition, though logical order is shown in flow charts, but in some cases, can
With with different from the shown or described step of order execution herein.
Term defines:
User behavior:Refer to the operation behavior that the user of account (i.e. account) is carried out using the account, for example, account is noted
Volume, authentication, web page browsing, account modification, account log in.
User behavior information:Refer to information caused by the operation behavior that the user of account is carried out using the account, for example,
The registration behavioural information (hereinafter referred to as account log-on message) of account, authentication behavioural information (hereinafter referred to as authentication letter
Breath), web page browsing behavioural information (hereinafter referred to as net page browse information), (the hereinafter referred to as account modification of account act of revision information
Information), account log in behavioural information (hereinafter referred to as account log-on message) etc..
User behavior feature:Refer to the data of reflection user behavior index of correlation, used for example, account carries out user behavior
Facility information, used phone number, used ship-to etc..
Characteristic value corresponding to user behavior feature:Refer to the specific data of user behavior feature, such as when user behavior feature
For during phone number, corresponding characteristic value can be phone number 1, phone number 2 etc., i.e., specific mobile phone used in user
Number.
User is accustomed to feature:Refer to the feature for meeting corresponding conditionses in one or more characteristic values corresponding to user behavior feature
Value;For example, when the ship-to that user uses includes ship-to A1, A2 and A3, ship-to A1 access times are most
More, then user, which is accustomed to feature, to be:Conventional ship-to is ship-to A1.
Linked character:Refer to build the characteristic value of the user behavior feature of relation between different accounts;For example, account A1
Respective behavior was carried out using phone number 1, account A2 also carried out respective behavior using phone number 1, then account A1 and account
Linked character between number A2 can be phone number 1.
Embodiment one
Fig. 1 is the flow chart for the account recognition methods that the embodiment of the present application one provides.The present embodiment, which provides a kind of account, to be known
Other method, for identifying the relevance between account and account user to be identified, such as identify the user of account to be identified
Whether (i.e. user) changes.The account recognition methods that the present embodiment provides can apply to service end computing device (for example, service
Device) or the virtual machine that is run on service end computing device.Illustrated below exemplified by applied to service end computing device.
In the present embodiment, multiple spies can be extracted from user behavior information of the account to be identified in scheduled duration
Sign.The multiple feature can include multiple users and be accustomed to feature, or, including customer relationship feature and at least one user
It is accustomed to feature.The multiple feature is used as the reference feature of follow-up account identification process.In addition, account is being carried out every time
During number identification, multiple features can be extracted from user behavior information of the account to be identified in scheduled duration again, or,
Can extracted obtained multiple features before use.The application is not limited this.Existed below with elder generation from account to be identified
Multiple features are extracted in user behavior information in scheduled duration, the process for recycling the multiple feature progress account identification is
Example illustrates.
As shown in figure 1, the account recognition methods that the present embodiment provides, may comprise steps of:
Step 101:Multiple features are extracted from user behavior information of the account to be identified in scheduled duration.
In the present embodiment, as user in internet platform (for example, the electric business platform such as Jingdone district store and QQ etc. are social flat
Platform) after register account number, the background server of internet platform can record the operation information that user is carried out by the account and (that is, use
Family behavioural information), and all user behavior information of each account are saved in database sequentially in time, such as with day
The mode of will is preserved.
In the present embodiment, step 101 can include:From the registration moment of account to be identified, in each acquisition of information
At the moment, multiple features are extracted in the user behavior information out of scheduled duration.Wherein, the scheduled duration can refer to account to be identified
Number registration the moment and the newest acquisition of information moment between duration.In the present embodiment, described information obtains the moment can be with
Periodically set, i.e., from the registration moment of account to be identified, user behavior information that can periodically out of scheduled duration
The middle multiple features of extraction.
Alternatively, can be in daily fixed time (i.e. foregoing acquisition of information from the registration moment of account to be identified
Moment is set for fixed setting or periodically), the user behavior out of duration between registration moment and newest fixed time
Multiple features are extracted in information.Or the described information acquisition moment can determine according to instruction triggers, i.e., from account to be identified
From registering the moment, when receiving triggering command, from scheduled duration (registering the duration between moment and instruction triggers moment)
Multiple features are extracted in interior user behavior information.However, the application is not limited this.
In the present embodiment, no matter being triggered the acquisition of information moment using which kind of mode, multiple features are from account to be identified
The registration moment extract into the user behavior information occurred during the newest acquisition of information moment.However, in other realities
In existing mode, the scheduled duration can refer to the interval duration between the adjacent acquisition of information moment.That is, multiple features are from most
Extracted in user behavior information during new acquisition of information moment and previous acquisition of information moment.
For example, from the registration moment T0 of account to be identified, at the T1 moment, user behavior that can be from T0 to T1 in the moment
Multiple features are extracted in information, afterwards, at the T2 moment, multiple spies can be extracted in the user behavior information from T0 to T2 in the moment
Sign, afterwards, at the T3 moment, can extract multiple features in the user behavior information from T0 to T3 in the moment, wherein, T3-T2 etc.
It is equal to T1-T0 in T2-T1;Or from the registration moment T0 of account to be identified, can be from T0 to T1 in the moment at the T1 moment
User behavior information in extract multiple features, afterwards, can be in the user behavior information from T1 to T2 in the moment at the T2 moment
Multiple features are extracted, afterwards, at the T3 moment, multiple features can be extracted in the user behavior information from T2 to T3 in the moment.
In the present embodiment, user behavior information can include following one or more:Account log-on message, authentication
Information, account log-on message, network browsing information, Transaction Information, account modification information, chat message:
Wherein, account log-on message can include following one or more:Account hour of log-on, phone number, mailbox
Location, account title, account number cipher, carry out account registration used in the network information (for example, IP address, WIFI addresses), carry out
Geographical location information when account is registered is (for example, GPS (Global Positioning System, global positioning system) believes
Breath);
Authentication information can include following one or more:Authenticated time, ID card No., name, identity card
Location, carry out authentication used in facility information, carry out authentication used in the network information, carry out authentication when
Geographical location information, real people's biological information (for example, fingerprint, face information);
Account log-on message can include following one or more:Login time, account log in used in facility information,
Geographical location information, real people's biological information used in account login when the network information, progress account login;
Network browsing information can include following one or more:The web page address that browses, browsing time, browse duration,
The network address of collection, facility information used in network browsing is carried out, the network information used in network browsing is carried out, carries out network
Geographical location information when browsing;
Transaction Information can include following one or more:Purchase information, sell information, carry out used in network trading
Facility information, the network information and geographical location information;Wherein, buying information includes information of receiving, and information of receiving includes:Receive
Name, cell-phone number of receiving, ship-to;
Account modification information can include following one or more:Account modification time, account modification content (for example,
Password, account title, Security Question etc.), carry out account modification used in facility information, the network information and geographical position
Confidence ceases;
Chat message can include following one or more:Chatting object, chatting time, carry out used equipment of chatting
Information, the network information and geographical location information.
Above-mentioned user behavior information is only for example, and the application is not limited this.
In the present embodiment, step 101 can include one below:
Multiple users are extracted from user behavior information of the account to be identified in scheduled duration and are accustomed to feature;
Customer relationship feature and at least one is extracted from user behavior information of the account to be identified in scheduled duration
User is accustomed to feature.
In the present embodiment, the user behavior information includes multiple user behavior features, each user behavior feature pair
There should be one or more characteristic values;
Wherein it is possible to extract user from user behavior information of the account to be identified in scheduled duration in the following manner
It is accustomed to feature:
In user behavior information out of scheduled duration, one or more characteristic values corresponding to user behavior feature are determined;
For each user behavior feature, determine to meet the first preparatory condition in one or more of characteristic values respectively
Characteristic value, and the characteristic value of the first preparatory condition of the satisfaction is defined as user and is accustomed to feature;Wherein, according to user's row
Being characterized can be to should determine that a user is accustomed to feature;
Wherein, first preparatory condition includes at least one of:Access times are most;It is most long using duration;Use
Number and the weighted sum maximum using duration.
In the present embodiment, for each account, after passage time axle combs user behavior, with account use when
Between growth, increasing user behavior can be recorded, and therefore, can go out use according to these user behavior informations
It is accustomed to feature in family.
In this, it is assumed that the matrix that user is accustomed to feature is H, then H can be represented such as following formula:
H=<H1,H2,......,Hm>;
Above formula represents that user has m kinds to be accustomed to feature, such as equipment accustomed to using, the webpage often linked, the ground that often goes
Manage position, conventional ship-to, conventional phone number etc..Wherein, m is the integer more than or equal to 1.
A kind of assuming that custom feature H in above-mentioned matrix HmConventional ship-to is represented, due to receiving for user
Address might have several, in this, take the most ship-to of access times as conventional ship-to, its calculation formula
It is as follows:
Hm=MAX (x1,x2,......,xn);
N ship-to, x are indicated in above formulanRepresent the access times of n-th of ship-to, i.e. xn=fn.Wherein, n
For positive integer.
A kind of in addition, it is assumed that custom feature H in above-mentioned matrix HmRepresent the link of equipment, now equipment accustomed to using
Referential is had more using time length ratio access times, therefore, takes link to be set using the most long equipment of duration as accustomed to using
Standby, its calculation formula is as follows:
Hm=MAX (x1,x2,......,xn);
Represent to have used n equipment, x in above formulanRepresent the use duration of n-th of equipment, i.e. xn=tn.Wherein, n is just
Integer.
For example, what ship-to, user login account of the user behavior feature for example including user's use used sets
It is standby;The ship-to that user uses for example corresponds to following value (i.e. characteristic value):Ship-to A1, ship-to A2, place of acceptance
Location A3;The equipment that user's login account uses for example corresponds to following value (i.e. characteristic value):Equipment B1, equipment B2, equipment B3.In
This, it is determined that user commonly use ship-to when, by the most receipts of access times in above three ship-to in scheduled duration
Goods address is as conventional ship-to, such as ship-to A1;It is determined that during equipment accustomed to using, by scheduled duration, on
State and the most long equipment of duration is used in three equipment as equipment accustomed to using, such as equipment B1.
Similarly, the user such as conventional name of receiving, conventional phone number, which is accustomed to feature, to use and conventional receipts
Goods address identical determination mode is determined.The users such as the webpage that often links, the geographical position often gone are accustomed to feature can be with
It is determined using with equipment identical determination mode accustomed to using.
However, the application is not limited this.In other embodiments, user behavior feature pair can be considered
The access times for the one or more characteristic values answered and using duration come determine user be accustomed to feature.For example, it is assumed that above-mentioned
A kind of custom feature H in matrix HmCalculation formula is as follows:
Hm=MAX (x1,x2,......,xn);
Represent that a user behavior feature is corresponding with n characteristic value, x in above formulanRepresent the use duration of n-th of characteristic value
With the weighted sum of access times, i.e. xn=a × fn+b×tn, fnRepresent access times, tnExpression uses duration;A, b is weight, can
To pre-set as needed;N is positive integer.
In some embodiments, the attribute of feature can be accustomed to according to user, select above-mentioned one or more modes
Determine that user is accustomed to feature.However, the application is not limited this.
In the present embodiment, customer relationship is extracted in the user behavior information from account to be identified in scheduled duration
Feature, including:
According to the user behavior information in scheduled duration, the first relation score value and the second relation score value are determined respectively;Wherein,
The first relation score value includes the relation score value between the account to be identified and each first linked character, and described first closes
Connection is characterized in the linked character being associated with the account to be identified;The second relation score value include relation account with it is associated
The first linked character between relation score value, the relation account refers in addition to the account to be identified, associated with any first
The related account of feature;
According to the first relation score value and the second relation score value, determine the account to be identified and each relation account it
Between the 3rd relation score value;
According to the 3rd relation score value, the customer relationship feature of the account to be identified is determined;
Wherein, the linked character refers to build the characteristic value of the user behavior feature of relation between different accounts.
In the present embodiment, the customer relationship feature can include:Meet the relation account of the second preparatory condition, its
In, second preparatory condition includes:The 3rd relation score value between relation account and the account to be identified is more than or equal to
First threshold, or, the 3rd relation score value belongs to top n in sequence from high to low, and N is positive integer.Wherein, described first
Threshold value and N values can be configured according to actual conditions, and the application is not limited this.
In the present embodiment, when several accounts all it is logged in same equipment, had user behavior when, then these
Account has been considered as certain relation;Similarly, some accounts have identical WIFI addresses, ship-to etc. to be regarded as
Relation.The relation that be can be seen that by the above-mentioned several relations enumerated between account is linked at by various linked characters
Together, i.e., the relation between account is built by linked character., can be as the user of linked character in the present embodiment
Behavioural characteristic can include following one or more:WIFI addresses, ship-to, email address, phone number.
It is determined that during relation between account, the relation between account and linked character is first determined.In the present embodiment, carve
During the relation score value of picture account and linked character, value weight is set according to respective rule or experience, i.e., account is in linked character
On operation can all have a value weight W.For example purchase operation of the account on a linked character is than addition shopping cart behaviour
The value weight of work will height.On be worth weight setting rule can be determined according to actual conditions, the application to this simultaneously
Do not limit.
In the present embodiment, the relation score value of account and linked character can be determined according in the following manner:
First and value of the value weight of operation behavior of the account on a linked character are calculated, calculates the linked character
Second He of the value weight of all operation behaviors in user behavior feature belonging to (i.e. the characteristic value of user behavior feature)
Value;
Calculate described first and value divided by described second with the result of value, and the result is multiplied by above-mentioned linked character institute
Corresponding to the user behavior feature of category valuable weight maximum;
The half power of above-mentioned result of product is taken, as the account and the relation score value of the linked character.
For example, as shown in Fig. 2 account A1 is on three cell-phone numbers (linked character M1, M2, M3) in scheduled duration
There is corresponding user behavior.Wherein, the value weight of buying behavior is W1, and the value weight of Modify password behavior is W2, is tied up
The value weight for determining mailbox behavior is W3.
The calculation formula of account A1 and linked character M1 (i.e. cell-phone number 1) relation score value is given below:
As can be seen from the above equation if user's operation behavior on some linked character is more, the value weight of behavior is got over
Important, then relation score value is just high.In above formula MAX be in order to reconcile this relation score value, due to latter one fractional factorial has can
Can be 1, extreme case is that the account only has a linked character and a behavior, but the R (A of this when1,M1) and differ
Surely it is maximum, also sees that this account does the value weight acted on this linked character.
It is determined that the relation score value (that is, the first relation score value and the second relation score value) of each account and linked character it
Afterwards, full dose account relating is got up by linked character.First illustrate some definition below:
M(Ai) represent account AiThe set of all associated linked characters, i.e.,
M(Ai)={ M1,M2,......,Mn};
U(Mk) represent linked character MkThe set of upper all accounts being associated, i.e.,
U(Mk)={ A1,A2,......,An};
Account AiAnd AjRelation score value (the 3rd i.e. foregoing relation score value) can be calculated according to following formula:
Wherein, the effect of log functions is to reconcile, and when in order to avoid denominator item number is a lot, the overall value of this fraction is anxious
Play is reduced.
Relation score value (the 3rd i.e. above-mentioned relation of each account and relation account can be calculated by formula above
Value).For an account to be identified, according to all 3rd relation score values being calculated between relation account, according to the 3rd
The descending order of relation score value takes relation list (that is, foregoing user pass of the top n relation account as account to be identified
It is feature).Wherein, N is positive integer, and N value is not limited this according to determination, the application is actually needed.
One example of brief description.As shown in figure 3, account A1 and A2 is associated by linked character M1 and M2, account
A1 also has relation with linked character M3, and account A3 and linked character M1 and M4 have relation.According to the relation score value between account
Calculation formula, the relation score value of account A1 and A2 herein can calculate according to following formula:
In the present embodiment, by the customer relationship feature of account, the user for being capable of auxiliary judgment account behind, if
Drastic change once occurs for customer relationship feature, then can illustrate that the user of this account is likely occurred transformation.
Step 102:According to the user behavior information to be judged of account to be identified, the to be judged of account to be identified is determined
Feature.
In the present embodiment, feature to be judged can include:Customer relationship feature to be judged and at least one use
Family behavioural characteristic;Or multiple user behavior features to be judged can be included.
In the present embodiment, the feature to be judged can determine according to following either type:
User behavior information to be judged is determined according to predetermined period;Now, for an account to be identified, in each week
At moment phase, according to apart from this moment in cycle recent user behavior information, obtain user behavior feature to be judged;According to
The registration moment of the account to be identified to the user behavior information between the current period moment, or, according to the current period moment
The last period duration in user behavior information, it is determined that customer relationship feature (for example, customer relationship list) to be judged;
User behavior information to be judged is determined according to instruction triggers;Now, for an account to be identified, touched in instruction
The moment is sent out, according to apart from this instruction triggers moment recent user behavior information, obtains user behavior feature to be judged;
According to the registration moment of the account to be identified to the user behavior information between the instruction triggers moment, or, according to the finger
The user behavior information in the last period duration of triggering moment is made, it is determined that customer relationship feature to be judged;
User behavior information to be judged is determined according to the newest user behavior of the account to be identified generation moment;Now, pin
To an account to be identified, according to the user behavior information of newest generation, user behavior feature to be judged is obtained;Treated according to this
Identify that the user behavior information between the moment occurs for registration moment to the newest user behavior of account, or, according to institute
The user behavior information in the last period duration at newest user behavior generation moment is stated, it is determined that customer relationship to be judged is special
Sign;
User behavior information to be judged is determined according to specific user's behavior of account to be identified generation moment;For example, root
Change the user behavior information at moment according to user's entry address of account to be identified, it is determined that user behavior to be judged is special
Sign;The user behavior information to be changed according to registration moment to the entry address of account to be identified between the moment, or,
The user behavior information to be changed according to the entry address in the last period duration at moment, it is determined that customer relationship to be judged
Feature.
However, the application is not limited this, in other embodiment, other conditions can also be arranged as required to
It is determined that feature to be judged.
Step 103:By the feature to be judged of account to be identified and the multiple feature respectively compared with, obtain multiple
Comparative result.
It is used for the multiple multiple features for being characterized as obtaining according to step 101 being compared in this step.In some realizations
In mode, when this account identifies and then carries out account identification, the multiple features that directly can be obtained using step 101,
It is not required to repeat step 101 and extracts multiple features again.
In the present embodiment, this step can include:
Include user behavior feature to be judged in the feature to be judged, and the multiple feature include with it is described
When user corresponding to the user behavior feature judged is accustomed to feature, the characteristic value of user behavior feature to be judged described in comparison
It is whether consistent with user custom feature, obtain comparative result;
Include customer relationship feature in the multiple feature, and the feature to be judged includes user to be judged and closed
When being feature, between the customer relationship feature that customer relationship feature to be judged and the multiple feature described in calculating include
Jie Kade distances, the Jie Kade distances and Second Threshold, obtain comparative result.
Wherein, include user corresponding with the user behavior feature to be judged in the multiple feature and be accustomed to feature
When, if the characteristic value of user behavior feature to be judged is consistent with user's custom feature, it is determined that the user is accustomed to feature pair
The comparative result answered is the first numerical value, if the characteristic value of user behavior feature to be judged and user's custom feature are inconsistent,
Determine that comparative result corresponding to user's custom feature is second value;
Include customer relationship feature in the multiple feature, and the feature to be judged includes user to be judged and closed
When being feature, the outstanding card between the customer relationship feature that customer relationship feature to be judged and the multiple feature include is calculated
Moral distance, if Jie Kade distances are less than Second Threshold, it is determined that comparative result corresponding to customer relationship feature is the 3rd number
Value, if Jie Kade distances are more than or equal to the Second Threshold, it is determined that comparative result corresponding to customer relationship feature is the
Four numerical value.
User that the multiple feature includes be accustomed to feature without it is corresponding wait the user behavior feature judged when, can
To determine that the user is accustomed to comparative result corresponding to feature as sky.
In the present embodiment, the first numerical value is 1, represents normal, second value 0, represents abnormal.Third value is 0, table
Show exception, the 4th numerical value is 1, represents normal.That is, the comparative result obtained is Boolean type.
For example, being accustomed to feature for user, newest user behavior feature is accustomed to feature progress with corresponding user
Match somebody with somebody, if newest user behavior feature and user are accustomed to characteristic matching, it is determined that the comparative result that the user is accustomed to feature is
1, otherwise, it is determined that it is 0 that the user, which is accustomed to feature comparative result,.For example, certain user is often done shopping in Hangzhou, it is current suddenly
There is buying behavior to occur in Beijing, then this does not just meet custom, therefore user's buying habit feature abnormalities, buying habit
Comparative result is 0.If user, which is accustomed to feature and newest user behavior feature, does not have corresponding relation, the user is set to be accustomed to special
Comparative result corresponding to sign is sky.For example, newest user behavior feature does not include ship-to information, then conventional receive is embodied
Comparative result corresponding to user's custom feature of address could be arranged to sky.
In practical application, it may be multiple to be accustomed to feature in the user that step 101 is extracted, and treated what this step obtained
The number of the user behavior feature of judgement is likely less than or is accustomed to equal to the user extracted in step 101 the number of feature.In
In the present embodiment, the feature to be judged that the number of comparative result obtains according to this step is (for example, according to newest user behavior
Information determines) number of the user behavior feature that includes and customer relationship feature determines that e.g., less than or equal to step 101 carries
The number for the feature got.
For example, being directed to customer relationship feature, calculate newest relation list (being designated as first set) and determined in step 101
Relation list (being designated as second set) between Jie Kade distance (Jaccard Distance).Wherein, by first set with
Account number in the common factor of second set is referred to as common factor number, by first set and second set and concentration account number
Referred to as union number, then Jie Kade distances are equal to common factor number divided by the value of union number.Ruo Jiekade distances are less than the second threshold
Value, then can be determined that the customer relationship feature abnormalities of the account to be identified, and corresponding comparative result can be 0;Ruo Jiekade away from
From more than or equal to the Second Threshold, then it can be determined that the customer relationship feature of the account to be identified is normal, corresponding relatively knot
Fruit can be 1.Wherein, Second Threshold can be configured according to actual conditions.
Obtain each user be accustomed to comparative result corresponding to comparative result corresponding to feature and customer relationship feature it
Afterwards, all comparative results calculated can be combined into feature vector, X.
Step 104:According to the multiple comparative result, the relevance between the account to be identified and user is identified.
In the present embodiment, by being calculated using user's identification model the multiple comparative result, to identify
State the relevance between account to be identified and user.
In the present embodiment, before step 104, methods described also includes:The user is obtained by following steps to know
Other model:
Sample data is chosen, the sample data includes the data for the account that known user changes;Using institute
State sample data to be trained machine learning algorithm model, obtain the user's identification model.
In the present embodiment, first pass through model training and obtain user's identification model, afterwards, be directed to the identification of account every time
During, the characteristic vector that directly can obtain step 103 is put into user's identification model and calculated, to identify account with making
Relevance between user.
In the present embodiment, case is falsely used using existing identity to precipitate the sample manually marked.Utilize these samples
Carry out model training.Afterwards with the user's identification model trained, full dose user's on-line prediction is carried out, it becomes possible to draw each
The authenticity score value of account identity.
In the present embodiment, machine learning algorithm can include following any:Logic returns to algorithm, random forest is calculated
Method, GBDT (Gradient Boosting Decision Tree, gradient lifting decision tree), decision Tree algorithms, SVM
(Support Vector Machine, SVMs) algorithm, neural network algorithm.
In the present embodiment, user's identification model is built by taking logistic regression algorithm as an example.By logistic regression algorithm come
The optimal weights of characteristic vector are solved, obtain the approximate optimal solution of non-convex optimization problem.
Logic-based regression algorithm structure user's identification model is described as follows:
First, sigmoid functions are provided, the function is used to model output result being smoothly mapped to two-value (0 and 1):
2nd, characteristic vector and weight vectors are given, in this, it is assumed that characteristic vector is tieed up for n, i.e. characteristic vector is X=(x1,
x2,......,xn);Due to carrying out supervised learning using artificial mark sample, can manually be provided after sample is audited whether
For the mark of real people (whether the user of account changes), it is assumed that annotation results are y after manual examination and verification sample
(answer), if dealing (i.e. the user of account changes) occurs for account, result 0, do not occur to buy and sell (i.e. account
User does not change) then result be 1, i.e. y ∈ (0,1);Therefore, it is as follows to give feature weight vector:
ΘTX=θ0+θ1x1+θ2x2+......+θnxn;
3rd, structure forecast function, above-mentioned feature weight vector are characterized the result after vector weighting, features described above power
Weight vector be put into be calculated in sigmoid functions model prediction result it is as follows:
The result h of above formula outputθ(X) expression value is 1 probability, i.e. the probability that user's identity of account does not morph
Value;
4th, loss function is constructed;Assuming that one shares m sample, then the loss function of logistic regression is as follows:
In this, target is to find out one group of optimal Θ, allows the value of loss function to become minimum;It is excellent with gradient descent method
Change the loss function that Θ draws minimum, it is as follows to directly give renewal function:
Wherein, a represents model training speed, i.e. learning rate in renewal function, and this value is smaller, more approaches optimal solution, but
It is that model training speed can be very slow, vice versa, and the value can determine as the case may be, and the value is, for example, in the present embodiment
0.01;Represent j-th of feature in i-th of sample.Θ is exactly the important achievement that model training is drawn.
, subsequently can be directly using the result Θ of model training to step after the user's identification model after being trained
Prediction result h is calculated after 103 obtained characteristic vector weightingsθ(X) prediction result, obtained is the user of account to be identified
The probable value not changed.
In the present embodiment, step 104 can include:
The multiple comparative result is calculated using user's identification model, obtains the user of the account to be identified
The probable value not changed;When the probable value is more than three threshold values, identify that the user of the account to be identified does not change
Become;When the probable value is less than or equal to three threshold values, identify that the user of the account to be identified changes.Wherein,
3rd threshold value can be configured according to actual conditions.
In some implementations, the comparison procedure of probable value and the 3rd threshold value can be also placed in user's identification model
In, i.e. after multiple comparative results are put into user's identification model, whether the user that can directly obtain account to be identified sends out
The raw result changed.
In the present embodiment, when the user for identifying account changes, user can be prompted to carry out real people's certification,
Such as, it is desirable to real people's biological information such as the fingerprint of certification user, face.Real people's certification is carried out by afterwards, just allowing in user
User is continuing with the account and operated.
In the present embodiment, the big data based on internet platform, take out user and be accustomed to feature, extraction customer relationship is special
Sign, when being subsequently monitored to the user behavior of account, feature is accustomed to based on user or user is accustomed to feature and user is closed
It is feature, by the calculating of user's identification model, the relevance carried out between account and user identifies.
Illustrate following scene:
Scene one
User in internet platform register account number and after carrying out real-name authentication, not using the account carry out web page browsing or
Shopping operation, just resells the account to other people, other people use the account.For such case, in user's registration account
And after carrying out real-name authentication, internet platform according to the information of the user's registration account and the information of real-name authentication, such as
The custom feature such as the geographical position of can determine conventional phone number, often going, conventional name and customer relationship feature (example
Such as, customer relationship list);Afterwards, internet platform monitors the service condition of the account in real time, is detecting the use of the account
During behavior (for example, other people carry out web page browsing using the account), it may be determined that the newest user behavior information of the account
(such as including currently used facility information, current geographical location information, and active user's relation list etc.).By user
The information recorded after register account number obtains multiple comparative results, to multiple compared with current newest user behavior information
Comparative result is calculated using user's identification model, judges whether the user of the account changes.Now, actually due to making
User changes, and therefore, user equipment information, geographical location information, customer relationship list etc. may change, according to
Above-mentioned change can identify that user changes.Internet platform prompts account after the user of identification account changes
User carries out In vivo detection, the authentication mode such as recognition of face, using ensure account user as authentication I.
Scene two
User logs in the account after internet platform register account number, in a place and carries out web page browsing, at another
Having logged in the account carries out shopping operation again in place.Internet platform can detect and use when user uses the account every time
Whether person changes.When user another place log in the account carry out shopping operation when, although geographical location information there occurs
Change, still, in the case where the customer relationship feature of the account and other custom features do not change, internet platform can
To identify that the user of the account does not change, so as to which user will not be prompted to carry out authentication, compared in correlation technique, mutually
Networked platforms find the mode for suggesting that user is verified that changed using equipment or login place of account, this implementation
The scheme that example provides can lift the usage experience of user.
In summary, in the present embodiment, by the action trail of electric quotient data take out user custom feature or
User is accustomed to feature and customer relationship feature, is accustomed to feature with reference to multiple users or combines user's custom feature and customer relationship
Feature, can continue, accurately identify relevance between account and user, for example, account whether occur identity dealing or whether
It is stolen.Once the user for monitoring account changes, then require that user carries out the body such as In vivo detection or recognition of face
Part verification mode, using ensure account user as authentication I, so as to improve internet security.
Embodiment two
The present embodiment provides a kind of account recognition methods, for identifying the relevance between account to be identified and user,
Such as identify whether the user (i.e. user) of account to be identified changes.
The account recognition methods and the difference of embodiment one that the present embodiment provides are:In the present embodiment, judgement is passed through
Whether the multiple comparative result meets predetermined condition, to identify the relevance between the account to be identified and user.
In the present embodiment, by judging whether the multiple comparative result meets predetermined condition, described wait to know to identify
Other relevance between account and user can include:
Product of each comparative result with corresponding weight is calculated respectively, calculate all products and value;
When described and value is more than four threshold values, identify that the user of the account to be identified does not change;
When described and value is less than or equal to four threshold values, identify that the user of the account to be identified changes.
In the present embodiment, each feature pair for being extracted from user behavior information of the account to be identified in scheduled duration
There should be a weight, the weight can be calculated by machine learning algorithm.
Wherein, machine learning algorithm can include following any:Logic returns to algorithm, random forests algorithm, GBDT
(Gradient Boosting Decision Tree, gradient lifting decision tree), decision Tree algorithms, SVM (Support
Vector Machine, SVMs) algorithm, neural network algorithm.For example, the mistake of weight is calculated using logistic regression algorithm
Journey is referred to embodiment one, wherein, Θ is obtained optimal weights vector.
In the present embodiment, obtaining, comparative result and customer relationship feature corresponding to each user's custom feature are corresponding
Comparative result after, all comparative results calculated can be combined into feature vector, X, according to feature vector, X and optimal
Weight vectors Θ, calculate feature weight vector:
ΘTX=θ0+θ1x1+θ2x2+......+θnxn;
Relevance between account to be identified and user is identified according to the result of feature weight vector.For example,
When feature weight vector is more than four threshold values, identify that the user of the account to be identified does not change;Weighed in feature
When weight vector is less than or equal to four threshold values, identify that the user of the account to be identified changes.Wherein, the 4th threshold value can
To be set according to actual conditions, the application is not limited this.
However, the present embodiment is not limited to this.In other embodiment, it can also meet in multiple comparative results
When the number of predetermined value is more than or equal to five threshold values, identifies that the user of account to be identified does not change, comparing
When meeting that the number of predetermined value is less than five threshold values in as a result, identify that the user of account to be identified changes.It is real herein
Apply in mode, based on the description of compared result in embodiment one, when comparative result is Boolean type, predetermined value can be
1 (representing normal).The number of normal characteristics is more than or equal to the 5th threshold value in feature i.e. to be judged, then identifies account to be identified
User do not change.Wherein, the 5th threshold value can be set according to actual conditions.
In addition, other explanations on the present embodiment are referred to described in embodiment one, therefore repeated no more in this.
Embodiment three
As shown in figure 4, the present embodiment provides a kind of account identification device, for identifying between account to be identified and user
Relevance, the account identification device includes:
First acquisition module, for extracting multiple spies from user behavior information of the account to be identified in scheduled duration
Sign;
Second acquisition module, for the user behavior information to be judged according to the account to be identified, it is determined that described treat
Identify the feature to be judged of account;
Comparison module, for by the feature to be judged of account to be identified and the multiple feature respectively compared with, obtain
To multiple comparative results;
Identification module, for according to the multiple comparative result, identifying the pass between the account to be identified and user
Connection property.
In the present embodiment, first acquisition module, it can made a reservation for by one below mode from account to be identified
Multiple features are extracted in user behavior information in duration:
Multiple users are extracted from user behavior information of the account to be identified in scheduled duration and are accustomed to feature;
Customer relationship feature and at least one is extracted from user behavior information of the account to be identified in scheduled duration
User is accustomed to feature.
Wherein, the user behavior information includes multiple user behavior features, and each user behavior feature is corresponding with one
Or multiple characteristic values;
First acquisition module can in the following manner from account to be identified in scheduled duration user behavior letter
User is extracted in breath and is accustomed to feature:
In user behavior information out of scheduled duration, one or more characteristic values corresponding to user behavior feature are determined;
For each user behavior feature, determine to meet the first preparatory condition in one or more of characteristic values respectively
Characteristic value, and the characteristic value of the first preparatory condition of the satisfaction is defined as user and is accustomed to feature;
Wherein, first preparatory condition includes at least one of:Access times are most;It is most long using duration;Use
Number and the weighted sum maximum using duration.
In the present embodiment, first acquisition module can be in the following manner from account to be identified in scheduled duration
Customer relationship feature is extracted in interior user behavior information:
According to the user behavior information in scheduled duration, the first relation score value and the second relation score value are determined respectively;Wherein,
The first relation score value includes the relation score value between the account to be identified and each first linked character, and described first closes
Connection is characterized in the linked character being associated with the account to be identified;The second relation score value include relation account with it is associated
The first linked character between relation score value, the relation account refers in addition to the account to be identified, associated with any first
The related account of feature;
According to the first relation score value and the second relation score value, determine the account to be identified and each relation account it
Between the 3rd relation score value;
According to the 3rd relation score value, the customer relationship feature of the account to be identified is determined;
Wherein, the linked character refers to build the characteristic value of the user behavior feature of relation between different accounts.
In the present embodiment, the customer relationship feature includes:Meet the relation account of the second preparatory condition, described second
Preparatory condition includes:The 3rd relation score value between relation account and the account to be identified is more than or equal to first threshold, or
Person, the 3rd relation score value belong to top n in sequence from high to low, and N is positive integer.
In the present embodiment, the comparison module, can be used in the following manner by feature to be judged with it is described more
Individual feature is compared respectively, obtains multiple comparative results:
Include user behavior feature to be judged in the feature to be judged, and the multiple feature include with it is described
When user corresponding to the user behavior feature judged is accustomed to feature, the characteristic value of user behavior feature to be judged described in comparison
It is whether consistent with user custom feature, obtain comparative result;
Include customer relationship feature in the multiple feature, and the feature to be judged includes user to be judged and closed
When being feature, between the customer relationship feature that customer relationship feature to be judged and the multiple feature described in calculating include
Jie Kade distances, the Jie Kade distances and Second Threshold, obtain comparative result.
In the present embodiment, the identification module, it can be used for according to the multiple comparative result being known in the following manner
Relevance between not described account to be identified and user:
By being calculated using user's identification model the multiple comparative result, come identify the account to be identified with
Relevance between user;Or
By judging whether the multiple comparative result meets predetermined condition, to identify the account to be identified and user
Between relevance.
In the present embodiment, the account identification device can also include:Model building module, for passing through following steps
Obtain the user's identification model:Sample data is chosen, the sample data includes the account that known user changes
Data;Machine learning algorithm model is trained using the sample data, obtains the user's identification model.
In the present embodiment, the identification module, it can be used in the following manner using the multiple comparative result
User's identification model is calculated, to identify the relevance between the account to be identified and user:
The multiple comparative result is calculated using user's identification model, obtains the user of the account to be identified
The probable value not changed;When the probable value is more than three threshold values, identify that the user of the account to be identified does not change
Become;When the probable value is less than or equal to three threshold values, identify that the user of the account to be identified changes.
In the present embodiment, first acquisition module, it can be used in the following manner from account to be identified predetermined
Multiple features are extracted in user behavior information in duration:
From the registration moment of the account to be identified, at each acquisition of information moment, user's row out of scheduled duration
To extract multiple features in information;
Wherein, the scheduled duration referred between the registration moment of the account to be identified and newest acquisition of information moment
Duration, or, the scheduled duration refers to the interval duration between the adjacent acquisition of information moment.
The handling process of the account identification device provided on the present embodiment is referred to the method described in embodiment one, therefore
Repeated no more in this.
Example IV
As shown in figure 5, the present embodiment provides a kind of account identifying system, for identifying between account to be identified and user
Relevance, the account identifying system includes:First device and second device;
The first device, for extracting multiple spies from user behavior information of the account to be identified in scheduled duration
Sign, according to the user behavior information to be judged of the account to be identified, the feature to be judged of the account to be identified is determined,
By feature to be judged and the multiple feature respectively compared with, obtain multiple comparative results, and relatively tie the multiple
Fruit is sent to the second device;
The second device, the multiple comparative result sent for receiving the first device, according to the multiple
Comparative result, identify the relevance between the account to be identified and user.
In the present embodiment, the second device is used to identify institute according to the multiple comparative result in the following manner
State the relevance between account to be identified and user:
By being calculated using user's identification model the multiple comparative result, come identify the account to be identified with
Relevance between user;Or
By judging whether the multiple comparative result meets predetermined condition, to identify the account to be identified and user
Between relevance.
In the present embodiment, the first device is additionally operable to obtain the subscriber identification module in the following manner:Choose
Sample data, the sample data include the data for the account that known user changes;Utilize the sample data pair
Machine learning algorithm model is trained, and obtains the user's identification model;
The user's identification model that the first device is additionally operable to obtain is sent to the second device.
In the present embodiment, first device can be service end computing device or be run on service end computing device
Virtual machine, second device can be client computing device or the service end computing device different from first device.
The present embodiment and the difference of embodiment one are:In the present embodiment, the determination process of comparative result and user
The executive agent of the training process of identification model is different from the executive agent of account identification process.
In the present embodiment, for example, user row of the service end computing device from account to be identified in scheduled duration
To extract multiple features in information, according to the user behavior information to be judged of account to be identified, treating for account to be identified is determined
The feature of judgement, by feature to be judged and the multiple feature respectively compared with, obtain multiple comparative results, and in real time will
The multiple comparative result is sent to the client computing device;Also, the service end computing device utilizes the sample chosen
Notebook data carries out model training, obtains user's identification model, and the user's identification model after training is sent into client and calculated
Equipment;Client computing device is calculated using user's identification model the multiple comparative results received, obtains described treat
The user of account is identified without the probable value changed, and according to the probable value, the user of the identification account to be identified
Whether change.Specific implementation on correlated process in the present embodiment is referred to described in embodiment one, therefore is repeated no more in this.
Embodiment five
The embodiment of the present invention provides a kind of data processing electronics for account identification, including:Memory and processing
Device;The memory is used to store the program for being used for account identification, and the program for account identification is by the processor
When reading execution, following operation is performed:According to the user behavior information to be judged of the account to be identified, it is determined that described wait to know
The feature to be judged of other account;By feature to be judged and multiple features respectively compared with, obtain multiple comparative results;Root
According to the multiple comparative result, the relevance between the account to be identified and user is identified, wherein, the multiple feature is
Extracted from user behavior information of the account to be identified in scheduled duration.
In addition, the concrete operations of the computing device are referred to described in embodiment one and embodiment two, therefore in this not
Repeat again.
In addition, the embodiment of the present invention also provides a kind of computer-readable recording medium, computer executable instructions are stored with,
Described account recognition methods is realized when the computer executable instructions are executed by processor.
One of ordinary skill in the art will appreciate that all or part of step in the above method can be instructed by program
Related hardware (such as processor) is completed, and described program can be stored in computer-readable recording medium, as read-only storage,
Disk or CD etc..Alternatively, all or part of step of above-described embodiment can also be come using one or more integrated circuits
Realize.Correspondingly, each module/unit in above-described embodiment can be realized in the form of hardware, such as pass through integrated circuit
To realize its corresponding function, can also be realized in the form of software function module, such as be stored in and deposited by computing device
Program/instruction in reservoir realizes its corresponding function.The application is not restricted to the knot of the hardware and software of any particular form
Close.
The advantages of general principle and principal character and the application of the application has been shown and described above.The application is not by upper
State the limitation of embodiment, the principle for simply illustrating the application described in above-described embodiment and specification, do not depart from the application
On the premise of spirit and scope, the application also has various changes and modifications, and these changes and improvements both fall within claimed
In the range of the application.
Claims (15)
1. a kind of account recognition methods, for identifying the relevance between account to be identified and user, wherein, from account to be identified
Multiple features can be extracted in user behavior information number in scheduled duration, the account recognition methods includes:
According to the user behavior information to be judged of the account to be identified, the spy to be judged of the account to be identified is determined
Sign;
By the feature to be judged of the account to be identified and the multiple feature respectively compared with, obtain multiple relatively tying
Fruit;
According to the multiple comparative result, the relevance between the account to be identified and user is identified.
2. according to the method for claim 1, it is characterised in that it is described according to the multiple comparative result, treated described in identification
The relevance between account and user is identified, including:
By being calculated using user's identification model the multiple comparative result, to identify the account to be identified with using
Relevance between person;Or
By judging whether the multiple comparative result meets predetermined condition, to identify between the account to be identified and user
Relevance.
3. according to the method for claim 1, it is characterised in that methods described also includes:One of in the following manner from treating
Identify in user behavior information of the account in scheduled duration and extract multiple features:
Multiple users are extracted from user behavior information of the account to be identified in scheduled duration and are accustomed to feature;
Customer relationship feature and at least one user are extracted from user behavior information of the account to be identified in scheduled duration
It is accustomed to feature.
4. according to the method for claim 3, it is characterised in that it is special that the user behavior information includes multiple user behaviors
Sign, each user behavior feature are corresponding with one or more characteristic values;
User is extracted from user behavior information of the account to be identified in scheduled duration in the following manner and is accustomed to feature:
In user behavior information out of scheduled duration, one or more characteristic values corresponding to user behavior feature are determined;
For each user behavior feature, the feature for meeting the first preparatory condition in one or more of characteristic values is determined respectively
Value, and the characteristic value of the first preparatory condition of the satisfaction is defined as user and is accustomed to feature;
Wherein, first preparatory condition includes at least one of:Access times are most;It is most long using duration;Access times
It is maximum with the weighted sum using duration.
5. according to the method for claim 4, it is characterised in that user's row from account to be identified in scheduled duration
To extract customer relationship feature in information, including:
According to the user behavior information in scheduled duration, the first relation score value and the second relation score value are determined respectively;Wherein, it is described
First relation score value includes the relation score value between the account to be identified and each first linked character, and first association is special
Sign is the linked character being associated with the account to be identified;The second relation score value includes relation account and associated the
Relation score value between one linked character, the relation account refer in addition to the account to be identified, with any first linked character
Related account;
According to the first relation score value and the second relation score value, determine between the account to be identified and each relation account
3rd relation score value;
According to the 3rd relation score value, the customer relationship feature of the account to be identified is determined;
Wherein, the linked character refers to build the characteristic value of the user behavior feature of relation between different accounts.
6. according to the method for claim 5, it is characterised in that the customer relationship feature includes:Meet the second default bar
The relation account of part, wherein, second preparatory condition includes:The 3rd relation between relation account and the account to be identified
Score value is more than or equal to first threshold, or, the 3rd relation score value belongs to top n in sequence from high to low, and N is just whole
Number.
7. according to the method for claim 1, it is characterised in that the feature to be judged by the account to be identified with
The multiple feature is compared respectively, obtains multiple comparative results, including:
Include user behavior feature to be judged in the feature to be judged, and the multiple feature includes waiting to sentence with described
When user corresponding to disconnected user behavior feature is accustomed to feature, the characteristic value of user behavior feature to be judged and institute described in comparison
Whether consistent state user's custom feature, obtain comparative result;
Include customer relationship feature in the multiple feature, and the feature to be judged includes customer relationship spy to be judged
During sign, the outstanding card between the customer relationship feature that customer relationship feature to be judged and the multiple feature described in calculating include
Moral distance, the Jie Kade distances and Second Threshold, obtain comparative result.
8. according to the method for claim 2, it is characterised in that described by being known to the multiple comparative result using user
Other model is calculated, come before identifying the relevance between the account to be identified and user, methods described also includes:It is logical
Cross following steps and obtain the user's identification model:
Sample data is chosen, the sample data includes the data for the account that known user changes;
Machine learning algorithm model is trained using the sample data, obtains the user's identification model.
9. according to the method for claim 2, it is characterised in that described that user's identification mould is used to the multiple comparative result
Type is calculated, to identify the relevance between the account to be identified and user, including:
The multiple comparative result is calculated using user's identification model, the user for obtaining the account to be identified does not have
The probable value of change;
When the probable value is more than three threshold values, identify that the user of the account to be identified does not change;
When the probable value is less than or equal to three threshold values, identify that the user of the account to be identified changes.
10. according to the method for claim 3, it is characterised in that the user from account to be identified in scheduled duration
Multiple features are extracted in behavioural information, including:
From the registration moment of the account to be identified, at each acquisition of information moment, the user behavior letter out of scheduled duration
Multiple features are extracted in breath;
Wherein, the scheduled duration refer to the account to be identified registration the moment and the newest acquisition of information moment between when
It is long, or, the scheduled duration refers to the interval duration between the adjacent acquisition of information moment.
A kind of 11. account identification device, for identifying the relevance between account to be identified and user, the account identification dress
Put including:
First acquisition module, for extracting multiple features from user behavior information of the account to be identified in scheduled duration;
Second acquisition module, for the user behavior information to be judged according to the account to be identified, determine described to be identified
The feature to be judged of account;
Comparison module, for by the feature to be judged of the account to be identified and the multiple feature respectively compared with, obtain
To multiple comparative results;
Identification module, for according to the multiple comparative result, identifying the relevance between the account to be identified and user.
12. device according to claim 11, it is characterised in that the identification module, for basis in the following manner
The multiple comparative result, identify the relevance between the account to be identified and user:
By being calculated using user's identification model the multiple comparative result, to identify the account to be identified with using
Relevance between person;Or
By judging whether the multiple comparative result meets predetermined condition, to identify between the account to be identified and user
Relevance.
A kind of 13. account identifying system, for identifying the relevance between account to be identified and user, the account identification system
System includes:First device and second device;
The first device, for extracting multiple features, root from user behavior information of the account to be identified in scheduled duration
According to the user behavior information to be judged of the account to be identified, the feature to be judged of the account to be identified is determined, by institute
Compared with stating feature to be judged and the multiple feature respectively, multiple comparative results are obtained, and relatively tie the multiple
Fruit is sent to the second device;
The second device, the multiple comparative result sent for receiving the first device, according to the multiple comparison
As a result, the relevance between the account to be identified and user is identified.
14. system according to claim 13, it is characterised in that the second device is used in the following manner according to institute
Multiple comparative results are stated, identify the relevance between the account to be identified and user:
By being calculated using user's identification model the multiple comparative result, to identify the account to be identified with using
Relevance between person;Or
By judging whether the multiple comparative result meets predetermined condition, to identify between the account to be identified and user
Relevance.
15. system according to claim 14, it is characterised in that the first device is additionally operable to obtain in the following manner
The subscriber identification module:Sample data is chosen, wherein, the sample data includes the account that known user changes
Data;Machine learning algorithm model is trained using the sample data, obtains the user's identification model;
The user's identification model that the first device is additionally operable to obtain is sent to the second device.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610857050.0A CN107872436B (en) | 2016-09-27 | 2016-09-27 | Account identification method, device and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610857050.0A CN107872436B (en) | 2016-09-27 | 2016-09-27 | Account identification method, device and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107872436A true CN107872436A (en) | 2018-04-03 |
CN107872436B CN107872436B (en) | 2020-11-24 |
Family
ID=61761075
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610857050.0A Active CN107872436B (en) | 2016-09-27 | 2016-09-27 | Account identification method, device and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107872436B (en) |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108898418A (en) * | 2018-05-31 | 2018-11-27 | 康键信息技术(深圳)有限公司 | User account detection method, device, computer equipment and storage medium |
CN108961034A (en) * | 2018-05-25 | 2018-12-07 | 中国建设银行股份有限公司 | System and method, storage medium based on user behavior certification |
CN109190874A (en) * | 2018-07-12 | 2019-01-11 | 阿里巴巴集团控股有限公司 | The implementation method and device of multiple-limb operation flow |
CN109242470A (en) * | 2018-08-14 | 2019-01-18 | 阿里巴巴集团控股有限公司 | Personal identification method, device, equipment and computer readable storage medium |
CN110008980A (en) * | 2019-01-02 | 2019-07-12 | 阿里巴巴集团控股有限公司 | Identification model generation method, recognition methods, device, equipment and storage medium |
CN110046783A (en) * | 2018-12-13 | 2019-07-23 | 阿里巴巴集团控股有限公司 | Falsely use account recognition methods, device, electronic equipment and storage medium |
CN110399538A (en) * | 2019-07-26 | 2019-11-01 | 腾讯科技(深圳)有限公司 | A kind of recognition methods of user account and relevant device |
CN110532957A (en) * | 2019-08-30 | 2019-12-03 | 北京市商汤科技开发有限公司 | Face identification method and device, electronic equipment and storage medium |
CN110995655A (en) * | 2019-11-06 | 2020-04-10 | 国网浙江武义县供电有限公司 | Method and device for monitoring corresponding relation between personnel and equipment behaviors |
CN111611616A (en) * | 2020-05-10 | 2020-09-01 | 蔡萍萍 | Data encryption method based on E-commerce live broadcast platform and big data server |
TWI703465B (en) * | 2018-05-28 | 2020-09-01 | 香港商阿里巴巴集團服務有限公司 | Core body method and device |
CN111652626A (en) * | 2020-06-18 | 2020-09-11 | 支付宝(杭州)信息技术有限公司 | Method and device for realizing service |
CN111708995A (en) * | 2020-06-12 | 2020-09-25 | 中国建设银行股份有限公司 | Service processing method, device and equipment |
CN111784355A (en) * | 2020-07-17 | 2020-10-16 | 支付宝(杭州)信息技术有限公司 | Transaction security verification method and device based on edge calculation |
CN111784468A (en) * | 2020-07-01 | 2020-10-16 | 支付宝(杭州)信息技术有限公司 | Account association method and device and electronic equipment |
CN112115455A (en) * | 2020-09-28 | 2020-12-22 | 中国银行股份有限公司 | Setting method, device, server and medium for incidence relation of multiple user accounts |
CN112667982A (en) * | 2019-10-16 | 2021-04-16 | 吴昌宇 | Fingerprint and face identification monitoring system |
CN113011889A (en) * | 2021-03-10 | 2021-06-22 | 腾讯科技(深圳)有限公司 | Account abnormity identification method, system, device, equipment and medium |
CN114531252A (en) * | 2020-10-30 | 2022-05-24 | 中国电信股份有限公司 | Account log security audit method and device |
CN116451201A (en) * | 2023-03-14 | 2023-07-18 | 电子科技大学 | Mobile communication identity authentication method and system based on artificial intelligence |
WO2024038239A1 (en) * | 2022-08-19 | 2024-02-22 | Worldline | Method and device for behavioural biometric authentication |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110197275A1 (en) * | 2005-02-28 | 2011-08-11 | Mcafee, Inc. | Stopping and remediating outbound messaging abuse |
CN103544289A (en) * | 2013-10-28 | 2014-01-29 | 公安部第三研究所 | Feature extraction achieving method based on deploy and control data mining |
CN104426884A (en) * | 2013-09-03 | 2015-03-18 | 深圳市腾讯计算机系统有限公司 | Method for authenticating identity and device for authenticating identity |
CN104574192A (en) * | 2013-10-25 | 2015-04-29 | 华为技术有限公司 | Method and device for identifying same user from multiple social networks |
CN104660594A (en) * | 2015-02-09 | 2015-05-27 | 中国科学院信息工程研究所 | Method for identifying virtual malicious nodes and virtual malicious node network in social networks |
CN104811428A (en) * | 2014-01-28 | 2015-07-29 | 阿里巴巴集团控股有限公司 | Method, device and system for verifying client identity by social relation data |
CN104917643A (en) * | 2014-03-11 | 2015-09-16 | 腾讯科技(深圳)有限公司 | Abnormal account detection method and device |
US20160182545A1 (en) * | 2008-12-02 | 2016-06-23 | The Trustees Of Columbia University In The City Of New York | Methods, systems, and media for masquerade attack detection by monitoring computer user behavior |
-
2016
- 2016-09-27 CN CN201610857050.0A patent/CN107872436B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110197275A1 (en) * | 2005-02-28 | 2011-08-11 | Mcafee, Inc. | Stopping and remediating outbound messaging abuse |
US20160182545A1 (en) * | 2008-12-02 | 2016-06-23 | The Trustees Of Columbia University In The City Of New York | Methods, systems, and media for masquerade attack detection by monitoring computer user behavior |
CN104426884A (en) * | 2013-09-03 | 2015-03-18 | 深圳市腾讯计算机系统有限公司 | Method for authenticating identity and device for authenticating identity |
CN104574192A (en) * | 2013-10-25 | 2015-04-29 | 华为技术有限公司 | Method and device for identifying same user from multiple social networks |
CN103544289A (en) * | 2013-10-28 | 2014-01-29 | 公安部第三研究所 | Feature extraction achieving method based on deploy and control data mining |
CN104811428A (en) * | 2014-01-28 | 2015-07-29 | 阿里巴巴集团控股有限公司 | Method, device and system for verifying client identity by social relation data |
CN104917643A (en) * | 2014-03-11 | 2015-09-16 | 腾讯科技(深圳)有限公司 | Abnormal account detection method and device |
CN104660594A (en) * | 2015-02-09 | 2015-05-27 | 中国科学院信息工程研究所 | Method for identifying virtual malicious nodes and virtual malicious node network in social networks |
Non-Patent Citations (2)
Title |
---|
S. HE, H. WANG AND Z. H. JIANG: ""Identifying user behavior on Twitter based on multi-scale entropy"", 《PROCEEDINGS 2014 IEEE INTERNATIONAL CONFERENCE ON SECURITY, PATTERN ANALYSIS, AND CYBERNETICS (SPAC), WUHAN, 2014》 * |
樊茜,许洪波,梁英: ""社交网络账号的马甲关系辨识方法"", 《中文信息学报》 * |
Cited By (34)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108961034A (en) * | 2018-05-25 | 2018-12-07 | 中国建设银行股份有限公司 | System and method, storage medium based on user behavior certification |
TWI703465B (en) * | 2018-05-28 | 2020-09-01 | 香港商阿里巴巴集團服務有限公司 | Core body method and device |
US10938812B2 (en) | 2018-05-28 | 2021-03-02 | Advanced New Technologies Co., Ltd. | Identity verification method and apparatus |
US11153311B2 (en) | 2018-05-28 | 2021-10-19 | Advanced New Technologies Co., Ltd. | Identity verification method and apparatus |
CN108898418A (en) * | 2018-05-31 | 2018-11-27 | 康键信息技术(深圳)有限公司 | User account detection method, device, computer equipment and storage medium |
CN109190874A (en) * | 2018-07-12 | 2019-01-11 | 阿里巴巴集团控股有限公司 | The implementation method and device of multiple-limb operation flow |
CN109242470A (en) * | 2018-08-14 | 2019-01-18 | 阿里巴巴集团控股有限公司 | Personal identification method, device, equipment and computer readable storage medium |
CN110046783A (en) * | 2018-12-13 | 2019-07-23 | 阿里巴巴集团控股有限公司 | Falsely use account recognition methods, device, electronic equipment and storage medium |
CN110046783B (en) * | 2018-12-13 | 2023-04-28 | 创新先进技术有限公司 | Method and device for identifying fraudulent account, electronic equipment and storage medium |
CN110008980A (en) * | 2019-01-02 | 2019-07-12 | 阿里巴巴集团控股有限公司 | Identification model generation method, recognition methods, device, equipment and storage medium |
CN110399538A (en) * | 2019-07-26 | 2019-11-01 | 腾讯科技(深圳)有限公司 | A kind of recognition methods of user account and relevant device |
CN110399538B (en) * | 2019-07-26 | 2023-11-07 | 腾讯科技(深圳)有限公司 | User account identification method and related equipment |
CN110532957A (en) * | 2019-08-30 | 2019-12-03 | 北京市商汤科技开发有限公司 | Face identification method and device, electronic equipment and storage medium |
CN112667982A (en) * | 2019-10-16 | 2021-04-16 | 吴昌宇 | Fingerprint and face identification monitoring system |
CN110995655A (en) * | 2019-11-06 | 2020-04-10 | 国网浙江武义县供电有限公司 | Method and device for monitoring corresponding relation between personnel and equipment behaviors |
CN110995655B (en) * | 2019-11-06 | 2022-08-23 | 国网浙江武义县供电有限公司 | Method and device for monitoring corresponding relation between personnel and equipment behaviors |
CN111611616B (en) * | 2020-05-10 | 2021-04-13 | 湖北微果网络科技有限公司 | Data encryption method based on E-commerce live broadcast platform and big data server |
CN111611616A (en) * | 2020-05-10 | 2020-09-01 | 蔡萍萍 | Data encryption method based on E-commerce live broadcast platform and big data server |
CN111708995A (en) * | 2020-06-12 | 2020-09-25 | 中国建设银行股份有限公司 | Service processing method, device and equipment |
CN111652626A (en) * | 2020-06-18 | 2020-09-11 | 支付宝(杭州)信息技术有限公司 | Method and device for realizing service |
CN111652626B (en) * | 2020-06-18 | 2023-03-24 | 支付宝(杭州)信息技术有限公司 | Method and device for realizing service |
CN111784468B (en) * | 2020-07-01 | 2022-11-18 | 支付宝(杭州)信息技术有限公司 | Account association method and device and electronic equipment |
CN111784468A (en) * | 2020-07-01 | 2020-10-16 | 支付宝(杭州)信息技术有限公司 | Account association method and device and electronic equipment |
CN111784355B (en) * | 2020-07-17 | 2023-03-10 | 支付宝(杭州)信息技术有限公司 | Transaction security verification method and device based on edge calculation |
CN111784355A (en) * | 2020-07-17 | 2020-10-16 | 支付宝(杭州)信息技术有限公司 | Transaction security verification method and device based on edge calculation |
CN112115455B (en) * | 2020-09-28 | 2023-10-24 | 中国银行股份有限公司 | Method, device, server and medium for setting association relation of multiple user accounts |
CN112115455A (en) * | 2020-09-28 | 2020-12-22 | 中国银行股份有限公司 | Setting method, device, server and medium for incidence relation of multiple user accounts |
CN114531252A (en) * | 2020-10-30 | 2022-05-24 | 中国电信股份有限公司 | Account log security audit method and device |
CN114531252B (en) * | 2020-10-30 | 2024-05-10 | 中国电信股份有限公司 | Security audit method and security audit device for account log |
CN113011889A (en) * | 2021-03-10 | 2021-06-22 | 腾讯科技(深圳)有限公司 | Account abnormity identification method, system, device, equipment and medium |
CN113011889B (en) * | 2021-03-10 | 2023-09-15 | 腾讯科技(深圳)有限公司 | Account anomaly identification method, system, device, equipment and medium |
WO2024038239A1 (en) * | 2022-08-19 | 2024-02-22 | Worldline | Method and device for behavioural biometric authentication |
FR3138955A1 (en) * | 2022-08-19 | 2024-02-23 | Worldline | Behavioral biometric authentication method and device |
CN116451201A (en) * | 2023-03-14 | 2023-07-18 | 电子科技大学 | Mobile communication identity authentication method and system based on artificial intelligence |
Also Published As
Publication number | Publication date |
---|---|
CN107872436B (en) | 2020-11-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107872436A (en) | A kind of account recognition methods, apparatus and system | |
US20230013306A1 (en) | Sensitive Data Classification | |
CN106453061B (en) | A kind of method and system identifying network fraudulent act | |
US20140351109A1 (en) | Method and apparatus for automatically identifying a fraudulent order | |
CN104504335B (en) | Fishing APP detection methods and system based on page feature and URL features | |
CN104217160A (en) | Method and system for detecting Chinese phishing website | |
CN108229963A (en) | The Risk Identification Method and device of user's operation behavior | |
CN107437223A (en) | Credit information checking method, device and equipment | |
CN107153847A (en) | Predict method and computing device of the user with the presence or absence of malicious act | |
WO2016201938A1 (en) | Multi-stage phishing website detection method and system | |
CN113722611B (en) | Recommendation method, device and equipment for government affair service and computer readable storage medium | |
CN109214914A (en) | A kind of loan information checking method and device based on communication open platform | |
CN107134074A (en) | Storing collaborative share method, storing equipment and computer-readable recording medium | |
TWI752349B (en) | Risk identification method and device | |
CN107071193A (en) | The method and apparatus of interactive answering system accessing user | |
CN110020002A (en) | Querying method, device, equipment and the computer storage medium of event handling scheme | |
CN108073645A (en) | A kind of job-hunter of recruitment platform recommends page display method and device | |
CN107622197A (en) | Device identification method and device, weighing computation method and device for equipment identification | |
CN107809370B (en) | User recommendation method and device | |
CN111127185A (en) | Credit fraud identification model construction method and device | |
HaratiNik et al. | FUZZGY: A hybrid model for credit card fraud detection | |
CN106910135A (en) | User recommends method and device | |
CN115577172A (en) | Article recommendation method, device, equipment and medium | |
CN117114514A (en) | Talent information analysis management method, system and device based on big data | |
CN111582757A (en) | Fraud risk analysis method, device, equipment and computer-readable storage medium |
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 |