Summary of the invention
In view of this, the embodiment of the present invention provides a kind of user identification method and device, can be used in multi-user sameIn the case where virtual account, multiple user models are constructed, and according to the similarity identification user between multiple user models, are reducedThe superposition behavior of multiple users improves the accuracy of user model to building user model and the interference of user is identifiedAnd the accuracy of identification user.
To achieve the above object, according to an aspect of an embodiment of the present invention, a kind of user identification method is provided, is wrappedIt includes: classifying to the historical data of virtual account, obtain the historical data of multiple classifications；According to the historical data of each classificationUser model is constructed, to obtain multiple user models corresponding with the virtual account；According between the multiple user modelSimilarity, carry out user's identification.
Optionally, carrying out classification to the historical data of virtual account includes: the history according to the time cycle to virtual accountData are classified, and are a classification with the historical data in a period of time；Or according to time cycle and unique identification to voidThe historical data of quasi- account is classified, and is a classification with the historical data of unique identification identical in a period of time.
Optionally, according to the similarity between the multiple user model, carrying out user's identification includes: for each userModel calculates the similarity between the user model and other users model；By similarity less than the first similarity thresholdSimilar users model of the other users model as the user model；Calculate user's average similarity of the user model withAnd user's average similarity of the similar users model of the user model；According to user's average similarity of the user modelAnd user's average similarity of the similar users model, carry out user's identification.
Optionally, for each user model, the similarity between the user model and other users model is according to such asUnder type calculates:
Wherein, K[N, M]Represent the similarity between user model userN and user model userM, SuserNRepresent user's mouldThe target weight of type userN, SuserMRepresent the target weight of user model userM.
Optionally, the difference between the target weight of each user model and the preset target weight referring to user modelLess than or equal to weight threshold.
To achieve the above object, according to another aspect of an embodiment of the present invention, a kind of customer identification device is provided, is wrappedInclude: data categorization module is classified for the historical data to virtual account, obtains the historical data of multiple classifications；ModelModule is constructed, it is corresponding more with the virtual account to obtain for constructing user model according to the historical data of each classificationA user model；Subscriber identification module, for carrying out user's identification according to the similarity between the multiple user model.
Optionally, the data categorization module is also used to: being divided according to historical data of the time cycle to virtual accountClass is a classification with the historical data in a period of time；Or virtual account is gone through according to time cycle and unique identificationHistory data are classified, and are a classification with the historical data of unique identification identical in a period of time.
Optionally, the subscriber identification module is also used to: being directed to each user model, is calculated the user model and otherSimilarity between user model；Other users model using similarity less than the first similarity threshold is as the user modelSimilar users model；Calculate user's average similarity of the user model and the similar users model of the user modelUser's average similarity；It is flat according to the user of user's average similarity of the user model and the similar users modelEqual similarity carries out user's identification.
Optionally, for each user model, the similarity between the user model and other users model is according to such asUnder type calculates:
Wherein, K[N, M]Represent the similarity between user model userN and user model userM, SuserNRepresent user's mouldThe target weight of type userN, SuserMRepresent the target weight of user model userM.
Optionally, the difference between the target weight of each user model and the preset target weight referring to user modelLess than or equal to weight threshold.
To achieve the above object, according to an embodiment of the present invention in another aspect, providing a kind of electronic equipment, comprising: oneA or multiple processors；Storage device, for storing one or more programs, when one or more of programs are oneOr multiple processors execute, so that one or more of processors realize user identification side provided by any of the above-described embodimentMethod.
To achieve the above object, another aspect according to an embodiment of the present invention, provides a kind of computer-readable medium,On be stored with computer program, user identification side provided by any of the above-described embodiment is realized when described program is executed by processorMethod.
One embodiment in foregoing invention has the following advantages that or the utility model has the advantages that because using the history to virtual accountData are classified, and the historical data of multiple classifications is obtained；User model is constructed according to the historical data of each classification, to obtainMultiple user models corresponding with the virtual account；According to the similarity between the multiple user model, user's knowledge is carried outOther technological means, so overcoming the technical problem of user model inaccuracy in the prior art, and then it is multiple to have reached reductionInterference of the superposition behavior of user to user model improves the technology of the accuracy of user model and the accuracy of user's identificationEffect.
Further effect possessed by above-mentioned non-usual optional way adds hereinafter in conjunction with specific embodimentWith explanation.
Specific embodiment
Below in conjunction with attached drawing, an exemplary embodiment of the present invention will be described, including the various of the embodiment of the present inventionDetails should think them only exemplary to help understanding.Therefore, those of ordinary skill in the art should recognizeIt arrives, it can be with various changes and modifications are made to the embodiments described herein, without departing from scope and spirit of the present invention.TogetherSample, for clarity and conciseness, descriptions of well-known functions and structures are omitted from the following description.
Fig. 2 is the schematic diagram of the main flow of user identification method according to an embodiment of the present invention, as shown in Fig. 2, the partyMethod includes:
Step S201: classifying to the historical data of virtual account, obtains the historical data of multiple classifications；
Step S202: constructing user model according to the historical data of each classification, corresponding with the virtual account to obtainMultiple user models；
Step S203: according to the similarity between the multiple user model, user's identification is carried out.
Step S201 can be passed through referring to FIG. 4, historical data can be user internet internet log dataWeb crawlers obtains.The channel for obtaining internet internet log data includes but is not limited to: portal website, video website, electric businessWebsite, tour site, forum, social media (such as microblogging, wechat) etc..By taking electric business website as an example, the historical data of acquisition canTo include the detail information (such as type of merchandise, price, purposes) for browsing commodity, the same commodity duration of browsing, history purchase noteIt records, single act of gathering records, (such as exchanges duration using the preferential frequency, search record, with contact staff's communication records, exchanges content)With comment on commodity record etc..By taking video website as an example, historical data may include: search record, browsing record, watch record,Video content details (such as plot, director, performer etc.).
Specifically, can be classified according to historical data of the default rule to virtual account.For example, can according to whenBetween the period classify to the historical data of acquisition, with the historical data in a period of time be a classification.Wherein, week timePhase can flexible setting according to demand, the present invention is herein with no restrictions.
As an example, obtaining 30 days internet log data of some virtual account as historical data, by above-mentioned 30 daysIt is divided into N (N is greater than or equal to 2) a time cycle every day, is one kind with the historical data in a period of time.Further,Above-mentioned 30 days every day is divided into 4 time cycles: T1、T2、T3And T4, wherein T1For 8:00-1400, T2For 14:00-20:00, T3For 20:00-2:00, T4For 02:00-8:00.The historical data that then will acquire is divided into 4 classes, in a period of timeHistorical data is one kind.
In other alternative-embodiments, the historical data of acquisition can be divided according to time cycle and unique identificationClass is a classification with the historical data of unique identification identical in a period of time.Wherein, the unique identification can be equipmentNumber or IP address.
Above equipment number can be IMEI, and (International Mobile Equipment Identit6, the world are mobileEIC equipment identification code).
Above equipment number is also possible to UUID (Universall6Unique Identifier, globally unique identifier),It is also possible to the address MAC (Media Access Control or Medium Access Control).
In an alternate embodiment of the invention, with reference to Fig. 4, before classifying to historical data, this method further include: to historyData are cleaned.
Specifically, can use incomplete data detection method, mistake value detection method, repeating value detection method, consistencyDetection method etc. handles historical data, to improve the quality of data and make data be more suitable for excavating, analyze.
Further, Text Mining Technology, natural language processing method, machine learning algorithm and pre- measuring and calculating can also be utilizedMethod etc. carries out text analyzing to historical data.
For step S202, may include: according to the step of historical data of each classification building user model
According to preset various dimensions tag library, the historical data of each classification is matched to obtain various dimensions label；
Multiple user models corresponding with the historical data of the multiple classification are constructed according to the various dimensions label.
Wherein, various dimensions tag library can be pre-set according to application demand, so as to for specific application,Product and project demands construct user model, and accuracy is higher.Various dimensions label refers to from multiple dimensions reflection user characteristicsLabel.By taking electric business website as an example, various dimensions label includes but is not limited to essential attribute label, behavioural characteristic in the embodiment of the present inventionLabel, purchasing power label, hobby label, psychological characteristics label, social attribute label.Each label includes multiple sonsLabel, such as essential attribute label include but is not limited to following subtab: gender, age bracket, height, weight, marital status, familyFront yard income, occupation；Behavioural characteristic label includes but is not limited to following subtab online hours, login time, purchase frequency etc.；PurchaseThe ability label of buying includes but is not limited to time of following subtab purchase mid and low-end products, high-end product, light luxurious commodity, luxury goodsNumber；Hobby label includes but is not limited to following subtab outdoor sports, body-building, drawing, music, reading etc.；Psychological characteristicsLabel includes but is not limited to resolute following subtab working, indecision, introversion, extroversion etc..Social attribute label includes but notIt is limited to following subtab: family party, friend's party, telephonic communication, common social application.
In the present embodiment, user model is referred to as user's portrait, i.e. user information labeling, is exactly that enterprise passes throughIt collects with after the data of main informations such as analysis consumer's society attribute, living habit, consumer behavior, takes out a userOverall picture.User model is the virtual representations of real user, is built upon on a series of truthful datas.User model is to answerWith the basic mode of big data technology.User model can provide enough Information bases for enterprise, and enterprise can be helped fastSpeed finds the more extensive feedback information such as accurate user group and user demand.
According to preset various dimensions tag library, sorted historical data is matched to obtain every a kind of historical dataVarious dimensions label constructs the corresponding multiple user models of such historical data according to the various dimensions label, to obtain the voidThe quasi- corresponding multiple user models of account.
For step S203, in practical application scene, for the same virtual account, it may be possible to which a real user makesWith, it is also possible to multiple real users use.In the present embodiment, for the same virtual account, multiple user's moulds are constructedType identifies that the user of which user model is the same real user by calculating the similarity between multiple user model.
Specifically, carrying out user's identification according to the similarity between multiple user models may include steps of:
For each user model, the similarity between the user model and other users model is calculated；
Other users model using similarity less than the first similarity threshold is as the similar users mould of the user modelType；
Calculate the user of user's average similarity of the user model and the similar users model of the user modelAverage similarity；
According to user's average similarity of the user model and user's average similarity of the similar users model,Carry out user's identification.
Further, for each user model, similarity between the user model and other users model according toAs under type calculates:
Wherein, K[N, M]Represent the similarity between user model userN and user model userM；SuserNRepresent user's mouldThe target weight of type userN, SuserMRepresent the target weight of user model userM.
Further, between the target weight of each user model and the preset target weight referring to user modelDifference is less than or equal to weight threshold.
Above-mentioned various dimensions tag library includes label weight, and the sum of described label weight is 100.According to going through for each classificationThe matched label of history data institute, weight summation to the matched label of institute, so that it is determined that corresponding to the historical data of the categoryUser model initial weight (the sum of i.e. above-mentioned weight).Judge each user model initial weight whether with preset ginsengWhether it is less than or equal to weight threshold according to the difference between the target weight of user model, if the initial power of all user modelsDifference between value and the preset target weight referring to user model, then will be described first both less than or equal to the weight thresholdTarget weight of the beginning weight as user model.Otherwise, the label weight is adjusted, until the target power of all user modelsValue and the difference of the preset target weight referring to user model are both less than or are equal to the weight threshold.Due to the behavior of userIt can change over time, attributive character can also follow variation, and therefore, adjusting label weight when calculating user model target weight canWith optimization algorithm, to improve accuracy.
Wherein, user's average similarity of user model is by the similar users model of the user model and the user modelWhat the similarity of bracket determined, for example, it may be the average value of the similarity, is also possible to the variance etc. of above-mentioned similarity.In the present embodiment, user's average similarity of user model is the similar users model branch of the user model Yu the user modelThe average value of the similarity of frame.The method for calculating the average similarity of the similar users model of the user model is same as above.
According to user's average similarity of the user model and user's average similarity of the similar users modelCarry out user's identification the step of may include:
Calculate the user model user's average similarity and the similar users model user's average similarity itBetween difference；
If the difference is less than user's average similarity threshold value, the user of the user model and the similar users mouldThe user of type is same real user.
As an example, user model useraSimilar users model include userb、userc, user model usercPhaseIt include user like user modeld、userf.Then user model useraUser's average similarityUser model useraSimilar users model usercUser's average similarityCalculate S(usera) and S (userc) difference, if the difference be less than or equal to user's average similarity threshold value, user model useraWithUser model usercUser be same real user.
User identification method provided in an embodiment of the present invention, because being classified using the historical data to virtual account,Obtain the historical data of multiple classifications；User model is constructed according to the historical data of each classification, to obtain and the virtual accountNumber corresponding multiple user models；According to the similarity between the multiple user model, the technological means of user's identification is carried out,So overcoming the technical problem of user model inaccuracy in the prior art, and then the superposition behavior of the multiple users of reduction is reachedInterference to user model improves the technical effect of the accuracy of user model and the accuracy of user's identification.
Fig. 3 and Fig. 4 are please referred to, for a virtual account, the method for the embodiment of the present invention constructs multiple user's moulds firstThen type judges whether the user of above-mentioned multiple user models is same true by the similarity between multiple user modelsUser constructs deviation, accuracy caused by a user draws a portrait by the behavior of multi-user's superposition in the prior art to solveLow technical problem.
Illustrate user identification method provided by the embodiment of the present invention below with reference to specific example.
For the same virtual account, the corresponding multiple user models of the virtual account are constructed according to above-mentioned steps:User1, user2, user3, user4 and user5.Its target weight is respectively as follows: Suser1=80, Suser2=82, Suser3=96,Suser4=86, Suser5=77.First similarity threshold is 5%, and the second similarity threshold is 1%.In the present invention, the first phaseCan there are demand or experience setting according to project like degree threshold value and the second similarity threshold, the present invention is herein with no restrictions.
It is as shown in table 1 below according to the similarity between above formula (1) each user model:
Table 1:
Then according to upper table it is found that user model user1 and user model user2 is similar users model, user modelUser1 and user model user5 is similar users model, and user model user2 and user model user4 are similar users mouldType.
Then, user's average similarity between above-mentioned similar users model is calculated.
User's average similarity of user model user1 isThe user of user model user2 is flatEqual similarity isUser's average similarity of user model user4 is 4.6%；User modelUser's average similarity of user5 is 3.8%.
Difference between user's average similarity of user model user1 and user's average similarity of user model user2Value is 0.533%, then the user of the user of user model user1 and user model user2 are the same real users；User's mouldDifference between user's average similarity of type user1 and user's average similarity of user model user4 is 1.475%, thenThe user of user model user1 and the user of user model user4 are not the same real users；The use of user model user1Difference between family average similarity and user's average similarity of user model user5 is 0.675%, then user modelThe user of user1 and the user of user model user5 are same each real user.In conclusion user model user1, userThe user of model user2 and user model user5 are the same real users, and the user of user model user3 is that another is trueReal user, the user of user model user4 are another real users.
User identification method provided in an embodiment of the present invention, because being classified using the historical data to virtual account,Obtain the historical data of multiple classifications；User model is constructed according to the historical data of each classification, to obtain and the virtual accountNumber corresponding multiple user models；According to the similarity between the multiple user model, the technological means of user's identification is carried out,So overcoming the technical problem of user model inaccuracy in the prior art, and then the superposition behavior of the multiple users of reduction is reachedTo the technical effect of interference, the accuracy of the accuracy and user's identification that improve user model of user model.Further,Strong data base can be provided for the fast accurate advertisement dispensing of operator/enterprise, personalized recommendation, steal-number risk monitoring and controlPlinth.
Fig. 5 is the schematic diagram of the main modular of user model construction device according to an embodiment of the present invention.As shown in figure 5,The device includes: data categorization module 501, is classified to the historical data of virtual account, and the history number of multiple classifications is obtainedAccording to；Model construction module 502 constructs user model according to the historical data of each classification, to obtain and the virtual account pairThe multiple user models answered；Subscriber identification module 503 carries out user's knowledge according to the similarity between the multiple user modelNot.
Optionally, the data categorization module 501 is also used to: being carried out according to historical data of the time cycle to virtual accountClassification is a classification with the historical data in a period of time；Or according to time cycle and unique identification to virtual accountHistorical data is classified, and is a classification with the historical data of unique identification identical in a period of time.
Optionally, subscriber identification module is also used to: being directed to each user model, is calculated the user model and other usersSimilarity between model；Other users model using similarity less than the first similarity threshold is as the phase of the user modelLike user model；Calculate the use of user's average similarity of the user model and the similar users model of the user modelFamily average similarity；It is averaged phase according to the user of user's average similarity of the user model and the similar users modelLike degree, user's identification is carried out.
Optionally, for each user model, the similarity between the user model and other users model is according to such asUnder type calculates:
Wherein, K[N, M]Represent the similarity between user model userN and user model userM, SuserNRepresent user's mouldThe target weight of type userN, SuserMRepresent the target weight of user model userM.
Optionally, the difference between the target weight of each user model and the preset target weight referring to user modelLess than or equal to weight threshold.
Method provided by the embodiment of the present invention can be performed in above-mentioned apparatus, has the corresponding functional module of execution method and hasBeneficial effect.The not technical detail of detailed description in the present embodiment, reference can be made to method provided by the embodiment of the present invention.
Fig. 6 is shown can be using the user identification method of the embodiment of the present invention or the exemplary system of customer identification deviceFramework 600.
As shown in fig. 6, system architecture 600 may include terminal device 601,602,603, network 604 and server 605.Network 604 between terminal device 601,602,603 and server 605 to provide the medium of communication link.Network 604 can be withIncluding various connection types, such as wired, wireless communication link or fiber optic cables etc..
User can be used terminal device 601,602,603 and be interacted by network 604 with server 605, to receive or send outSend message etc..Various telecommunication customer end applications, such as the application of shopping class, net can be installed on terminal device 601,602,603The application of page browsing device, searching class application, instant messaging tools, mailbox client, social platform software etc..
Terminal device 601,602,603 can be the various electronic equipments with display screen and supported web page browsing, packetInclude but be not limited to smart phone, tablet computer, pocket computer on knee and desktop computer etc..
Server 605 can be to provide the server of various services, such as utilize terminal device 601,602,603 to userThe shopping class website browsed provides the back-stage management server supported.Back-stage management server can believe the product receivedThe data such as breath inquiry request carry out the processing such as analyzing, and processing result (such as target push information, product information) is fed back toTerminal device.
It should be noted that user identification method provided by the embodiment of the present invention is generally executed by server 605, accordinglyGround, customer identification device are generally positioned in server 605.
It should be understood that the number of terminal device, network and server in Fig. 6 is only schematical.According to realization needIt wants, can have any number of terminal device, network and server.
Below with reference to Fig. 7, it illustrates the computer systems 700 for the terminal device for being suitable for being used to realize the embodiment of the present inventionStructural schematic diagram.Terminal device shown in Fig. 7 is only an example, function to the embodiment of the present invention and should not use modelShroud carrys out any restrictions.
As shown in fig. 7, computer system 700 includes central processing unit (CPU) 701, it can be read-only according to being stored inProgram in memory (ROM) 702 or be loaded into the program in random access storage device (RAM) 703 from storage section 708 andExecute various movements appropriate and processing.In RAM 703, also it is stored with system 700 and operates required various programs and data.CPU 701, ROM 702 and RAM 703 are connected with each other by bus 704.Input/output (I/O) interface 705 is also connected to alwaysLine 704.
I/O interface 705 is connected to lower component: the importation 706 including keyboard, mouse etc.；It is penetrated including such as cathodeThe output par, c 707 of spool (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage section 708 including hard disk etc.；And the communications portion 709 of the network interface card including LAN card, modem etc..Communications portion 709 via such as becauseThe network of spy's net executes communication process.Driver 710 is also connected to I/O interface 705 as needed.Detachable media 711, such asDisk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 710, in order to read from thereonComputer program be mounted into storage section 708 as needed.
Particularly, disclosed embodiment, the process described above with reference to flow chart may be implemented as counting according to the present inventionCalculation machine software program.For example, embodiment disclosed by the invention includes a kind of computer program product comprising be carried on computerComputer program on readable medium, the computer program include the program code for method shown in execution flow chart.?In such embodiment, which can be downloaded and installed from network by communications portion 709, and/or from canMedium 711 is dismantled to be mounted.When the computer program is executed by central processing unit (CPU) 701, system of the invention is executedThe above-mentioned function of middle restriction.
It should be noted that computer-readable medium shown in the present invention can be computer-readable signal media or meterCalculation machine readable storage medium storing program for executing either the two any combination.Computer readable storage medium for example can be --- but notBe limited to --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor system, device or device, or any above combination.MeterThe more specific example of calculation machine readable storage medium storing program for executing can include but is not limited to: have the electrical connection, just of one or more conducting wiresTaking formula computer disk, hard disk, random access storage device (RAM), read-only memory (ROM), erasable type may be programmed read-only storageDevice (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device,Or above-mentioned any appropriate combination.In the present invention, computer readable storage medium can be it is any include or storage journeyThe tangible medium of sequence, the program can be commanded execution system, device or device use or in connection.And at thisIn invention, computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal,Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including but unlimitedIn electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer canAny computer-readable medium other than storage medium is read, which can send, propagates or transmit and be used forBy the use of instruction execution system, device or device or program in connection.Include on computer-readable mediumProgram code can transmit with any suitable medium, including but not limited to: wireless, electric wire, optical cable, RF etc. are above-mentionedAny appropriate combination.
Flow chart and block diagram in attached drawing are illustrated according to the system of various embodiments of the invention, method and computer journeyThe architecture, function and operation in the cards of sequence product.In this regard, each box in flowchart or block diagram can generationA part of one module, program segment or code of table, a part of above-mentioned module, program segment or code include one or moreExecutable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in boxThe function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practicalOn can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it wantsIt is noted that the combination of each box in block diagram or flow chart and the box in block diagram or flow chart, can use and execute ruleThe dedicated hardware based systems of fixed functions or operations is realized, or can use the group of specialized hardware and computer instructionIt closes to realize.
Being described in module involved in the embodiment of the present invention can be realized by way of software, can also be by hardThe mode of part is realized.Described module also can be set in the processor, for example, can be described as: a kind of processor packetIt includes sending module, obtain module, determining module and first processing module.Wherein, the title of these modules is under certain conditions simultaneouslyThe restriction to the module itself is not constituted, for example, sending module is also described as " sending picture to the server-side connectedThe module of acquisition request ".
As on the other hand, the present invention also provides a kind of computer-readable medium, which be can beIncluded in equipment described in above-described embodiment；It is also possible to individualism, and without in the supplying equipment.Above-mentioned calculatingMachine readable medium carries one or more program, when said one or multiple programs are executed by the equipment, makesObtaining the equipment includes:
Classify to the historical data of virtual account, obtains the historical data of multiple classifications；
User model is constructed according to the historical data of each classification, to obtain multiple users corresponding with the virtual accountModel；
According to the similarity between the multiple user model, user's identification is carried out.
Technical solution according to an embodiment of the present invention is obtained because being classified using the historical data to virtual accountThe historical data of multiple classifications；User model is constructed according to the historical data of each classification, to obtain and the virtual account pairThe multiple user models answered；According to the similarity between the multiple user model, the technological means of user's identification is carried out, soOvercome the technical problem of user model inaccuracy in the prior art, so reached the superposition behavior of the multiple users of reduction toThe interference of family model improves the technical effect of the accuracy of user model and the accuracy of user's identification.
Above-mentioned specific embodiment, does not constitute a limitation on the scope of protection of the present invention.Those skilled in the art should be brightIt is white, design requirement and other factors are depended on, various modifications, combination, sub-portfolio and substitution can occur.It is anyMade modifications, equivalent substitutions and improvements etc. within the spirit and principles in the present invention, should be included in the scope of the present inventionWithin.