Invention content
The embodiment of the present invention provides recognition methods and the device of a kind of advertisement cheating, with solve one in the prior art orMultiple technical problems.
In a first aspect, an embodiment of the present invention provides a kind of recognition methods of advertisement cheating, including:
Sample client is obtained from the charging failure daily record of advertisement;
According to sample Customer Acquisition cheating feature and cheating material characteristics;
Using cheating feature construction cheating model;
According to cheating model and cheating material characteristics, cheating client is identified.
With reference to first aspect, the embodiment of the present invention is lost in the first realization method of first aspect from the charging of advertisementAcquisition sample client in daily record is lost, including:
It obtains to surpass in the charging failure daily record of advertisement and throws ranking of sum of money in the sample client of top n, N is positive integer, and N is pre-If value, the super amount of money of throwing is the amount of money more than advertising budget.
With reference to first aspect, the embodiment of the present invention is in second of realization method of first aspect, according to sample visitorFamily obtains cheating feature and cheating material characteristics, including:
It is launched in behavioral data from the corresponding advertisement of the sample client, extracts abnormal budget, abnormal bid, abnormal consumptionAt least one of speed and abnormal dispensing frequency, as cheating feature;
It has launched in advertisement, has been extracted in similar title, picture, sound, video, text from the sample client is correspondingAt least one, as cheating material characteristics.
With reference to first aspect, the embodiment of the present invention is in the third realization method of first aspect, according to cheating model andCheating material characteristics identify cheating client, including:
The historical operation record of client is identified using cheating model, obtains risk client;
Whether the detection risk client material used when creating advertisement meets cheating material characteristics;
If met, risk client is determined as the client that practises fraud.
With reference to first aspect or its any one realization method, the embodiment of the present invention is in the 4th kind of realization side of first aspectIn formula, this method further includes:
Cheating label is added for cheating client;
Behavior is launched according to the Instant Ads of the cheating label limitation cheating client.
Second aspect, an embodiment of the present invention provides a kind of identification devices of advertisement cheating, including:
First acquisition module, for obtaining sample client from the charging failure daily record of advertisement;
Second acquisition module, for according to the sample Customer Acquisition cheating feature and cheating material characteristics;
Model module is built, for using the cheating feature construction cheating model;
Identification module, for according to cheating model and cheating material characteristics, identifying cheating client.
In conjunction with second aspect, the embodiment of the present invention is in the first realization method of second aspect, the first acquisition module packetIt includes:First acquisition submodule surpasses in the charging failure daily record for obtaining advertisement and throws ranking of sum of money in the sample client of top n, NFor positive integer, N is preset value, and the super amount of money of throwing is the amount of money more than advertising budget.
In conjunction with second aspect, for the embodiment of the present invention in second of realization method of second aspect, described second obtains mouldBlock includes:
First extracting sub-module, for being launched in behavioral data from the corresponding advertisement of the sample client, extraction is abnormal pre-At least one of calculation, abnormal bid, abnormal consumption rate and abnormal dispensing frequency, as cheating feature;
Second extracting sub-module extracts similar title, figure for having been launched in advertisement from the sample client is correspondingAt least one of piece, sound, video, text, as cheating material characteristics.
In conjunction with second aspect, the embodiment of the present invention is in the third realization method of second aspect, the identification module packetIt includes:
It identifies submodule, for the historical operation record of client to be identified using cheating model, obtains risk client;
Check submodule, it is special whether the material used when creating advertisement for detection risk client meets cheating materialSign;
Decision sub-module is used for, if meeting cheating material characteristics, risk client is determined as the client that practises fraud.
In conjunction with second aspect or its any one realization method, the embodiment of the present invention is in the 4th kind of realization side of second aspectIn formula, which further includes:
Mark module, for adding cheating label for cheating client;
Module is limited, for launching behavior according to the Instant Ads of the cheating label limitation cheating client.
The third aspect, an embodiment of the present invention provides a kind of identification device of advertisement cheating, the function of described device can be withBy hardware realization, corresponding software realization can also be executed by hardware.The hardware or software include it is one or more withThe above-mentioned corresponding module of function.
In a possible design, the structure of the identification device of advertisement cheating includes processor and memory, describedMemory is used to store the program for the recognition methods for supporting the identification device of advertisement cheating to execute above-mentioned advertisement cheating, the processingDevice is configurable for executing the program stored in the memory.The identification device of the advertisement cheating can also include communicationInterface, identification device and other equipment or communication for advertisement cheating.
Fourth aspect, an embodiment of the present invention provides a kind of computer readable storage mediums, for storing advertisement cheatingComputer software instructions used in identification device comprising for executing the journey involved by the recognition methods of above-mentioned advertisement cheatingSequence.
A technical solution in above-mentioned technical proposal has the following advantages that or advantageous effect:The training sample of cheating modelFrom the cheating feature for actually resulting in charging failure, sufficient data verification is have passed through, identification accuracy is pushed away better than theoryDisconnected and case analysis;By practising fraud, material characteristics assist in identifying, and further increase identification accuracy, therefore identify cheating client'sAccuracy is high.
Another technical solution in above-mentioned technical proposal has the following advantages that or advantageous effect:Client has cheating specialCheating is stamped when sign and marks this high-risk label so that retrieval end can be suppressed in real time, no longer need to take time to estimate, therefore timelinessProperty it is high.Furthermore, it is possible to regularly update charging failure daily record, constantly inputted using new data, it is automatic to improve and correct cheating mouldType improves the rule of cheating model, corrects erroneous judgement and omits.
Above-mentioned general introduction is merely to illustrate that the purpose of book, it is not intended to be limited in any way.Except foregoing descriptionSchematical aspect, except embodiment and feature, by reference to attached drawing and the following detailed description, the present invention is furtherAspect, embodiment and feature, which will be, to be readily apparent that.
Specific implementation mode
Hereinafter, certain exemplary embodiments are simply just described.As one skilled in the art will recognize thatLike that, without departing from the spirit or scope of the present invention, described embodiment can be changed by various different modes.Therefore, attached drawing and description are considered essentially illustrative rather than restrictive.
Fig. 1 shows the flow chart of the recognition methods of advertisement cheating according to an embodiment of the invention.
As shown in Figure 1, the recognition methods of advertisement cheating may comprise steps of:
Step S110, sample client is obtained from the charging failure daily record of advertisement.
Step S120, according to sample Customer Acquisition cheating feature and cheating material characteristics.
Step S130, using cheating feature construction cheating model.
Step S140, according to cheating model and cheating material characteristics, cheating client is identified.
In advertisement release process, if charging failure, charging failure daily record can be generated.It can be in charging failure daily recordIncluding but not limited to the client of charging failure account (such as:User name, social network sites account, cell-phone number, affiliated company etc.),Failure cause (such as:Budget is too low, the super throwing amount of money is excessive, monovalent abnormal promotion, launches excessive velocities, frequent starting, pause throwingPut), Time To Failure etc..Can within a period of time such as in one month, in 1 year charging failure daily record in, carryTake the account of several clients as sample client.It is then possible to which behavioural characteristic is launched in the advertisement for searching these accounts.Such as from thisRelevant dispensing behavioural characteristic is searched in the failure cause of the charging failure daily record of account.For another example, from the operation behavior of the accountRelevant dispensing behavioural characteristic is searched in record.These are launched behavioural characteristic and belong to cheating feature.By these cheatingsFeature can train to obtain cheating model as training sample.Then, identify that client's is real-time or history using cheating modelAdvertisement launches whether behavior meets cheating feature, to identify cheating client.
In one possible implementation, step S110 includes:
It obtains to surpass in the charging failure daily record of advertisement and throws ranking of sum of money in the sample client of top n, N is positive integer, and N is pre-If value, the super amount of money of throwing is the amount of money more than advertising budget.
For example, can be from big to small ranked up according to the super throwing amount of money to each client in charging failure daily record, then fromAccount of the head extraction ranking in top n.Such as N=20, the super ranking for throwing the amount of money is chosen in charging failure daily record preceding 20(TOP20) account is launched in the advertisement of client.
In one possible implementation, step S120 includes:
It is launched in behavioral data from the corresponding advertisement of sample client, extracts abnormal budget, abnormal bid, abnormal consumption rateAt least one of frequency is launched with abnormal, as cheating feature;
It has launched in advertisement, has been extracted in similar title, picture, sound, video, text extremely from sample client is correspondingFew one kind, as cheating material characteristics.
Wherein, advertisement dispensing behavioral data can include but is not limited to client in real time or the operation behavior of history records.FromThe budget of client can be found in operation behavior record, bid, consumption rate, launch the features such as frequency.It can be budget, go outThe relevant threshold values of setting such as valence, consumption rate, dispensing frequency.If some exceeds dependent thresholds, indicate that this feature is abnormal.It can be withOut it regard abnormal feature extraction as cheating feature.
For example, cheating feature can include but is not limited to:
1) unit price is substantially improved, for example makes 20 yuan into from 1 yuan, and bid promotion amplitude is more than threshold value, and the behavior is abnormal.
2) the frequent pause of plan starts, if frequency is greater than the set value, the behavior is abnormal.
3) period is promoted in modification plan.
In embodiments of the present invention, title, picture, sound, video, text etc. used when advertisement are created, belongs to wideThe material of announcement.Similar material characteristics can be extracted from the advertisement of dispensing of the account of multiple charging failures, these are similarMaterial characteristics as cheating material characteristics, to be subsequently compared, to screen.
For example, cheating material characteristics can include but is not limited to:
1) intention is similar:The intention height of the account of multiple charging failures is similar, for example, picture is the same, title is similar etc.;
2) it is similar to plan title, unit title:Such as the title formed with meaningless pure digi-tal.
In one possible implementation, as shown in Fig. 2, step S140 includes:
Step S141, the historical operation record of client is identified using cheating model, obtains risk client;
Step S142, whether the detection risk client material used when creating advertisement meets cheating material characteristics;
If step S143, met, risk client is determined as the client that practises fraud.
In addition it is also possible to identify a part of risk client by cheating model, by practising fraud, material recognition goes out a partRisk client, then the risk client that all includes by this two parts risk client are determined as the client that practises fraud.
In one possible implementation, as shown in figure 3, this method further includes:
Step S310, it is that cheating client adds cheating label;
Step S320, behavior is launched according to the Instant Ads of the cheating label limitation cheating client.
Referring to Fig. 4, cheating identification end can be that cheating client adds cheating label, deposit after identifying cheating clientEnter advertisement base.Practise fraud label realization method can there are many.Such as the account is indicated by field such as " ZB " of individually practising fraudFamily is cheating client.For another example, cheating client is marked by user name, the account etc. of modification cheating client.
Identification end can also will add the account of cheating label to be synchronized to retrieval end.End is retrieved to implement these cheating usersIt suppresses.For example, the budget minimum of limitation cheating client, bid, dispensing speed etc., or forbid client's dispensing advertisement of practising fraud.
The training sample of the cheating model of the embodiment of the present invention is from the cheating feature for actually resulting in charging failure, warpSufficient data verification is crossed, identification accuracy is better than theoretical implications and case analysis;By practising fraud, material characteristics assist in identifying,Identification accuracy is further increased, therefore identifies that the accuracy of cheating client is high.
In addition, client stamps cheating when having cheating feature mark this high-risk label so that retrieval end can be real-timeIt suppresses, no longer needs to take time to estimate, therefore timeliness is high.Furthermore, it is possible to regularly update charging failure daily record, constantly utilize newlyData input, automatic to improve and correct cheating model, improve the rule of cheating model, correct erroneous judgement and omit.
Client's such as advertiser's cheating is usually with a series of abnormal behaviour.In a kind of application example, it may be usedCheating pattern and cheating material characteristics are excavated in charging failure daily record offline, online real-time according to regular (such as cheating model)Identification cheating client adds output after marking for cheating client and is suppressed to retrieval end.The knowledge of the advertisement cheating of the embodiment of the present inventionOther method can be applied in system as shown in Figure 4.The system may include with lower module:
Offline behavior patterns mining module 1:The module exports cheating pattern and cheating material characteristics.Material of practising fraud is specialSign be referred to as high-risk material characteristics, indicate these materials belong to cheating risk it is higher.The module can obtain often offlineDay charging failure daily record (can also obtain charging failure daily record) according to other times.Being obtained from daily record has super throwing amount of money rowFor maximum top n client.The super investment bank for analyzing these clients is, puts into material, the account features etc. of advertisement, extraction operation sequenceRow pattern.At this point, if it find that new behavior pattern, can return to historical data and be verified, to ensure accuracy rateUnder the premise of, it recalls in advance.Further, it is also possible to synchronized update cheating material characteristics library.
Online real-time mode matching module 2:The cheating feature that the module will can offline be excavated, is abstracted as cheating mouldType.By subscribing to advertisement base real-time incremental, compares old value and new value, obtain managing operation history.The module can will be usedHigh-risk sensitive behavior is abstracted as behavior event in family historical operation record, using cheating model judge client operation behavior whetherFor cheating.In addition, when client creates advertisement material, which can also judge whether used material is high-risk objectMaterial can assist identifying cheating client in real time in this way.
Advertisement base 3:Online real-time mode matching module 2 can be that cheating client adds label, and is stored in advertisement base.AdvertisementLibrary can give label increment real-time synchronization to retrieval end.
It retrieval end 4 can be by search strategy to there is the client that cheating marks to implement to suppress.Such as does not go out or go out this lessThe advertisement of a little clients.Retrieval end can take corresponding means to limit the advertisement behavior that the client launches according to cheating label.It suppressesEffect can be embodied in charging failure daily record, periodically cheating model can be updated with new charging failure daily record, to optimize mouldType improves judgement accuracy.
The recognition methods practised fraud using the advertisement of the embodiment of the present invention, recognition accuracy can reach 95% or more, can drawReturn huge click loss and consumption loss.
Fig. 5 shows the block diagram of the identification device of advertisement cheating according to another embodiment of the present invention.As shown in figure 5, the dressIt sets and may include:
First acquisition module 510, for obtaining sample client from the charging failure daily record of advertisement;
Second acquisition module 520, for according to the sample Customer Acquisition cheating feature and cheating material characteristics;
Model module 530 is built, for using the cheating feature construction cheating model;
Identification module 540, for according to cheating model and cheating material characteristics, identifying cheating client.
In one possible implementation, the first acquisition module 510 includes:
First acquisition submodule surpasses the sample for throwing ranking of sum of money in top n in the charging failure daily record for obtaining advertisementClient, N are positive integer, and N is preset value, and the super amount of money of throwing is the amount of money more than advertising budget.
In one possible implementation, as shown in fig. 6, the second acquisition module 520 includes:
First extracting sub-module 521, for being launched in behavioral data from the corresponding advertisement of the sample client, extraction is abnormalAt least one of budget, abnormal bid, abnormal consumption rate and abnormal dispensing frequency, as cheating feature;
Second extracting sub-module 522 extracts similar mark for having been launched in advertisement from the sample client is correspondingAt least one of topic, picture, sound, video, text, as cheating material characteristics.
In one possible implementation, as shown in fig. 7, identification module 540 includes:
It identifies submodule 541, for the historical operation record of client to be identified using cheating model, obtains risk visitorFamily;
Check whether submodule 542, the material used when creating advertisement for detection risk client meet cheating materialFeature;
Decision sub-module 543, is used for, if meeting cheating material characteristics, risk client is determined as the client that practises fraud.
In one possible implementation, as shown in figure 8, the device further includes:
Mark module 810, for adding cheating label for cheating client;
Module 820 is limited, for launching behavior according to the Instant Ads of the cheating label limitation cheating client.
The function of each module in each device of the embodiment of the present invention may refer to the corresponding description in the above method, herein notIt repeats again.
Fig. 9 shows the block diagram of apparatus according to an embodiment of the present invention.As shown in figure 9, the device includes:Memory 910With processor 920, the computer program that can be run on processor 920 is stored in memory 910.The processor 920 is heldThe recognition methods of the advertisement cheating in above-described embodiment is realized when the row computer program.The memory 910 and processor920 quantity can be one or more.
The device further includes:
Communication interface 930 carries out data interaction for being communicated with external device.
Memory 910 may include high-speed RAM memory, it is also possible to further include nonvolatile memory (non-Volatile memory), a for example, at least magnetic disk storage.
If memory 910, processor 920 and the independent realization of communication interface 930, memory 910,920 and of processorCommunication interface 930 can be connected with each other by bus and complete mutual communication.The bus can be Industry Standard ArchitectureStructure (ISA, Industry Standard Architecture) bus, external equipment interconnection (PCI, PeripheralComponent) bus or extended industry-standard architecture (EISA, Extended Industry StandardComponent) bus etc..The bus can be divided into address bus, data/address bus, controlling bus etc..For ease of indicating, Fig. 9In only indicated with a thick line, it is not intended that an only bus or a type of bus.
Optionally, in specific implementation, if memory 910, processor 920 and communication interface 930 are integrated in one piece of coreOn piece, then memory 910, processor 920 and communication interface 930 can complete mutual communication by internal interface.
An embodiment of the present invention provides a kind of computer readable storage mediums, are stored with computer program, the program quiltProcessor realizes any method in above-described embodiment when executing.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically showThe description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or examplePoint is included at least one embodiment or example of the invention.Moreover, particular features, structures, materials, or characteristics describedIt may be combined in any suitable manner in any one or more of the embodiments or examples.In addition, without conflicting with each other, thisThe technical staff in field can be by the spy of different embodiments or examples described in this specification and different embodiments or examplesSign is combined.
In addition, term " first ", " second " are used for description purposes only, it is not understood to indicate or imply relative importanceOr implicitly indicate the quantity of indicated technical characteristic." first " is defined as a result, the feature of " second " can be expressed or hiddenInclude at least one this feature containing ground.In the description of the present invention, the meaning of " plurality " is two or more, unless otherwiseClear specific restriction.
Any process described otherwise above or method description are construed as in flow chart or herein, and expression includesIt is one or more for realizing specific logical function or process the step of executable instruction code module, segment or portionPoint, and the range of the preferred embodiment of the present invention includes other realization, wherein can not press shown or discuss suitableSequence, include according to involved function by it is basic simultaneously in the way of or in the opposite order, to execute function, this should be of the inventionEmbodiment person of ordinary skill in the field understood.
Expression or logic and/or step described otherwise above herein in flow charts, for example, being considered useIn the order list for the executable instruction for realizing logic function, may be embodied in any computer-readable medium, forInstruction execution system, device or equipment (system of such as computer based system including processor or other can be held from instructionThe instruction fetch of row system, device or equipment and the system executed instruction) it uses, or combine these instruction execution systems, device or setIt is standby and use.For the purpose of this specification, " computer-readable medium " can any can be included, store, communicating, propagating or passingDefeated program is for instruction execution system, device or equipment or the dress used in conjunction with these instruction execution systems, device or equipmentIt sets.The more specific example (non-exhaustive list) of computer-readable medium includes following:Electricity with one or more wiringInterconnecting piece (electronic device), portable computer diskette box (magnetic device), random access memory (RAM), read-only memory(ROM), erasable edit read-only storage (EPROM or flash memory), fiber device and portable read-only memory(CDROM).In addition, computer-readable medium can even is that the paper that can print described program on it or other suitable JieMatter, because can be for example by carrying out optical scanner to paper or other media, then into edlin, interpretation or when necessary with otherSuitable method is handled electronically to obtain described program, is then stored in computer storage.
It should be appreciated that each section of the present invention can be realized with hardware, software, firmware or combination thereof.Above-mentionedIn embodiment, software that multiple steps or method can in memory and by suitable instruction execution system be executed with storageOr firmware is realized.It, and in another embodiment, can be under well known in the art for example, if realized with hardwareAny one of row technology or their combination are realized:With the logic gates for realizing logic function to data-signalDiscrete logic, with suitable combinational logic gate circuit application-specific integrated circuit, programmable gate array (PGA), sceneProgrammable gate array (FPGA) etc..
Those skilled in the art are appreciated that realize all or part of step that above-described embodiment method carriesSuddenly it is that relevant hardware can be instructed to complete by program, the program can be stored in a kind of computer-readable storage mediumIn matter, which includes the steps that one or a combination set of embodiment of the method when being executed.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing module, it can alsoThat each unit physically exists alone, can also two or more units be integrated in a module.Above-mentioned integrated mouldThe form that hardware had both may be used in block is realized, can also be realized in the form of software function module.The integrated module is such asFruit is realized in the form of software function module and when sold or used as an independent product, can also be stored in a computerIn readable storage medium storing program for executing.The storage medium can be read-only memory, disk or CD etc..
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, anyThose familiar with the art in the technical scope disclosed by the present invention, can readily occur in its various change or replacement,These should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with the guarantor of the claimIt protects subject to range.