CN103577541B - The ranking fraud detection method and ranking fraud detection system of application program - Google Patents
The ranking fraud detection method and ranking fraud detection system of application program Download PDFInfo
- Publication number
- CN103577541B CN103577541B CN201310469920.3A CN201310469920A CN103577541B CN 103577541 B CN103577541 B CN 103577541B CN 201310469920 A CN201310469920 A CN 201310469920A CN 103577541 B CN103577541 B CN 103577541B
- Authority
- CN
- China
- Prior art keywords
- application program
- ranking
- history
- user
- evaluation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 127
- 238000011156 evaluation Methods 0.000 claims abstract description 85
- 238000000034 method Methods 0.000 claims abstract description 46
- 238000009826 distribution Methods 0.000 claims description 26
- 238000010200 validation analysis Methods 0.000 claims description 18
- 239000000284 extract Substances 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 description 8
- 238000004458 analytical method Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 7
- 230000008901 benefit Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000007689 inspection Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 230000002045 lasting effect Effects 0.000 description 2
- 241000209202 Bromus secalinus Species 0.000 description 1
- 206010036086 Polymenorrhoea Diseases 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000005314 correlation function Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 201000007094 prostatitis Diseases 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 230000036642 wellbeing Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/10—Protecting distributed programs or content, e.g. vending or licensing of copyrighted material ; Digital rights management [DRM]
- G06F21/12—Protecting executable software
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2221/00—Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/21—Indexing scheme relating to G06F21/00 and subgroups addressing additional information or applications relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/2127—Bluffing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2221/00—Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/21—Indexing scheme relating to G06F21/00 and subgroups addressing additional information or applications relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/2135—Metering
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Technology Law (AREA)
- Computer Hardware Design (AREA)
- Computer Security & Cryptography (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Storage Device Security (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a kind of ranking fraud detection method of application program and ranking fraud detection system.Methods described includes：Active period detecting step, the active period of the application program is detected based on history ranking information；Ranking fraud detection step, is detected to the active period based at least one evidence related to user's evaluation, obtains ranking fraud detection result.The method and system of the present invention can automatically identify the ranking fraud relevant with application program, so that application user obtains real application program ranking information.
Description
Technical field
The present invention relates to ranking fraud detection method and ranking the fraud inspection of network field, more particularly to a kind of application program
Examining system.
Background technology
User application, the mobile applications especially installed and run on mobile terminal are quickly grown in recent years.
User selects and installs application program for convenience, and many application program websites or application program shop can intensively provide application
Inquiry, download, evaluation of program etc. are serviced, while can also regularly, for example daily, release application program ranking list
（Application Leaderboard）To embody some current application programs popular with users.In fact, the ranking list is
One of most important means of application program are promoted, under application program ranking very high in ranking list would generally stimulate user a large amount of
The application program is carried, and huge economic well-being of workers and staff is brought for application developer.Therefore, application developer is highly desirable to
Its application program occupies higher ranking in ranking list.
The ranking fraud of application program（Ranking Fraud）Refer to that purpose is that improve application program arranges in application program
Ranking on row list and the deceptive practices carried out.In fact, improving application program row different from relying on traditional market means
Name, application developer has implemented the behavior of ranking fraud by exaggerating its product sales volume or the false product evaluation of issue
Through more and more universal, for example, employ " waterborne troops（human water armies）" lift the download of application program in a short time
Amount and evaluation number of times etc..
Industry, which has appreciated that, prevents ranking from cheating so that application user obtains real application program ranking information
Importance.In order to prevent the ranking of application program from cheating, existing method is risen according to application program ranking in one day
Degree directly locks whole application program to infer the presence of ranking fraud when judging and ranking fraud occur
Ranking, this mode is excessively simple and crude, it is difficult to accurately judges ranking fraud and has injured the row of normal application
Name rises.It can be seen that, this area is also very limited for the understanding and research of the ranking fraud detection problem of application program, so far also
In the absence of the correlation technique of the ranking fraud of effective detection application program.
The content of the invention
It is an object of the invention to provide the detection technique that a kind of ranking of application program is cheated, so as to automatically effectively know
Do not go out the ranking fraud relevant with application program, so that application user obtains real application program ranking information.
In order to solve the above technical problems, cheating inspection there is provided a kind of ranking of application program according to an aspect of the present invention
Survey method, methods described includes：
Active period detecting step, the active period of the application program is detected based on history ranking information；
Ranking fraud detection step, is tested the active period based at least one evidence related to user's evaluation
Card, obtains ranking fraud the result.
According to another aspect of the present invention, a kind of ranking fraud detection system of application program, the system are also provided
Including：
Active period detection unit, the active period for detecting the application program based on history ranking information；
Ranking fraud detection unit, for being carried out based at least one evidence related to user's evaluation to the active period
Checking, obtains ranking fraud the result.
According to another aspect of the present invention, a kind of ranking fraud detection method of application program, methods described are also provided
Including：
The active period of application program is verified based at least one evidence related to user's evaluation, ranking is obtained and takes advantage of
Cheat the result.
According to another aspect of the present invention, a kind of ranking fraud detection system of application program, the system are also provided
Including：
Ranking fraud detection unit, for being enlivened based at least one evidence related to user's evaluation to application program
Phase is verified, obtains ranking fraud the result.
The method and apparatus of the present invention automatically can effectively identify the ranking fraud relevant with application program, from
And application user is obtained real application program ranking information.
Brief description of the drawings
Fig. 1 is the flow chart of the active period detection method of application program in the specific embodiment of the invention；
Fig. 2 a are an examples of the Active event in application program ranking list；
Fig. 2 b are an examples of the active period in application program ranking list；
Fig. 3 is the system construction drawing of the ranking fraud detection system of application program in the specific embodiment of the invention；
Fig. 4 is the structural representation of the ranking fraud detection system of application program in another embodiment of the present invention.
Embodiment
With reference to the accompanying drawings and examples, the embodiment to the present invention is described in further detail.Implement below
Example is used to illustrate the present invention, but is not limited to the scope of the present invention.
The present invention is studied for the technical problem related to application program ranking, therefore those skilled in the art are to this
" application program " in invention should be interpreted broadly, and it includes what can be published on internet and be available for user to download, evaluate, performing
Various programs or file, the i.e. Mobile solution including running on the legacy application in PC, running on mobile terminal
Program, also including multimedia files such as the picture that can be downloaded and play, audio, video etc..
When detecting the ranking fraud of application program, there are several major issues for needing to solve.First, in application program
Can't always occur ranking fraud in whole life cycle, therefore the time that ranking is cheated is likely to occur firstly the need of detection；The
Two, because number of applications is huge, it is difficult to manually be demarcated each to there is the application program of ranking fraud, therefore need
A kind of technology of automatic detection ranking fraud is provided；3rd, uncertain it can be detected in the prior art and based on which kind of foundation
The presence of ranking fraud.
Ranking fraud of the specific embodiment of the present invention to application program carried out globality analysis and
Study there is provided the technology that a kind of ranking of detectable application program is cheated, it can be believed by the history ranking to application program
The analysis of breath detects " active period " of application program, for the evaluating characteristic of application program in active period, based on being commented with user
Valency related evidence carries out the detection of ranking fraud.
Found according to the analysis of inventor, the application program that there is ranking fraud can't be occupied very in billboard for a long time
High ranking, the higher situation of ranking is concentrated in one relatively short period only as some independent events, this
Show that ranking fraud exactly occurred within this period.In the present invention, application program can be continued ranking it is higher when
Phase is referred to as the " Active event of application program（Leading Event）", the frequent period for occurring Active event can be referred to as application
" the active period of program（Leading Session）”.Therefore, for ranking cheat detection firstly the need of detection each apply journey
Sequence there may exist Active event and the active period of ranking fraud.
Possess the history ranking information of application program at application program shop operator, at application program shop operator
Directly obtain, or ranked by the application program persistently issued within one section of longer period of history to application program shop operator
List information is analyzed and handled, it is also possible to obtain the history ranking information of application program.Because the history of application program is arranged
Name information describes the historical information and related user's evaluation information about application program ranking, therefore of the invention specific real
Apply in mode, the Active event of each application program and the detection of active period can be carried out based on the history ranking information, and
And then realize the detection cheated ranking.Behavior discovery is evaluated by analyzing the user of application program, compared to normal application
For program, there is the application program of ranking fraud can be rendered into different users evaluation spies in Active event and active period
Levy.Arranged therefore, it is possible to extract some be used for judgements related to user's evaluation from the history ranking information of application program
The evidence of name fraud, and these evidences are obtained, so as to realize the detection cheated ranking.
As shown in figure 1, providing a kind of ranking fraud detection of application program in the specific embodiment of the present invention
Method, methods described includes：
Active period detecting step S10, the active period of the application program is detected based on history ranking information；Ranking fraud inspection
Survey step S20, the evidence related to user's evaluations based at least one is detected to the active period, is obtained ranking and is cheated
Testing result.
Below, each step stream of above-mentioned ranking fraud detection method in the specific embodiment of the invention is illustrated with reference to accompanying drawing
Journey and function.
Because history ranking information is the data basis of the ranking fraud of detection application program in the present invention, therefore it is used as this
One preferred embodiment of invention, the ranking fraud detection method may also include a history ranking information obtaining step, obtain
History ranking information of the application program in application program ranking list.
After an application program is published, any user can be evaluated it.In fact, user evaluate for
Application program is one of most important feature for promoting.Application program with more high praise will attract more users to come
Buy or download it, and cause higher ranking of the application program in ranking list.Thus in history ranking information, it can wrap
Include the evaluation information that the user of application program in history evaluation information, i.e. history each period makes to the application program.
Application program ranking list can generally show the application program of K before welcome ranking, such as first 1000.And
And, application program ranking list would generally be regularly updated, for example, be updated daily.Therefore, for each application program a
There is its history ranking information, the history ranking information can include being expressed as a ranking sequence corresponding with discrete-time series
Ra={ r1 a,…,ri a,…,rn a, the interval between time point in the discrete-time series is fixed, i.e. application program ranking list
Update cycle.Wherein, ri aIt is application program a in time tiWhen ranking, ri a∈ { 1 ..., K ... ,+∞ } ,+∞ are represented should
With the row of program a K positions not before ranking list ranking；N represents the time point sum corresponding to all history ranking informations.For example,
In the case where ranking list updates daily, tiI-th day in this phase of history is meant that, n is exactly corresponding to history ranking information
Total number of days.As can be seen that ri aValue it is smaller, illustrate that i-th day ranking in ranking list of application program a is higher.
In the history ranking information obtaining step, the history ranking information can be obtained in many ways.For example, can be from
The history ranking information is directly obtained at application program shop operator, can also be from application program shop in one phase of longer history
The history ranking information etc. is extracted in the data persistently issued in period.
S10：Active period detecting step, the active period of the application program is detected based on history ranking information.
Active period represents that application program ranking in application program ranking list is higher, that is, user's attention rate is higher
One period, therefore these active periods are only appeared in the ranking fraud that application program market can be affected greatly
It is interior.So in the specific embodiment of the invention, the history ranking letter from application program is first had to for the detection that ranking is cheated
The active period of application program is detected in breath.
In a preferred embodiment of the invention, an Active event can be further comprised in the active period detecting step
Detecting step, the Active event of the application program is detected based on the history ranking information.
Due to application developer it is desirable to which its application program occupies higher ranking in ranking list, therefore apply journey
The means that sequence developer cheats possible with ranking make its application program rank among ranking list prostatitis.Found, applied by analysis
Program can't always occupy very high ranking in billboard, and it is " Active event " to occur lasting ranking higher period,
Show that transverse axis represents the corresponding time series of history ranking information in the example of the Active event of application program, figure in Fig. 2 a
（Date Index）, the longitudinal axis represents the ranking of application program（Ranking）, the event 1 in figure（Event1）With event 2
（Event2）Represent two Active events appeared in the application program placement history, its profile is respectively during the Active event
Ranking point be formed by connecting.
In the specific embodiment of the invention, application program higher standard of ranking in application program ranking list is that this should
It is not more than a rank threshold K* with the ranking of program.Due to application program ranking before ranking list the row of K* positions be considered as row
Name is higher, thus the period of the lasting row in preceding K* positions of ranking of application program can be considered as an Active event, should
Active event should be initially entered before ranking list since the application program the row of K* positions, continueed to that the application program is fallen and ranking list
The row of preceding K* positions terminate.
Preferably, the step of method in embodiment of the present invention may also include setting rank threshold K*, so that really
Determine application program higher standard of ranking in application program ranking list.Because the application program total quantity K in ranking list is usual
It is very big, for example, 1000 etc., therefore above-mentioned rank threshold K* is typically smaller than K values.According to application program in application program ranking list
Total quantity K and those skilled in the art the factor such as analysis demand, rank threshold K* can be whole between such as 1~500
Several values.It will be understood by those skilled in the art that K* value is smaller, application program is considered as the higher standard of ranking and got over
It is high.In fig. 2 a, the value of the K* is 300.
According to the above-mentioned character express for Active event, application program a Active event e can be with equation below table
State：
A given rank threshold K* is used as the higher standard of ranking, wherein K* ∈ [1, K]；Application program a Active event e
Include the time range of time from the beginning a to end timeCorresponding application program a ranking expires
FootAndAndIt is satisfied by rk a≤K*。
Detection for Active event can be seen that according to above-mentioned statement it is only important that the ranking of detection application program is held
Between continuing at the beginning of a period of time of the row of preceding K* positions and the end time, and by between a pair of time starteds and end time
Period is defined as Active event.Therefore, in the specific embodiment of the invention, the Active event detecting step can further comprise
Following steps：
Time started identification step S101：In this step, identifying Active event since history ranking information
Time.Specifically, in the time started identification step, application that can be in sequential search history ranking information on each time point
Program ranking, is more than rank threshold K* when the ranking of current point in time was not more than rank threshold K* and the ranking at a upper time point
When, identification current point in time be Active event at the beginning of between.It will be understood by those skilled in the art that due to being arranged in application program
Multiple Active events are potentially included in name history, therefore multiple time starteds can be can recognize that in the time started identification step
Point.
End time identification step S102：In this step, the end of active time is identified from history ranking information
Time.Specifically, in the end time identification step, application that can be in sequential search history ranking information on each time point
Program ranking, when the ranking of current point in time was more than the rank threshold K* and ranking at a upper time point is not more than rank threshold K*
When, recognized the end time that a upper time point is Active event.It will be understood by those skilled in the art that due to being arranged in application program
Multiple Active events are potentially included in name history, therefore multiple end times can be can recognize that in the end time identification step
Point.
Active event identification step S103：In this step by adjacent end time after each time started and its it
Between period be identified as Active event, thus detected all Active events of the application program in placement history.
What deserves to be explained is, as a kind of special circumstances, if in first of the period of history analyzed and handled
Between on point, such as first day in historical record, the ranking of application program just before ranking list K* positions row, now described
In time started identification step S101, first time point is defined as a time started.Similarly, if analyzed
On last time point of the period of history of processing, such as today, the ranking of the application program K* positions still before ranking list
Row, are now defined as an end time in the end time identification step S102 by last time point.
The mode of Active event in detection application program is described above, it is on this basis, preferred real in the present invention one
Apply in mode, adjoining Active event can be merged in the active period detecting step to constitute the active period.
By further study show that, some application programs can continuously occur repeatedly adjacent to each other near within one period
Active event, this period is exactly " active period " of application program in the present invention.It can be seen that, adjoining Active event is merged
Just to constitute active period.Specifically, the time interval of two neighboring Active event can be less than to an interval threshold φ as general
Two Active events merge the standard in same active period, and the time interval of two neighboring Active event then refers to adjacent two
In individual Active event at the beginning of the end time of previous Active event and latter Active event between interval.
Preferably, the step of method in embodiment of the present invention may also include setting interval threshold φ, so that really
It is fixed that two Active events are merged into the standard in same active period.The factors such as the analysis demand according to those skilled in the art,
Interval threshold φ value can be the integer value in 2~10 times of the update cycle of application program ranking list.This area skill
Art personnel are appreciated that interval threshold φ value is smaller, by standard of two Active events merging in same active period just
It is higher.
Show that transverse axis represents the history ranking information corresponding time in the example of the active period of application program, figure in Fig. 2 b
Sequence（Date Index）, the longitudinal axis represents the ranking of application program（Ranking）, in figure during 1（Session1）With period 2
（Session2）Two active periods appeared in the application program placement history are represented, each active period is by multiple Active events
Constitute.
According to the above-mentioned character express for active period, application program a active period s can be stated with equation belowization：
Application program a active period s includes a time rangeAdjacent Active event { the e with n1,…,
en, it meetsAnd cause in the absence of other active period s*In addition,HaveWherein φ is default Active event interval threshold, be used to judge between Active event neighboring extent with
It is incorporated into the criterion of same active period.
Detection for active period can be seen that according to above-mentioned statement it is only important that journey will be applied based on interval threshold φ
Adjoining Active event merges to form active period in sequence placement history.Specifically, in the work of the specific embodiment of the invention
In jump phase detecting step, the Active event that sequential search is each detected since the initial time point in history ranking information,
When current active event and the time interval of a upper Active event are less than interval threshold φ, the two Active events are merged
In same active period, until having searched for all Active events detected to detect the application program in placement history
All active periods.
What deserves to be explained is, as a kind of special circumstances, if an Active event not with any other Active event
Adjoining, the Active event itself is also contemplated as constituting an active period.In this case, in the active period detecting step
In, when the time interval of an Active event and a upper Active event is not less than the interval threshold φ, and the Active event is with
When the time interval of one Active event is not less than the interval threshold φ, detect the Active event from as an active period.
As previously mentioned, detected above-mentioned active period represents application program ranking in application program ranking list
It is higher, that is, one period welcome by user, the detected active period, which can be used as, to be included detecting that ranking fraud exists
The data basis of interior various application program services.Therefore, after the active period of application program is detected, it is used as the present invention one
The active period information of detected application program, can also be sent to application developer, answer by individual preferred embodiment
With program shop operator or the terminal user of application program.
For application developer, it can become according to the development of the active period information analysis correlative technology field
The demand of gesture or application user, so as to instruct the exploitation and operation of application program；For application program shop operator
Speech, it can further analyze the ranking that false high ranking in ranking list is obtained using fraudulent mean according to the active period information
Fraud etc., so as to improve the operation of application program shop；And for application terminal user, they can basis
The active period information voluntarily judges that application program there is a possibility that ranking fraud or selection meet the application of self-demand
Program etc..
In addition, being used as the Active event and a kind of specific implementation of active period of detection application program, following algorithm 1
Show the example of a program code of detection active period in given application program a history ranking information.
In above-mentioned algorithm 1, each Active event e is defined asActive period s is defined as
Wherein EsIt is the set of the Active event in active period s.Especially, first between at the beginning of history ranking information extract should
With program a each Active event e（Step 2-5 in algorithm 1）.For each Active event e extracted, detection e is with before
Time interval between one Active event e* is to judge whether they belong to same active period.Specifically, ifActive event e is then considered to belong to a new active period（Step 7-13 in algorithm 1）.So, it is above-mentioned
Algorithm 1 can recognize Active event and active period by the single pass of the history ranking information to application program a.
Ranking fraud detection step S20, the evidence related to user's evaluation based at least one enters to the active period
Row detection, obtains ranking fraud detection result.
As introduction above to history ranking information, it includes should in history evaluation information, i.e. history each period
The user made with the user of program to the application program evaluates.Meanwhile, active period is that application program is likely to occur ranking and taken advantage of
The period of swindleness.Therefore, the evaluating characteristic of history ranking information in application program active period can be analyzed, extract some with
User evaluates related information as the evidence for detecting ranking fraud.
Specifically, after an application program is published, any download user can be evaluated it, such as to this
Application program provides 1~5 point of scoring, and usual 5 points to represent user very satisfied to the application program（Highest is evaluated）, and 1 point
Represent very dissatisfied（It is minimum to evaluate）.In fact, user evaluate for application program popularization for be most important feature it
One.Application program with more high praise will attract more users to buy or download it, and cause the application program to exist
Higher ranking in ranking list.Therefore, false evaluation is also the important behaviour form in ranking fraud.If the work of application program
There is ranking fraud in jump phase s, the evaluation within active period s period will be different with the evaluation from other historical stages
Off-note, this feature can be used for building the evidence related to user's evaluation for being used for detecting ranking fraud.
As a preferred embodiment of the present invention, the ranking fraud detection step can further comprise a proof validation
Step, is verified to the active period based at least one evidence related to user's evaluation and obtains a fraud parameter.This
Sample, after the evidence relevant with user's evaluation is extracted, can calculate fraud parameter corresponding with the evidence, the fraud parameter sheet
Body can as the ranking fraud detection method in present embodiment ranking fraud detection result.Due to commenting for influence application program
The factor of valency feature is complex, and only relying on one or more evidences related to user's evaluation possibly can not accurately judge one
Application program is cheated with the presence or absence of ranking but only obtains a detected value for reference（Cheat parameter）, but art technology
Personnel can judge that application program there is a possibility that ranking fraud according to the fraud parameter completely.
For normal application program, the average user evaluation within the particular active phase should be commented with its all history
Average ratings in valency record are consistent.On the contrary, for exist ranking fraud application program, in its active period compared to
Its history evaluation can have surprising high praise.It is used as a preferred embodiment of the present invention, the card related to user's evaluation
According to can be based on the average user evaluation in active periodWith history average ratingsTo constitute, and based on the evidence meter constituted
An evidence value is calculated as the fraud parameter for judging ranking fraud.
For example intuitively, the average value that all users evaluate in active period can be calculatedWith history average ratingsBetween
Difference, or the average value that all users evaluateWith history average ratingsBetween ratio, be used as the fraud parameter.
For another example the average value that all users evaluate in active period can also be calculatedWith history average ratingsBetween
Difference and history average ratingsRatio, be used as the fraud parameter.By formulating description, fraud parameter, Δ RsIt is as follows：
WhereinIt is the average user evaluation of estimate in active period,It is application program a history evaluation average value.Therefore,
Compared to the active period of other applications in ranking list, if the active period s of an application program includes significantly greater Δ
RsValue, with regard to there is a strong possibility there is ranking fraud in property to the application program.
In the evaluation information of application program, each evaluate can be classified as a discrete opinion rating system | L |
In, such as including from 1~5 this five grades, which represent fancy grade of the user for the application program.It is normal for one
Application program a for, its opinion rating l in active period siDistribution p (li|Rs,a) should with its history evaluation record in
Distribution p (li|Ra) it is consistent.As a preferred embodiment of the present invention, the evidence related to user's evaluation can be based on
Opinion rating distribution in opinion rating distribution and history evaluation information of the application program in active period is constituted, and based on institute's structure
Into the evidence calculate an evidence value as judge ranking fraud fraud parameter.
For example, can calculate opinion rating of the application program in active period distribution and history evaluation information in opinion rating
Distribution between difference, be used as the fraud parameter.Specifically, it can pass through firstTo calculate p (li|
Rs,a) value, whereinBe active period inner evaluation grade be liUser evaluate number,It is total in active period s comment
Valence mumber mesh；Simultaneously similar mode can be used to calculate p (li|Ra)；Then the evaluation of application program in active period etc. is calculated
Difference in the distribution of level and history evaluation information between the distribution of opinion rating.As a kind of specific implementation, it can make
With p (li|Rs,a) and p (li|Ra) between COS distance D (s) estimate the difference between them., should by formulating description
Cheat parameter D (s) as follows：
It can be seen that, compared to the active period of other applications in ranking list, if the active period s of an application program is included
Significantly greater D (s) values, with regard to there is a strong possibility there is ranking fraud in property to the application program.
A variety of evidences related to user's evaluation are described above, except being used alone in above-mentioned each preferred embodiment
One of which is carried out outside ranking fraud detection, in the preferred embodiment of proof validation step, can be with
Consider multiple in the above-mentioned evidence related to user's evaluation, will to be obtained based on these proof validations correspondence fraud parameters
It is weighted, so as to obtain a final fraud parameter.It is possible in view of above-mentioned a variety of evidences with different dimensions, this
Art personnel can be returned according to the attention degree in actual analysis demand for each evidence based on commonly known in the art
One change method and Weight Determination determine the weighted value of each fraud parameter, will not be repeated here.
The proof validation step in ranking fraud detection step is described above, it can be commented based at least one with user
Valency it is related evidence is verified to the active period and obtains a fraud parameter, the fraud parameter can serve as ranking in itself
The ranking fraud detection result of fraud detection method.But in order that those skilled in the art more easily carry out ranking fraud
Detection, in a preferred embodiment, ranking fraud detection step can further include a fraud parameter judgment step,
Obtained fraud parameter will be calculated according to evidence to be compared with a threshold value, so as to intuitively judge to judge that application program is
It is no to there is ranking fraud.
It will be understood by those skilled in the art that based on a variety of evidences related to user's evaluation described in above, this
Art personnel can set corresponding threshold value respectively according to the heterogeneity and detection demand of evidence, according to set threshold
Value whether there is the judgement that ranking is cheated to carry out application program, and the final result that will determine that is used as specific embodiment party of the present invention
The ranking fraud detection result of ranking fraud detection method in formula.For example, being evaluated for a variety of and user described in above
For related evidence, when the fraud parameter calculated exceedes set threshold value, judge that the application program has ranking
Cheat phenomenon.
Obtained in ranking fraud detection step after ranking fraud detection result, in a preferred embodiment of the invention
In, the terminal that resulting ranking fraud detection result can also be sent to application program shop operator or application program is used
Family.For application program shop operator, it can improve application program shop according to the ranking fraud detection result
Operation；And for application terminal user, they can select to meet itself according to the ranking fraud detection result
Application program of demand etc..
As shown in figure 3, additionally providing a kind of ranking fraud detection system of application program in the specific embodiment of the invention
100, the system 100 includes：
Active period detection unit 110, the active period for detecting the application program based on history ranking information；Ranking is taken advantage of
Detection unit 120 is cheated, for being detected based at least one evidence related to user's evaluation to the active period, is arranged
Name fraud detection result.
Below, each unit function of said detecting system is illustrated with reference to accompanying drawing.
Because history ranking information is the data basis of the ranking fraud of detection application program in the present invention, therefore it is used as this
One preferred embodiment of invention, the ranking fraud detection system 100 may also include a history ranking information acquiring unit, use
In history ranking information of the acquisition application program in application program ranking list.
The history ranking information acquiring unit can obtain the history ranking information in many ways.For example, can be from application
The history ranking information is directly obtained at the operator of program shop, can also be from application program shop in one section of longer period of history
The history ranking information etc. is extracted in the data inside persistently issued.
Active period detection unit 110, the active period for detecting the application program based on history ranking information.
In a preferred embodiment of the invention, the active period detection unit 110 can further comprise an Active event
Detection module, the Active event for detecting the application program based on the history ranking information.
Preferably, the system in embodiment of the present invention may also include a rank threshold setting unit, for setting ranking
Threshold k * value, so that it is determined that application program higher standard of ranking in application program ranking list.Rank threshold K*'s takes
Value can be the integer between 1~500.
In the specific embodiment of the invention, the Active event detection module further comprises：
Time started identification module 111, between being identified from history ranking information at the beginning of Active event.Specifically
Ground, the application program ranking that the time started identification module can be in sequential search history ranking information on each time point, when working as
When the ranking at preceding time point was not more than rank threshold K* and the ranking at a upper time point and is more than rank threshold K*, current time is recognized
Point be Active event at the beginning of between.
End time identification module 112, the end time for identifying active time from history ranking information.Specifically
Ground, the application program ranking that the end time identification module can be in sequential search history ranking information on each time point, when working as
When the ranking at preceding time point was more than rank threshold K* and the ranking at a upper time point and is not more than rank threshold K*, the upper time was recognized
Point is the end time of Active event.
Active event identification module 113, for by between each time started and adjacent end time after it when
Between section be identified as Active event, thus detected all Active events of the application program in placement history.
What deserves to be explained is, as a kind of special circumstances, if in first of the period of history analyzed and handled
Between on point, such as first day in historical record, the ranking of application program just before ranking list K* positions row, now this start
First time point is defined as a time started by time identification module 111.Similarly, if analyzing and handling
On last time point of period of history, such as today, the row of ranking K* positions still before ranking list of application program now should
Last time point is defined as an end time by end time identification module 112.
In a preferred embodiment of the invention, the active period detection unit 110, which is used to merging, adjoining enlivens thing
Part is to constitute the active period of the application program.
Preferably, the ranking fraud detection system 100 in embodiment of the present invention may also include an interval threshold and set single
Member, the value for setting interval threshold φ, so that it is determined that two Active events are merged into the standard in same active period.Should
Interval threshold φ value can be the integer value in 2~10 times of the update cycle of application program ranking list.
In the specific embodiment of the invention, active period detection unit 110 is from the initial time point in history ranking information
Start the Active event that sequential search is each detected, be somebody's turn to do when the time interval of current active event and a upper Active event is less than
During interval threshold φ, by the two Active events merge in same active period, until searched for it is all detect enliven thing
Part is to detect all active periods of the application program in placement history.
What deserves to be explained is, as a kind of special circumstances, if an Active event not with any other Active event
Adjoining, the Active event itself is also contemplated as constituting an active period.In this case, the active period detection unit 110
For being not less than the interval threshold φ when the time interval of an Active event and a upper Active event, and the Active event is with
When the time interval of one Active event is not less than the interval threshold φ, detect the Active event from as an active period.
As a preferred embodiment of the invention, ranking fraud detection system 100 can also include an active period and send
Unit, application developer, application program shop operator are sent to by the active period information of detected application program
Or application user.
Ranking fraud detection unit 120, for the evidence related to user's evaluation based at least one come to described active
Phase is detected, obtains ranking fraud detection result.
As a preferred embodiment of the present invention, the ranking fraud detection unit 120 can further comprise an evidence
Authentication module, for being verified based at least one evidence related to user's evaluation to the active period and obtaining a fraud
Parameter.
In a preferred embodiment, the evidence related to user's evaluation can be based on the average user evaluation in active periodWith history average ratingsTo constitute, and an evidence value is calculated as judging that ranking is taken advantage of based on the evidence constituted
The fraud parameter of swindleness.In another preferred embodiment, the evidence related to user's evaluation can be based on application program active
The opinion rating distribution in opinion rating distribution and history evaluation information in phase is constituted, and is calculated based on the evidence constituted
Go out an evidence value as the fraud parameter for judging ranking fraud.
Except one of which is used alone in above-mentioned each preferred embodiment come in addition to carrying out ranking fraud detection,
Proof validation module can also consider multiple in the above-mentioned evidence related to user's evaluation, will be based on these proof validations
Obtained correspondence fraud parameter is weighted, so as to obtain a final fraud parameter.
In order that those skilled in the art more easily carry out ranking fraud detection, in a preferred embodiment,
Ranking fraud detection unit 120 can further include a fraud parameter judge module, will calculate what is obtained according to evidence
Fraud parameter is compared with a threshold value, so as to intuitively judge to judge that application program is cheated with the presence or absence of ranking.
Obtained in ranking fraud detection step after ranking fraud detection result, in a preferred embodiment of the invention
In, ranking fraud detection system 100 also includes a ranking fraud detection result transmitting element, by resulting ranking fraud detection
As a result it is sent to the terminal user of application program shop operator or application program.
It will be understood by those skilled in the art that in the case of known to the Active event and active period information of application program,
Those skilled in the art can directly implement above-mentioned ranking fraud detection step according to above-mentioned Active event and active period information,
So as to realize the detection of application program ranking fraud.Therefore, one is additionally provided in the another embodiment of the present invention
The ranking fraud detection method of application program is planted, methods described includes：Based at least one evidence related to user's evaluation come
Active period to application program is detected, obtains ranking fraud detection result.Arranged in the application program of the embodiment
In name fraud detection method, the technology contents implemented are identical with ranking fraud detection step in embodiment before, this
Place is repeated no more.
Accordingly, a kind of ranking fraud inspection of application program is additionally provided in another embodiment of the present invention simultaneously
Examining system, the system includes：Ranking fraud detection unit, for based at least one evidence related to user's evaluation to institute
State active period to be detected, obtain ranking fraud detection result.In the application program ranking fraud detection of the embodiment
In system, the technology contents implemented are identical with ranking fraud detection unit in embodiment before, and here is omitted.
Fig. 4 is a kind of structural representation of the ranking fraud detection system 400 of application program provided in an embodiment of the present invention,
The specific embodiment of the invention is not limited implementing for ranking fraud detection system 400.As shown in figure 4, the ranking is taken advantage of
Swindleness detecting system 400 can include：
Processor (processor) 410, communication interface (Communications Interface) 420, memory
(memory) 430 and communication bus 440.Wherein：
Processor 410, communication interface 420 and memory 430 complete mutual communication by communication bus 440.
Communication interface 420, communicates for the network element with such as client etc..
Processor 410, for configuration processor 432, can specifically realize ranking fraud detection in embodiment described in above-mentioned Fig. 3
The correlation function of system.
Specifically, program 432 can include program code, and described program code includes computer-managed instruction.
Processor 410 is probably a central processor CPU, or specific integrated circuit ASIC（Application
Specific Integrated Circuit）, or it is arranged to implement one or more integrated electricity of the embodiment of the present invention
Road.
Memory 430, for depositing program 432.Memory 430 may include high-speed RAM memory, it is also possible to also include
Nonvolatile memory（non-volatile memory）, for example, at least one magnetic disk storage.Program 432 can specifically be wrapped
Include：
Active period detection unit, the active period for detecting the application program based on history ranking information；
Ranking fraud detection unit, for being carried out based at least one evidence related to user's evaluation to the active period
Detection, obtains ranking fraud detection result.
Program 432 can also specifically include：
Ranking fraud detection unit, for being examined based at least one evidence related to user's evaluation to active period
Survey, obtain ranking fraud detection result.
Each unit implements the corresponding units that may refer in above embodiment in program 432, will not be described here.
It is apparent to those skilled in the art that, for convenience and simplicity of description, the equipment of foregoing description
With the specific work process of module, the correspondence description in aforementioned means embodiment is may be referred to, be will not be repeated here.
Those of ordinary skill in the art are it is to be appreciated that the list of each example described with reference to the embodiments described herein
Member and method and step, can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
Performed with hardware or software mode, depending on the application-specific and design constraint of technical scheme.Professional and technical personnel
Described function can be realized using distinct methods to each specific application, but this realization is it is not considered that exceed
The scope of the present invention.
If the function is realized using in the form of SFU software functional unit and is used as independent production marketing or in use, can be with
It is stored in a computer read/write memory medium.Understood based on such, technical scheme is substantially in other words
The part contributed to original technology or the part of the technical scheme can be embodied in the form of software product, the meter
Calculation machine software product is stored in a storage medium, including some instructions are to cause a computer equipment（Can be individual
People's computer, server, or network equipment etc.）Perform all or part of step of each embodiment methods described of the invention.
And foregoing storage medium includes：USB flash disk, mobile hard disk, read-only storage（ROM, Read-Only Memory）, arbitrary access deposits
Reservoir（RAM, Random Access Memory）, magnetic disc or CD etc. are various can be with the medium of store program codes.
Embodiment of above is merely to illustrate the present invention, and not limitation of the present invention, about the common of technical field
Technical staff, without departing from the spirit and scope of the present invention, can also make a variety of changes and modification, therefore all
Equivalent technical scheme falls within scope of the invention, and scope of patent protection of the invention should be defined by the claims.
Claims (35)
1. the ranking fraud detection method of a kind of application program, it is characterised in that methods described includes：
Active period detecting step, the active period of the application program is detected based on history ranking information；The work of the application program
The jump phase is the period for the Active event for frequently occurring application program；The Active event of the application program is that application program is persistently arranged
Name higher period；
Ranking fraud detection step, is verified to the active period based at least one evidence related to user's evaluation, obtained
The result is cheated to ranking；
The ranking fraud detection step further comprises：
Proof validation step, is verified to the active period based at least one evidence related to user's evaluation and obtains one
Cheat parameter；The evidence related to user's evaluation is based on the average ratings in the active periodWith history average ratingsConstitute.
2. according to the method described in claim 1, it is characterised in that
The fraud parameter is the average ratings in the active periodWith history average ratingsDifference or ratio.
3. according to the method described in claim 1, it is characterised in that
The fraud parameter is the average ratings in the active periodWith history average ratingsDifference averagely commented with history
ValencyRatio.
4. according to the method described in claim 1, it is characterised in that
The distribution of opinion rating of the evidence related to user's evaluation based on application program in the active period and history
The distribution of opinion rating is constituted in evaluation information.
5. method according to claim 4, it is characterised in that
Described cheat is evaluated in the distribution and history evaluation information that parameter is opinion rating of the application program in the active period
Difference between the distribution of grade.
6. method according to claim 5, it is characterised in that by calculating evaluation of the application program in the active period
COS distance in the distribution of grade and history evaluation information between the distribution of opinion rating calculates the difference between them.
7. described according to the method described in claim 1, it is characterised in that in the proof validation step, considering extremely
A few evidence related to user's evaluation, pair that will be obtained based at least one described proof validation related to user's evaluation
Parameter should be cheated to be weighted, so as to obtain the fraud parameter.
8. the method according to any one of claim 1-7, it is characterised in that the ranking fraud detection step is further
Including：
Parameter judgment step is cheated, the fraud parameter is compared with a threshold value, so as to whether judge the application program
There is ranking fraud.
9. according to the method described in claim 1, it is characterised in that methods described also includes：
History ranking information obtaining step, obtains history ranking letter of the application program in application program ranking list
Breath.
10. method according to claim 9, it is characterised in that in the history ranking information obtaining step, from application
Program shop operator extracts the history row in obtaining the history ranking information, or the data issued from application program shop
Name information.
11. according to the method described in claim 1, it is characterised in that the history ranking information was included in history each period
The user that the user of the application program makes to the application program evaluates.
12. according to the method described in claim 1, it is characterised in that methods described also includes：By the detected application
The active period of program is sent in application developer, application program shop operator, application user at least
One.
13. according to the method described in claim 1, it is characterised in that methods described also includes：By the detected ranking
Fraud detection result is sent at least one in application program shop operator, application user.
14. the ranking fraud detection system of a kind of application program, it is characterised in that the system includes：
Active period detection unit, the active period for detecting the application program based on history ranking information；The application program
Active period for frequently occur application program Active event period；The Active event of the application program is held for application program
Continuous ranking higher period；
Ranking fraud detection unit, for being tested based at least one evidence related to user's evaluation the active period
Card, obtains ranking fraud the result；The ranking fraud detection unit further comprises：
Proof validation module, for the active period to be verified and obtained based at least one evidence related to user's evaluation
To a fraud parameter；The evidence related to user's evaluation is based on the average ratings in the active periodAveragely commented with history
ValencyConstitute.
15. system according to claim 14, it is characterised in that the evidence related to user's evaluation is based on applying journey
The distribution of opinion rating is constituted in the distribution of opinion rating of the sequence in the active period and history evaluation information.
16. system according to claim 14, it is characterised in that the proof validation module, described for considering
At least one evidence related to user's evaluation, by what is obtained based at least one described proof validation related to user's evaluation
Correspondence fraud parameter is weighted, so as to obtain the fraud parameter.
17. the system according to any one of claim 14-16, it is characterised in that the ranking fraud detection unit enters
One step includes：
Parameter judge module is cheated, for the fraud parameter to be compared with a threshold value, so as to judge the application program
With the presence or absence of ranking fraud.
18. system according to claim 14, it is characterised in that the system also includes：
History ranking information acquiring unit, for obtaining the history ranking of the application program in application program ranking list
Information.
19. system according to claim 18, it is characterised in that the history ranking information acquiring unit, for from should
The history is extracted in obtaining the history ranking information, or the data issued from application program shop with program shop operator
Ranking information.
20. system according to claim 14, it is characterised in that the system also includes an active period transmitting element, is used
In the active period of the detected application program is sent into application developer, application program shop operation
At least one in business, application user.
21. system according to claim 14, it is characterised in that the system also includes a ranking fraud detection result and sent out
Unit is sent, for the detected ranking fraud detection result to be sent into application program shop operator, application program
At least one in user.
22. the ranking fraud detection method of a kind of application program, it is characterised in that methods described includes：
The active period of application program is verified based at least one evidence related to user's evaluation, ranking fraud is obtained and tests
Demonstrate,prove result；The active period of the application program is the period for the Active event for frequently occurring application program；The application program
Active event is that application program continues ranking higher period；
Methods described further comprises：
Proof validation step, is verified to the active period based at least one evidence related to user's evaluation and obtains one
Cheat parameter；The evidence related to user's evaluation is based on the average ratings in the active periodWith history average ratingsConstitute.
23. method according to claim 22, it is characterised in that
The fraud parameter is the average ratings in the active periodWith history average ratingsDifference or ratio.
24. method according to claim 22, it is characterised in that
The fraud parameter is the average ratings in the active periodWith history average ratingsDifference averagely commented with history
ValencyRatio.
25. method according to claim 22, it is characterised in that
The distribution of opinion rating of the evidence related to user's evaluation based on application program in the active period and history
The distribution of opinion rating is constituted in evaluation information.
26. method according to claim 25, it is characterised in that
Described cheat is evaluated in the distribution and history evaluation information that parameter is opinion rating of the application program in the active period
Difference between the distribution of grade.
27. method according to claim 26, it is characterised in that by calculating application program commenting in the active period
COS distance in the distribution of valency grade and history evaluation information between the distribution of opinion rating calculates the difference between them.
28. method according to claim 22, it is characterised in that in the proof validation step, considers described
At least one evidence related to user's evaluation, by what is obtained based at least one described proof validation related to user's evaluation
Correspondence fraud parameter is weighted, so as to obtain the fraud parameter.
29. the method according to any one of claim 22-28, it is characterised in that methods described further comprises：
Parameter judgment step is cheated, the fraud parameter is compared with a threshold value, so as to whether judge the application program
There is ranking fraud.
30. method according to claim 22, it is characterised in that methods described also includes：By the detected row
Name fraud detection result is sent at least one in application program shop operator, application user.
31. the ranking fraud detection system of a kind of application program, it is characterised in that the system includes：
Ranking fraud detection unit, for being entered based at least one evidence related to user's evaluation to the active period of application program
Row checking, obtains ranking fraud the result；The active period of the application program is the frequent Active event for occurring application program
Period；The Active event of the application program is that application program continues ranking higher period；
The ranking fraud detection unit further comprises：
Proof validation module, for the active period to be verified and obtained based at least one evidence related to user's evaluation
To a fraud parameter；The evidence related to user's evaluation is based on the average ratings in the active periodAveragely commented with history
ValencyConstitute.
32. system according to claim 31, it is characterised in that the evidence related to user's evaluation is based on applying journey
The distribution of opinion rating is constituted in the distribution of opinion rating of the sequence in the active period and history evaluation information.
33. system according to claim 31, it is characterised in that the proof validation module, described for considering
At least one evidence related to user's evaluation, by what is obtained based at least one described proof validation related to user's evaluation
Correspondence fraud parameter is weighted, so as to obtain the fraud parameter.
34. the system according to any one of claim 31-33, it is characterised in that the ranking fraud detection unit enters
One step includes：
Parameter judge module is cheated, for the fraud parameter to be compared with a threshold value, so as to judge the application program
With the presence or absence of ranking fraud.
35. system according to claim 31, it is characterised in that the system also includes a ranking fraud detection result and sent out
Unit is sent, for the detected ranking fraud detection result to be sent into application program shop operator, application program
At least one in user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310469920.3A CN103577541B (en) | 2013-10-10 | 2013-10-10 | The ranking fraud detection method and ranking fraud detection system of application program |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310469920.3A CN103577541B (en) | 2013-10-10 | 2013-10-10 | The ranking fraud detection method and ranking fraud detection system of application program |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103577541A CN103577541A (en) | 2014-02-12 |
| CN103577541B true CN103577541B (en) | 2017-10-10 |
Family
ID=50049317
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310469920.3A Active CN103577541B (en) | 2013-10-10 | 2013-10-10 | The ranking fraud detection method and ranking fraud detection system of application program |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103577541B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103530796B (en) | 2013-10-10 | 2016-06-01 | 北京智谷睿拓技术服务有限公司 | The active period detection method of application program and active period detection system |
| CN103559208B (en) * | 2013-10-10 | 2017-03-01 | 北京智谷睿拓技术服务有限公司 | The ranking fraud detection method of application program and ranking fraud detection system |
| CN106528525B (en) * | 2016-09-30 | 2021-02-12 | 广州酷狗计算机科技有限公司 | Method and device for identifying cheating on ranking list |
| CN107391548B (en) * | 2017-04-06 | 2020-08-04 | 华东师范大学 | Mobile application market examination user group detection method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103177109A (en) * | 2013-03-27 | 2013-06-26 | 四川长虹电器股份有限公司 | Application ranking optimization method |
| CN103235815A (en) * | 2013-04-25 | 2013-08-07 | 北京小米科技有限责任公司 | Display method and display device for application software |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8682888B2 (en) * | 2011-12-14 | 2014-03-25 | Joseph Smith | System and methods for tasking, collecting, and dispatching information reports |
-
2013
- 2013-10-10 CN CN201310469920.3A patent/CN103577541B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103177109A (en) * | 2013-03-27 | 2013-06-26 | 四川长虹电器股份有限公司 | Application ranking optimization method |
| CN103235815A (en) * | 2013-04-25 | 2013-08-07 | 北京小米科技有限责任公司 | Display method and display device for application software |
Non-Patent Citations (2)
| Title |
|---|
| Review Spam Detection via Temporal Pattern Discovery;Sihong Xie等;《In Proceedings of the 18th ACM SIGKDD international conference on knowledge discovery and data mining,KDD’12》;20121231;第823-831页 * |
| 刷榜那些事;虎嗅网，TMT青年沙龙笔记;《http://www.huxiu.com/article/5382/1.html》;20121101;第1-10页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103577541A (en) | 2014-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103559208B (en) | The ranking fraud detection method of application program and ranking fraud detection system | |
| CN103853948B (en) | The identification of user identity, the filtering of information and searching method and server | |
| AU2013209248B2 (en) | Automated discovery of gaming preferences | |
| CN108665159A (en) | A kind of methods of risk assessment, device, terminal device and storage medium | |
| CN103577541B (en) | The ranking fraud detection method and ranking fraud detection system of application program | |
| KR101427233B1 (en) | System and Method of Recommendation Number of Lotto Lottery Number for Providing Lotto Lottery for Increasing Winning Ration Using Data Mining | |
| CN103577542B (en) | The ranking fraud detection method and ranking fraud detection system of application program | |
| CN103577543B (en) | The ranking fraud detection method and ranking fraud detection system of application program | |
| CN103559210B (en) | The ranking fraud detection method and ranking fraud detection system of application program | |
| CN117291649B (en) | Intensive marketing data processing method and system | |
| CN107909516A (en) | A kind of problem source of houses recognition methods and system | |
| CN110399559A (en) | Intelligence insurance recommender system and computer storage medium | |
| CN102982048B (en) | A kind of method and apparatus for being used to assess junk information mining rule | |
| CN104618347B (en) | A kind of game events processing unit and method, the network platform | |
| CN109842858A (en) | A kind of service exception order detection method and device | |
| CN109214634A (en) | A kind of information processing method, device and information processing readable medium | |
| CN110442852B (en) | Target cost compiling method, storage medium and intelligent terminal thereof | |
| CN111489190A (en) | Anti-cheating method and system based on user relationship | |
| CN103593355A (en) | User original content recommending method and device | |
| CN114065051A (en) | Private domain platform video recommendation method and device, electronic equipment and medium | |
| KR20140114161A (en) | System and Method for Processing Number of Lotto Lottery for Increasing Winning Ration for Member Recommendation | |
| CN104252540B (en) | A kind of game configuration method for pushing graded based on computer performance | |
| CN110009012A (en) | A kind of risk specimen discerning method, apparatus and electronic equipment | |
| CN107430590A (en) | Data compare | |
| CN106600432B (en) | Inter-behavior influence evaluation method and device based on social attribute behavior data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |