Specific embodiment
The application is described in further detail with reference to the accompanying drawing.
In a typical configuration of this application, terminal, the equipment of service network include one or more processors
(CPU), input/output interface, network interface and memory.
Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/or
The forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable medium
Example.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media, can be by any side
Method or technology realize that information stores.Information can be the device or other numbers of computer readable instructions, data structure, program
According to.The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory
(SRAM), dynamic random access memory (DRAM), other kinds of random access memory (RAM), read-only memory
(ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory techniques, CD-ROM (CD-
ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storages
Equipment or any other non-transmission medium, can be used for storage can be accessed by a computing device information.
The embodiment of the present application provides trade company's recognition methods, and this method can be based on the trading activity of user, determining and sample
This trade company may have the association trade company of like attribute, such as sample trade company is risk trade company, then is associated with trade company and is likely to
It is association trade company.On this basis, the association trade company and institute are further determined according to the incidence relation between trade company and user
The similarity between sample trade company is stated, and is denoised based on similarity, the lower association trade company of similarity is excluded, to be promoted
The accuracy of trade company's identification.
In actual scene, the identification equipment for executing this method can be user equipment, the network equipment or user equipment
Constituted equipment is integrated by network with the network equipment.Wherein, the user equipment include but is not limited to personal computer,
All kinds of terminal devices such as smart phone, tablet computer, the network equipment include but is not limited to that such as network host, single network take
Device, multiple network server collection or the set of computers based on cloud computing etc. of being engaged in are realized, can be used to implement when alarm clock is arranged
Part processing function.Here, cloud is made of a large amount of hosts or network server for being based on cloud computing (Cloud Computing),
Wherein, cloud computing is one kind of distributed computing, a virtual machine consisting of a loosely coupled set of computers.
Fig. 1 shows a kind of process flow of trade company's recognition methods provided by the embodiments of the present application, includes at least following place
Manage step:
Step S101 obtains the user that transaction occurred with sample trade company in first time window, as association user.
Wherein, the sample trade company is the sample according to selected by the demand of identification scene, these samples can have certain specific
Attribute, in order to identify the target trade company for obtaining and there is like attribute.For example, the attribute, which can be, is accused of gambling, then may be used
It, can be with if user needs to identify the trade company for being accused of fraudulent act using the gambling trade company that will have been acknowledged as sample trade company
Using the swindle trade company having been acknowledged as sample trade company.
In actual scene, the set of trade company can be denoted as merchant cluster, and the set of user can be denoted as user collection
Group, at this point, the trade company in the merchant cluster is sample trade company, determining association user can be put into user cluster.
The first time window refers to a preset period, and the length of the period can be according to the demand of actual scene
Setting, for example, it can be set to being 1 hour, 90 minutes, 2 hours etc..
If in the embodiment of the present application, the first time window that sets as the period of 20 points to 21 points of hour,
Merchant cluster includes 2 sample trade company m1 and m2, the two sample trade companies are gambling trade company.By transferring the two samples
The transaction data of payment platform where this trade company can count and all friendship occur between 20 points to 21 points with m1 or m2
Easy user, such as the result of association user obtained in embodiment is to be denoted as u1-10, i.e. user cluster comprising 10 users
In include tetra- elements of u1-10.
Step S102 determines that transaction count is more in the association user according to the transaction count of the association user
Central user.
In actual scene, for certain a kind of trade company with like attribute, generally all can some frequently trade
User group, such as trade company of gambling, the user of transaction may include some users to gamble once in a while, it also can be comprising frequent
The user to gamble.Wherein, the user to gamble once in a while participates in gambling at other gambling trade companies since the number of gambling is few
Probability is also relatively small, conversely, the user frequently to gamble is more likely to since the number for participating in gambling is more in different gamblings
Bo Shanghuchu participates in gambling.It therefore, is frequency by the more central user of transaction count under the scene of gambling trade company identification
A possibility that numerous user to gamble, the other gambling trade companies that can be found by these users, can be bigger, needs according to transaction time
Number determines central user in association user.
In actual scene, it can determine that the more center of transaction count is used in the association user in the following way
Family.For example, can be ranked up according to the transaction count of association user, then chosen from sequence according to the result of sequence it is N number of,
Centered on user.By taking the user in aforementioned user cluster as an example, the transaction count of each user can be as shown in the table:
u1 |
u2 |
u3 |
u4 |
u5 |
u6 |
u7 |
u8 |
u9 |
u10 |
2 |
3 |
21 |
105 |
99 |
6 |
1 |
5 |
63 |
1 |
It follows that according to the ranking results of transaction count from greatly to it is small successively are as follows: u4, u5, u9, u3, u6, u8, u2,
U1, u7 and u10.Wherein, N is positive integer, and a fixed value can be redefined for according to the empirical value of application scenarios, can also be with
It is that an identified dynamic value is adjusted according to the real trade number dynamic of user each in sequence.For example, N can be set as
Fixed value 5, thus the result of central user is u4, u5, u9, u3 and u6.In the present embodiment, since the setting of fixed value is general
Depending on the empirical value under the scene, it is thus possible to appearance and situation unmatched in certain processing, such as aforementioned N is to fix
The case where when value 5, the transaction count of central user u6 and the gap of non-central user are smaller, and the gap with other central users
It is larger instead, it is thus possible to the association user weaker with sample trade company relevance to be determined as central user.
In order to enable correlation degree when determining central user between more more enough reflections and sample trade company, it can be based on each
The dynamic value that the order of magnitude of a association user determines.Firstly, determining transaction count according to the transaction count of the association user
The order of magnitude.For example, being indicated with the metric order of magnitude, wherein 2=2 × 100, its order of magnitude be 0,21=2.1 × 101, its
The order of magnitude is 1,105=1.05 × 102, its order of magnitude be 2, and so on, can get the number of each association user transaction count
Magnitude, as shown in the table:
u1 |
u2 |
u3 |
u4 |
u5 |
u6 |
u7 |
u8 |
u9 |
u10 |
0 |
0 |
1 |
2 |
1 |
0 |
0 |
0 |
1 |
0 |
Then, according to the order of magnitude of the transaction count, the higher center of quantification grade is used in the association user
Association user of the order of magnitude 1 or more can be chosen in family, such as the present embodiment, the N thereby determined that is 4, corresponding center
User is u3, u4, u5, u9.Compared to the mode of fixed value, when determining central user, can be obtained more based on the order of magnitude
Representational data, the correlation degree between the central user determined and sample trade company can be significantly hotter than other non-central use
Family, so that subsequent processing knot is more accurate.Here, it will be appreciated by those skilled in the art that above-mentioned according to the order of magnitude
Determine that concrete mode when central user is only for example, the existing or other deformations or expansion based on similar principles that occur from now on
If exhibition mode can be suitable for the application, also should include comprising within the scope of protection of this application, and in the form of reference
In this.For example, the threshold value of transaction count can be set, such as 20,80, if the transaction count of association user is more than the threshold value,
Determine it as central user.2 × 10 be can be understood as due to 201, 80 can be understood as=8 × 101, therefore, such mode
It can be understood as the extension of the aforementioned mode based on the order of magnitude, thus include within the scope of protection of this application.
After determining central user, user cluster can be updated with this, such as in the user cluster include at this time
User can be updated to u3, u4, u5, u9.
Step S103 obtains the trade company that transaction occurred with the central user in the second time window, as association
Trade company.Wherein, second time window is a period for being different from the first time window, can be by selecting again
It takes the mode of a period to determine, can also be determined by the way of sliding first time window.For example, when described first
Between window be 20 points to 21 points, window can be slided backward into half an hour, thereby determine that the second time window be 20 thirty extremely
21 thirty.
Since identified central user is that there are the stronger users of correlation degree with sample trade company, with these centers
User occurred have the biggish trade companies may comprising other and sample trade company with like attribute in the trade company of transaction.Example
Such as, when sample trade company is gambling trade company, in the association user based on determined by respective center user, having very big may include
Others gambling trade company, therefore new gambling trade company can be found in this manner.But due to the user of related to gambling activities in addition into
It bribes except winning, it is also possible to other normal trading activities, such as purchase daily life articles etc. are had, it is identified at this time
It equally can include that normal trade company needs to lead to avoid for normal trade company being erroneously identified as gambling trade company in association user
Later continuous step is accurately identified.
Step S104 determines the association trade company and the sample trade company according to the incidence relation between trade company and user
Between similarity.The quantity of identified association trade company is 3 in the embodiment of the present application, and respectively m3, m4 and m5, Fig. 2 is
The schematic diagram for successively obtaining central user by sample trade company, being associated with trade company, node therein is trade company or user, between node
Line indicate that trading activity occurred between trade company and user.Trading activity is able to reflect out being associated between trade company and user
Relationship, such as multiple central users all only be associated withs trade company with one trading activity have occurred, it may be considered that the association trade company and
Meeting similarity with higher, logic are between sample of users: central user is higher with sample of users correlation degree
User, the incidence relation that trading activity is shown can reflect the attribute of trade company, such as gambling quotient to a certain extent
The corresponding central user in family is some users frequently to gamble, these users are likely to also join at other gambling trade companies
Gambling, if multiple central users are all only traded with a trade company in a period of time, then this trade company is gambling trade company
Probability will be high, i.e. the similarity of the trade company and sample trade company is higher.
When judging similarity, can be according to following principle: the quantity of multiple central users be more, then is associated with trade company and sample
Trade company has the probability of like attribute also can be higher, and the similarity for being associated with trade company and sample trade company is higher.In addition, if multiple centers
Also with other users transaction occurred for part or all users in user, then other than transaction occurs with the trade company
Probability can reduce, and the similarity for being associated with trade company and sample trade company is higher.It as a result, can be according to trade company and use based on above principle
Incidence relation between family determines the similarity between the association trade company and the sample trade company.
It, can be between the expression trade company that formalized based on Swing structure and user in some embodiments of the present application
Incidence relation.Friendship occurred to being associated with trade company with one for the user of two central users of Swing representation composition
Easily, the structure that Swing structure shows as the node of two users in Fig. 2, the node of trade company and its line are constituted, example
As u1, u2 and m3 and its between the structure that is constituted of line.
Based on the Swing structure, in some embodiments of the present application, the phase between association trade company and sample trade company is being determined
When seemingly spending, processing step as shown in Figure 3 can be used, comprising:
Step S301 obtains the first quantity of Swing structure relevant to the first association trade company.Due to calculating association quotient
Similarity between family and the sample trade company needs to calculate the similarity between each association trade company and sample trade company one by one,
First association trade company is currently calculated trade company, such as when calculating the similarity between m3 and sample trade company, should
M3 is the first association trade company.By taking scene shown in Fig. 2 as an example, wherein the quantity of Swing structure relevant to m3 is 1, i.e.,
The Swing structure being made of u1, u2 and m3.
Step S302 is obtained to the first user to the second quantity of relevant other Swing structures.Wherein, described first
User is to for user couple corresponding to the first relevant Swing structure of association trade company, such as the first association trade company m3, with it
Relevant Swing structure is the Swing structure being made of u1, u2 and m3, in the Swing structure corresponding user to for u1 and
U2. it follows that corresponding first user is to as u1 and u2 for the first association trade company m3.
To the first user in this present embodiment to u1 and u2, relevant other Swing structures are by u1, u2 and m5 institute
The Swing structure of composition, it is possible thereby to determine that the second quantity of other Swing structures relevant to u1 and u2 to the first user is
1。
Step S303, according to first quantity and the second quantity, determine the first association trade company and the sample trade company it
Between similarity.Based on principle when judging similarity it was determined that first is associated with the phase between trade company and the sample trade company
It is positively correlated like degree and first quantity, and negatively correlated with the second quantity.Therefore, first can be calculated based on following mode
The similarity being associated between trade company and the sample trade company: S=nX-mY, wherein n is the first quantity, and m is the second quantity,
X is positively related parameter, for indicating each positive influences of the Swing structure relevant to the first association trade company for similarity
Degree, Y are negative relevant parameter, for indicate each to the first user to relevant other Swing structures for similarity
Negative effect degree.X and Y can be based on specific application scenarios, in conjunction with the difference in scene because usually determining.
Here, those skilled in the art are it should be understood that the mode of above-mentioned calculating similarity is only for example, it is existing or modern
If the other deformations based on similar principles occurred afterwards or extended mode can be suitable for the application, it should also be included in this
In the protection scope of application, and it is incorporated herein in the form of reference.For example, Swing structure is for similarity in actual scene
Influence may be more complicated, therefore can be in conjunction with the other information of user and/or trade company in related Swing structure, such as
Evaluation information, other transaction records, credit value in transaction platform etc. individually assess each Swing structure for similarity
Positive influences or negative effect etc., similarity final is determined with this.
By taking scene shown in Fig. 2 as an example, for being associated with trade company m3, corresponding first quantity is that the 1, second quantity is 1;It is right
In association trade company m4, corresponding first quantity is that the 1, second quantity is 0;For being associated with trade company m5, corresponding first quantity
It is 1 for the 3, second quantity.It is possible thereby to which the similarity calculated between each association trade company and sample trade company is as shown in the table:
Trade company |
m3 |
m4 |
m5 |
First quantity |
1 |
1 |
3 |
Second quantity |
1 |
0 |
1 |
Similarity |
X-Y |
X |
3X-Y |
Step S105 determines that similarity meets the target trade company of preset condition in the association trade company.Wherein, described pre-
If condition can be set according to the demand of practical application scene, such as can be ranked up according to the specific value of similarity, choosing
It takes the forward association trade company that sorts as target trade company, if in the present embodiment, the value for setting the X is greater than Y, then can determine
Sequence after similarity S1, S2 and S3 sequence corresponding to three associations trade company m3, m4 and m5 are as follows: S3 > S2 > S1, if choosing
The first two data for wherein sorting forward, then identified target trade company is m4 and m5.
Alternatively, a threshold value can also be set, the association trade company that similarity is more than or equal to the threshold value is determined as target quotient
Family, such as given threshold Z, then can be by being respectively compared the size of S1, S2, S3 and Z, so that it is determined that target trade company.Here, this
Field technical staff it should be understood that the concrete form of above-mentioned preset condition is only for example, it is existing or occur from now on based on
If other deformations of similar principles or extended mode can be suitable for the application, it should also be included in the protection model of the application
In enclosing, and it is incorporated herein in the form of reference.
In some embodiments of the present application, according to the incidence relation between trade company and user, the association trade company is determined
When similarity between the sample trade company, can also in conjunction with the aforementioned mode for filtering out central user from association user,
First from association trade company in filter out center trade company, then with center trade company substitution script association trade company, calculate similarity and from
In identify target trade company.Fig. 4 shows another trade company recognition methods provided by the embodiments of the present application, including at least following
Processing step:
Step S401 obtains the user that transaction occurred with sample trade company in first time window, as association user.
Step S402 determines that transaction count is more in the association user according to the transaction count of the association user
Central user.
Step S403 obtains the trade company that transaction occurred with the central user in the second time window, as association
Trade company.
Step S404 determines that transaction count is more in the association trade company according to the transaction count of the association trade company
Center trade company.
Step S405 determines the center trade company and the sample trade company according to the incidence relation between trade company and user
Between similarity.
Step S406 determines that similarity meets the target trade company of preset condition in the center trade company.
Wherein, described when determining the center trade company, can according to association trade company transaction count use in determination
Mode as heart user class.For example, first determining the order of magnitude of transaction count, then according to the transaction count of the association trade company
According to the order of magnitude of the transaction count, the higher center trade company of quantification grade in the association trade company, so that quotient
The recognition result at family is more accurate.
In order to enable recognition result is more accurate, in some embodiments of the present application, can be based on above-mentioned any one
Kind scheme is iterated processing, i.e., the sample that will execute target trade company identifying processing as after determined by an identifying processing
This trade company, and slide after the first time window makes time window change, it is iterated processing, until processing obtains
The target trade company convergence obtained, it is possible thereby to which the error rate of identification is substantially reduced.Fig. 5 is the schematic diagram of iterative processing.Wherein,
When judging whether target trade company restrains, can be compared based on the result of this identification with result before at least once,
If these results are consistent or difference is within the scope of setting, it may be considered that the target trade company convergence that processing obtains.
For example, 50 target trade companies are identified after the 10th iterative processing, 50 target trade companies and preceding primary knowledge
The target trade company (i.e. this sample trade company) not obtained is identical, it is possible thereby to which the target trade company for thinking that this processing obtains receives
It holds back, as final recognition result.Also such as, after the 7th iterative processing, 100 target trade companies are identified, with the 6th
The recognition result of secondary processing all only has 1 target trade company difference, has 2 target trade companies different from the recognition result of the 5th processing,
Judge that convergent condition is the difference with recognition result twice before within 3% if setting, it may be considered that the 7th iteration
The target trade company that processing obtains has restrained, and can be used as final recognition result.Here, those skilled in the art should can manage
Solution, it is above-mentioned to judge that the whether convergent mode of target trade company is only for example, it is existing or occur from now on based on the other of similar principles
It, also should be comprising within the scope of protection of this application, and with reference if deformation or extended mode can be suitable for the application
Form be incorporated herein.
In addition, can be used in the row of identification with user the embodiment of the present application also provides a kind of data object recognition methods
For there are associated data objects, such as can be the trade company with user there are trading activity, be furthermore also possible to can by with
Webpage that family thumbs up, can be by commodity of user's evaluation etc..Similar with the scene that trade company identifies, other behaviors with user, which exist, closes
The data object of connection can also be based on sample data object, find central user, and then find there is class with sample data object
Like the associated data object of attribute, and according between user and data object be based on user behavior caused by incidence relation,
The similarity between associated data object and sample data object is determined, to complete to identify.
Fig. 6 shows the process flow of data object recognition methods provided by the embodiments of the present application, may include following place
Manage step:
Step S601 obtains the user that correlating event occurred with sample data object in first time window, as
Association user.
Step S602 determines association according to the correlating event frequency of the association user in the association user
The more central user of event frequency.
Step S603 obtains the data object that correlating event occurred with the central user in the second time window,
As associated data object.
Step S604, according to the incidence relation between data object and user, determine the associated data object with it is described
Similarity between sample data object.
Step S605 determines that similarity meets the target data objects of preset condition in the associated data object.
The program can extend to other certain behaviors with user there are associated data objects as a result, to identify
The data object with particular community is accurately identified, so that application scenarios are more extensive.Such as it is disliked for identification by user
Anticipate the webpage thumbed up, by commodity etc. of user's malice evaluation, carry out wind so as to the behavior to malice brush temperature, brush evaluation
Dangerous prevention and control.It is similar with the scheme that trade company identifies, it can be used in improving the extension side of identification accuracy in aforementioned trade company's identification scene
The schemes such as case, such as iterative processing, extraction center trade company also can be applied in the identifying schemes of data object, since it is related to
Principle it is similar, details are not described herein again.
Based on the same inventive concept, a kind of trade company's identification equipment is additionally provided in the embodiment of the present application, the equipment is corresponding
Method be trade company's recognition methods in previous embodiment, and its principle solved the problems, such as is similar to this method.
Trade company's identification equipment provided by the embodiments of the present application can be based on the trading activity of user, and determining and sample trade company can
There can be the association trade company of like attribute, such as sample trade company is risk trade company, then be associated with trade company and be likely to be association quotient
Family.On this basis, the association trade company and the sample quotient are further determined according to the incidence relation between trade company and user
Similarity between family, and denoised based on similarity, the lower association trade company of similarity is excluded, to promote trade company's identification
Accuracy.
In actual scene, trade company identification equipment can be user equipment, the network equipment or user equipment and network
Equipment is integrated constituted equipment by network.Wherein, the user equipment includes but is not limited to personal computer, intelligent hand
All kinds of terminal devices such as machine, tablet computer, the network equipment include but is not limited to as network host, single network server,
Multiple network server collection or the set of computers based on cloud computing etc. are realized, at part when can be used to implement setting alarm clock
Manage function.Here, cloud is made of a large amount of hosts or network server for being based on cloud computing (Cloud Computing), wherein cloud
Calculating is one kind of distributed computing, a virtual machine consisting of a loosely coupled set of computers.
Fig. 7 shows a kind of structure of trade company's identification equipment provided by the embodiments of the present application, includes at least: the first association mould
Block 710, screening module 720, the second relating module 730 and denoising module 740.
First relating module 710 is used to obtain the use that transaction occurred with sample trade company in first time window
Family, as association user.Wherein, the sample trade company is the sample according to selected by the demand of identification scene, these samples can
To obtain the target trade company with like attribute in order to identify with certain specific attributes.For example, the attribute can be and relate to
Dislike gambling, then it can be using the gambling trade company having been acknowledged as sample trade company, if user, which needs to identify, is accused of swindle row
For trade company, then can be using the swindle trade company having been acknowledged as sample trade company.
In actual scene, the set of trade company can be denoted as merchant cluster, and the set of user can be denoted as user collection
Group, at this point, the trade company in the merchant cluster is sample trade company, determining association user can be put into user cluster.
The first time window refers to a preset period, and the length of the period can be according to the demand of actual scene
Setting, for example, it can be set to being 1 hour, 90 minutes, 2 hours etc..
If in the embodiment of the present application, the first time window that sets as the period of 20 points to 21 points of hour,
Merchant cluster includes 2 sample trade company m1 and m2, the two sample trade companies are gambling trade company.By transferring the two samples
The transaction data of payment platform where this trade company can count and all friendship occur between 20 points to 21 points with m1 or m2
Easy user, such as the result of association user obtained in embodiment is to be denoted as u1-10, i.e. user cluster comprising 10 users
In include tetra- elements of u1-10.
The screening module 720 is used for the transaction count according to the association user, determines and hands in the association user
The more central user of easy number.
In actual scene, for certain a kind of trade company with like attribute, generally all can some frequently trade
User group, such as trade company of gambling, the user of transaction may include some users to gamble once in a while, it also can be comprising frequent
The user to gamble.Wherein, the user to gamble once in a while participates in gambling at other gambling trade companies since the number of gambling is few
Probability is also relatively small, conversely, the user frequently to gamble is more likely to since the number for participating in gambling is more in different gamblings
Bo Shanghuchu participates in gambling.It therefore, is frequency by the more central user of transaction count under the scene of gambling trade company identification
A possibility that numerous user to gamble, the other gambling trade companies that can be found by these users, can be bigger, needs according to transaction time
Number determines central user in association user.
In actual scene, screening module 720 can determine transaction count in the association user in the following way
More central user.For example, can be ranked up according to the transaction count of association user, then according to the result of sequence from sequence
Chosen in column it is N number of, centered on user.By taking the user in aforementioned user cluster as an example, the transaction count of each user can be as
Shown in following table:
u1 |
u2 |
u3 |
u4 |
u5 |
u6 |
u7 |
u8 |
u9 |
u10 |
2 |
3 |
21 |
105 |
99 |
6 |
1 |
5 |
63 |
1 |
It follows that according to the ranking results of transaction count from greatly to it is small successively are as follows: u4, u5, u9, u3, u6, u8, u2,
U1, u7 and u10.Wherein, N is positive integer, and a fixed value can be redefined for according to the empirical value of application scenarios, can also be with
It is that an identified dynamic value is adjusted according to the real trade number dynamic of user each in sequence.For example, N can be set as
Fixed value 5, thus the result of central user is u4, u5, u9, u3 and u6.In the present embodiment, since the setting of fixed value is general
Depending on the empirical value under the scene, it is thus possible to appearance and situation unmatched in certain processing, such as aforementioned N is to fix
The case where when value 5, the transaction count of central user u6 and the gap of non-central user are smaller, and the gap with other central users
It is larger instead, it is thus possible to the association user weaker with sample trade company relevance to be determined as central user.
In order to enable correlation degree when determining central user between more more enough reflections and sample trade company, it can be based on each
The dynamic value that the order of magnitude of a association user determines.Firstly, screening module can be according to the transaction count of the association user, really
Determine the order of magnitude of transaction count.For example, being indicated with the metric order of magnitude, wherein 2=2 × 100, its order of magnitude be 0,21
=2.1 × 101, its order of magnitude be 1,105=1.05 × 102, its order of magnitude be 2, and so on, can get each association user
The order of magnitude of transaction count, as shown in the table:
u1 |
u2 |
u3 |
u4 |
u5 |
u6 |
u7 |
u8 |
u9 |
u10 |
0 |
0 |
1 |
2 |
1 |
0 |
0 |
0 |
1 |
0 |
Then, screening module can be according to the order of magnitude of the transaction count, the quantification grade in the association user
Higher central user, such as association user of the order of magnitude 1 or more can be chosen in the present embodiment, the N thereby determined that is
4, corresponding central user is u3, u4, u5, u9.Compared to the mode of fixed value, based on the order of magnitude come when determining central user,
More representational data can be obtained, the correlation degree between the central user determined and sample trade company can be significantly hotter than
Other non-central users, so that subsequent processing knot is more accurate.Here, it will be appreciated by those skilled in the art that above-mentioned
Determine that concrete mode when central user is only for example according to the order of magnitude, existing or its based on similar principles that occurs from now on
It, also should be comprising within the scope of protection of this application, and to draw if it is deformed or extended mode can be suitable for the application
Form is incorporated herein.For example, the threshold value of transaction count can be set, such as 20,80, if the transaction count of association user is super
The threshold value is crossed, then determines it as central user.2 × 10 be can be understood as due to 201, 80 can be understood as=8 × 101, because
This, such mode is it can be appreciated that be the extension of the aforementioned mode based on the order of magnitude, thus be included in the protection model of the application
In enclosing.
After determining central user, user cluster can be updated with this, such as in the user cluster include at this time
User can be updated to u3, u4, u5, u9.
With the central user transaction occurred for second relating module 730 in the second time window for obtaining
Trade company, as association trade company.Wherein, second time window is a period for being different from the first time window,
It can be determined, can also be determined by the way of sliding first time window by way of choosing a period again.
For example, the first time window is 20 points to 21 points, window can be slided backward into half an hour, thereby determine that for the second time
Window is 20 thirty to 21 thirty.
Since identified central user is that there are the stronger users of correlation degree with sample trade company, with these centers
User occurred have the biggish trade companies may comprising other and sample trade company with like attribute in the trade company of transaction.Example
Such as, when sample trade company is gambling trade company, in the association user based on determined by respective center user, having very big may include
Others gambling trade company, therefore new gambling trade company can be found in this manner.But due to the user of related to gambling activities in addition into
It bribes except winning, it is also possible to other normal trading activities, such as purchase daily life articles etc. are had, it is identified at this time
It equally can include that normal trade company needs to lead to avoid for normal trade company being erroneously identified as gambling trade company in association user
Later continuous step is accurately identified.
The denoising module 740 is used to determine the association trade company and institute according to the incidence relation between trade company and user
State the similarity between sample trade company.The quantity of identified association trade company is 3, respectively m3, m4 in the embodiment of the present application
And m5, Fig. 2 are the schematic diagram for successively obtaining central user by sample trade company, being associated with trade company, node therein is trade company or use
Family, the line between node indicate that trading activity occurred between trade company and user.Trading activity is able to reflect out trade company and uses
Incidence relation between family, such as all only be associated with trade company with one has occurred trading activity to multiple central users, it may be considered that
Meeting similarity with higher, logic are between the association trade company and sample of users: central user is and sample of users is closed
The higher user of connection degree, the incidence relation that trading activity is shown can reflect the category of trade company to a certain extent
Property, such as the corresponding central user of gambling trade company is some users frequently to gamble, these users are likely in other gamblings
Bo Shanghuchu also gambles, if multiple central users are all only traded with a trade company in a period of time, then this quotient
Family is that the probability of gambling trade company will be high, i.e. the similarity of the trade company and sample trade company is higher.
When judging similarity, can be according to following principle: the quantity of multiple central users be more, then is associated with trade company and sample
Trade company has the probability of like attribute also can be higher, and the similarity for being associated with trade company and sample trade company is higher.In addition, if multiple centers
Also with other users transaction occurred for part or all users in user, then other than transaction occurs with the trade company
Probability can reduce, and the similarity for being associated with trade company and sample trade company is higher.It as a result, can be according to trade company and use based on above principle
Incidence relation between family determines the similarity between the association trade company and the sample trade company.
It, can be between the expression trade company that formalized based on Swing structure and user in some embodiments of the present application
Incidence relation.Friendship occurred to being associated with trade company with one for the user of two central users of Swing representation composition
Easily, the structure that Swing structure shows as the node of two users in Fig. 2, the node of trade company and its line are constituted, example
As u1, u2 and m3 and its between the structure that is constituted of line.
Based on the Swing structure, in some embodiments of the present application, denoising module is determining association trade company and sample trade company
Between similarity when, processing step as shown in Figure 3 can be used, comprising:
Step S301 obtains the first quantity of Swing structure relevant to the first association trade company.Due to calculating association quotient
Similarity between family and the sample trade company needs to calculate the similarity between each association trade company and sample trade company one by one,
First association trade company is currently calculated trade company, such as when calculating the similarity between m3 and sample trade company, should
M3 is the first association trade company.By taking scene shown in Fig. 2 as an example, wherein the quantity of Swing structure relevant to m3 is 1, i.e.,
The Swing structure being made of u1, u2 and m3.
Step S302 is obtained to the first user to the second quantity of relevant other Swing structures.Wherein, described first
User is to for user couple corresponding to the first relevant Swing structure of association trade company, such as the first association trade company m3, with it
Relevant Swing structure is the Swing structure being made of u1, u2 and m3, in the Swing structure corresponding user to for u1 and
U2. it follows that corresponding first user is to as u1 and u2 for the first association trade company m3.
To the first user in this present embodiment to u1 and u2, relevant other Swing structures are by u1, u2 and m5 institute
The Swing structure of composition, it is possible thereby to determine that the second quantity of other Swing structures relevant to u1 and u2 to the first user is
1。
Step S303, according to first quantity and the second quantity, determine the first association trade company and the sample trade company it
Between similarity.Based on principle when judging similarity it was determined that first is associated with the phase between trade company and the sample trade company
It is positively correlated like degree and first quantity, and negatively correlated with the second quantity.Therefore, first can be calculated based on following mode
The similarity being associated between trade company and the sample trade company: S=nX-mY, wherein n is the first quantity, and m is the second quantity,
X is positively related parameter, for indicating each positive influences of the Swing structure relevant to the first association trade company for similarity
Degree, Y are negative relevant parameter, for indicate each to the first user to relevant other Swing structures for similarity
Negative effect degree.X and Y can be based on specific application scenarios, in conjunction with the difference in scene because usually determining.
Here, those skilled in the art are it should be understood that the mode of above-mentioned calculating similarity is only for example, it is existing or modern
If the other deformations based on similar principles occurred afterwards or extended mode can be suitable for the application, it should also be included in this
In the protection scope of application, and it is incorporated herein in the form of reference.For example, Swing structure is for similarity in actual scene
Influence may be more complicated, therefore can be in conjunction with the other information of user and/or trade company in related Swing structure, such as
Evaluation information, other transaction records, credit value in transaction platform etc. individually assess each Swing structure for similarity
Positive influences or negative effect etc., similarity final is determined with this.
By taking scene shown in Fig. 2 as an example, for being associated with trade company m3, corresponding first quantity is that the 1, second quantity is 1;It is right
In association trade company m4, corresponding first quantity is that the 1, second quantity is 0;For being associated with trade company m5, corresponding first quantity
It is 1 for the 3, second quantity.It is possible thereby to which the similarity calculated between each association trade company and sample trade company is as shown in the table:
Trade company |
m3 |
m4 |
m5 |
First quantity |
1 |
1 |
3 |
Second quantity |
1 |
0 |
1 |
Similarity |
X-Y |
X |
3X-Y |
In addition, denoising module 740 is also used to determine that similarity meets the target quotient of preset condition in the association trade company
Family.Wherein, the preset condition can be set according to the demand of practical application scene, such as can be according to the specific number of similarity
Value is ranked up, and chooses the forward association trade company that sorts as target trade company, if in the present embodiment, the value for setting the X is big
Sequence in Y, then after can determining similarity S1, S2 and S3 sequence corresponding to three associations trade company m3, m4 and m5 are as follows: S3 >
S2 > S1, if choosing the first two data for wherein sorting forward, identified target trade company is m4 and m5.
Alternatively, a threshold value can also be set, the association trade company that similarity is more than or equal to the threshold value is determined as target quotient
Family, such as given threshold Z, then can be by being respectively compared the size of S1, S2, S3 and Z, so that it is determined that target trade company.Here, this
Field technical staff it should be understood that the concrete form of above-mentioned preset condition is only for example, it is existing or occur from now on based on
If other deformations of similar principles or extended mode can be suitable for the application, it should also be included in the protection model of the application
In enclosing, and it is incorporated herein in the form of reference.
In some embodiments of the present application, the second relating module is determined according to the incidence relation between trade company and user
When the similarity being associated between trade company and the sample trade company, center can also be filtered out from association user in conjunction with aforementioned
The mode of user first filters out center trade company from association trade company, then with the association trade company of center trade company substitution script, calculates
Similarity simultaneously therefrom identifies target trade company.As a result, in another trade company's identification equipment provided by the embodiments of the present application, described the
Two relating modules be used for according to it is described association trade company transaction count, in the association trade company determination transaction count it is more in
Heart trade company, and according to the incidence relation between trade company and user, determine the phase between the center trade company and the sample trade company
Determine that similarity meets the target trade company of preset condition like degree, and in the center trade company.
Wherein, described when determining the center trade company, can according to association trade company transaction count use in determination
Mode as heart user class.For example, first determining the order of magnitude of transaction count, then according to the transaction count of the association trade company
According to the order of magnitude of the transaction count, the higher center trade company of quantification grade in the association trade company, so that quotient
The recognition result at family is more accurate.
In order to enable recognition result is more accurate, in the trade company's identification equipment provided in some embodiments of the present application,
It can also include iteration control module, be used for using the target trade company as sample trade company, and slide the first time window,
It controls first relating module, screening module, the second relating module and denoising module is iterated processing until processing obtains
Target trade company convergence.
By iteration control module, it can will execute target trade company determined by an identifying processing and once be identified as after
The sample trade company of processing, and slide after the first time window makes time window change, it is iterated processing, directly
The target trade company convergence obtained to processing, it is possible thereby to which the error rate of identification is substantially reduced.Fig. 5 is the signal of iterative processing
Figure.Wherein, when judging whether target trade company restrains, can based on this identification result with before at least once result into
Row compares, if these results are consistent or difference is within the scope of setting, it may be considered that the target trade company that processing obtains receives
It holds back.
For example, 50 target trade companies are identified after the 10th iterative processing, 50 target trade companies and preceding primary knowledge
The target trade company (i.e. this sample trade company) not obtained is identical, it is possible thereby to which the target trade company for thinking that this processing obtains receives
It holds back, as final recognition result.Also such as, after the 7th iterative processing, 100 target trade companies are identified, with the 6th
The recognition result of secondary processing all only has 1 target trade company difference, has 2 target trade companies different from the recognition result of the 5th processing,
Judge that convergent condition is the difference with recognition result twice before within 3% if setting, it may be considered that the 7th iteration
The target trade company that processing obtains has restrained, and can be used as final recognition result.Here, those skilled in the art should can manage
Solution, it is above-mentioned to judge that the whether convergent mode of target trade company is only for example, it is existing or occur from now on based on the other of similar principles
It, also should be comprising within the scope of protection of this application, and with reference if deformation or extended mode can be suitable for the application
Form be incorporated herein.
In addition, the embodiment of the present application also provides a kind of data objects to identify equipment, it can be used in row of the identification with user
For there are associated data objects, such as can be the trade company with user there are trading activity, be furthermore also possible to can by with
Webpage that family thumbs up, can be by commodity of user's evaluation etc..Similar with the scene that trade company identifies, other behaviors with user, which exist, closes
The data object of connection can also be based on sample data object, find central user, and then find there is class with sample data object
Like the associated data object of attribute, and according between user and data object be based on user behavior caused by incidence relation,
The similarity between associated data object and sample data object is determined, to complete to identify.
The structure of the data object identification equipment is similar with trade company above-mentioned identification equipment, may include the first association mould
Block, screening module, the second relating module and denoising module.Wherein, the first relating module, for obtaining in first time window
The user that correlating event occurred with sample data object, as association user;Screening module is used for according to the association user
Correlating event frequency, the more central user of correlating event frequency is determined in the association user;Second closes
Gang mould block is for obtaining the data object that correlating event occurred with the central user in the second time window, as association
Data object;Module is denoised to be used for according to the incidence relation between data object and user, determine the associated data object with
Similarity between the sample data object, and determine that similarity meets the mesh of preset condition in the associated data object
Mark data object.
The program can extend to other certain behaviors with user there are associated data objects as a result, to identify
The data object with particular community is accurately identified, so that application scenarios are more extensive.Such as it is disliked for identification by user
Anticipate the webpage thumbed up, by commodity etc. of user's malice evaluation, carry out wind so as to the behavior to malice brush temperature, brush evaluation
Dangerous prevention and control.It is similar with the scheme that trade company identifies, it can be used in improving the extension side of identification accuracy in aforementioned trade company's identification scene
The schemes such as case, such as iterative processing, extraction center trade company also can be applied in the identifying schemes of data object, since it is related to
Principle it is similar, details are not described herein again.
In addition, a part of the application can be applied to computer program product, such as computer program instructions, when its quilt
When computer executes, by the operation of the computer, it can call or provide according to the present processes and/or technical solution.
And the program instruction of the present processes is called, it is possibly stored in fixed or moveable recording medium, and/or pass through
Broadcast or the data flow in other signal-bearing mediums and transmitted, and/or be stored according to program instruction run calculating
In the working storage of machine equipment.Here, include a calculating equipment as shown in Figure 8 according to some embodiments of the present application,
The equipment includes being stored with one or more memories 810 of computer-readable instruction and for executing computer-readable instruction
Processor 820, wherein when the computer-readable instruction is executed by the processor, so that the equipment, which executes, is based on aforementioned
The method and/or technology scheme of multiple embodiments of application.
In addition, some embodiments of the present application additionally provide a kind of computer-readable medium, it is stored thereon with computer journey
Sequence instruction, the computer-readable instruction can be executed by processor with the method for realizing multiple embodiments of aforementioned the application and/
Or technical solution.
It should be noted that the application can be carried out in the assembly of software and/or software and hardware, for example, can adopt
With specific integrated circuit (ASIC), general purpose computer or any other realized similar to hardware device.In some embodiments
In, the software program of the application can be executed by processor to realize above step or function.Similarly, the software of the application
Program (including relevant data structure) can be stored in computer readable recording medium, for example, RAM memory, magnetic or
CD-ROM driver or floppy disc and similar devices.In addition, hardware can be used to realize in some steps or function of the application, for example,
As the circuit cooperated with processor thereby executing each step or function.
It is obvious to a person skilled in the art that the application is not limited to the details of above-mentioned exemplary embodiment, Er Qie
In the case where without departing substantially from spirit herein or essential characteristic, the application can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and scope of the present application is by appended power
Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims
Variation is included in the application.Any reference signs in the claims should not be construed as limiting the involved claims.This
Outside, it is clear that one word of " comprising " does not exclude other units or steps, and odd number is not excluded for plural number.That states in device claim is multiple
Unit or device can also be implemented through software or hardware by a unit or device.The first, the second equal words are used to table
Show title, and does not indicate any particular order.