CN108596632A - A kind of anti-cheating recognition methods and system based on order attributes and user behavior - Google Patents
A kind of anti-cheating recognition methods and system based on order attributes and user behavior Download PDFInfo
- Publication number
- CN108596632A CN108596632A CN201710132470.7A CN201710132470A CN108596632A CN 108596632 A CN108596632 A CN 108596632A CN 201710132470 A CN201710132470 A CN 201710132470A CN 108596632 A CN108596632 A CN 108596632A
- Authority
- CN
- China
- Prior art keywords
- order
- cheating
- user
- passenger
- driver
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
- G06Q30/0185—Product, service or business identity fraud
Landscapes
- Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Finance (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present embodiments relate to Internet technical field more particularly to a kind of anti-cheating recognition methods and system based on order attributes and user behavior.The anti-cheating recognition methods based on order attributes and user behavior that an embodiment of the present invention provides a kind of, obtains the behavioural information of the attribute information and the relevant user of the order of the order；According to the attribute information and the behavioural information, the cheating probability of the order is obtained by prediction model of practising fraud；The cheating probability of the order is compared with first threshold, if the cheating probability of the order is more than the first threshold, judges the order for order of practising fraud.The present invention considers the attribute information of order and the behavioural information with the relevant user of the order, and identify whether the order is cheating order by practising fraud prediction model, accuracy is high, and then the accuracy of identification order can be effectively improved, also it is avoided that the user to be practised fraud utilizes simultaneously, to extend the use effect of cheating prediction model.
Description
Technical field
The present embodiments relate to technical field of data processing more particularly to a kind of based on order attributes and user behavior
Anti- cheating recognition methods and system.
Background technology
Currently, the anti-cheating for scene of calling a taxi is exactly order, driver and the passenger for going identification abnormal by technological means, skill
The core of art is how accurately to identify these abnormal orders and user.Common technological means is, according to historical information and
Experience, which is formulated, identifies the anti-rule practised fraud, if user behavior, by this rule coverage, which is exactly the user that practises fraud.Example
Such as, some passenger is used for a long time great number coupons and pays fare, then the suspicion of this passenger cheating is just very big.
Since the existing anti-rule practised fraud of identification is the single anti-cheating rule being manually set, according only to historical information
Judge with experience, departing from the self attributes information for judging order is treated, lacks flexibility and updating ability, one side accuracy
It is difficult to ensure；On the other hand, in order to obtain platform subsidy, while hiding the anti-cheating strategy of platform, driver and passenger usually can
The anti-cheating strategy of research platform, single anti-cheating rule are easy to be found, and then the driver and passenger's profit that can be practised fraud
With so single anti-cheating rule often begins to fail after a period of time, service life is short.
Therefore, the problem of order, passenger and driver for how accurately identifying cheating is current industry needs urgently to be resolved hurrily
Project.
Invention content
For the defects in the prior art, the embodiment of the present invention provides a kind of anti-work based on order attributes and user behavior
Disadvantage recognition methods and system.
On the one hand, the embodiment of the present invention provides a kind of anti-cheating recognition methods based on order attributes and user behavior, packet
It includes：
Obtain the behavioural information of the attribute information and the relevant user of the order of the order；
According to the attribute information and the behavioural information, the cheating that the order is obtained by prediction model of practising fraud is general
Rate；
The cheating probability of the order is compared with first threshold, if the cheating probability of the order is more than described the
One threshold value then judges the order for order of practising fraud.
On the other hand, the embodiment of the present invention provides a kind of anti-cheating identifying system based on order attributes and user behavior,
Including：
First acquisition unit, the behavior letter of attribute information and the relevant user of the order for obtaining the order
Breath；
Second acquisition unit, for according to the attribute information and the behavioural information, prediction model to obtain by practising fraud
The cheating probability of the order；
First judging unit, for the cheating probability of the order to be compared with first threshold, if the order
Probability of practising fraud is more than the first threshold, then judges the order for order of practising fraud.
Anti- cheating recognition methods and system provided in an embodiment of the present invention based on order attributes and user behavior, synthesis are examined
Consider the attribute information of order and the behavioural information with the relevant user of the order, and the order is identified by practising fraud prediction model
Whether it is cheating order, accuracy is high, and then can effectively improve the accuracy of identification order, while being also avoided that and being practised fraud
User utilizes, to extend the use effect of cheating prediction model.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is this hair
Some bright embodiments for those of ordinary skill in the art without creative efforts, can be with root
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is that the flow of the anti-cheating recognition methods provided in an embodiment of the present invention based on order attributes and user behavior is shown
It is intended to；
Fig. 2 is to establish to make in the anti-cheating recognition methods provided in an embodiment of the present invention based on order attributes and user behavior
The flow diagram of disadvantage prediction model.
Specific implementation mode
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical solution in the embodiment of the present invention is explicitly described, it is clear that described embodiment is the present invention
A part of the embodiment, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not having
The every other embodiment obtained under the premise of creative work is made, shall fall within the protection scope of the present invention.
Fig. 1 is that the flow of the anti-cheating recognition methods provided in an embodiment of the present invention based on order attributes and user behavior is shown
It is intended to, as shown in Figure 1, the present embodiment provides a kind of anti-cheating recognition methods based on order attributes and user behavior, including：
S1, obtain the order attribute information and the relevant user of the order behavioural information；
Specifically, it is provided by the present application it is anti-cheating recognition methods considered order attribute information and with the order phase
The behavioural information of the user of pass can effectively prevent the user that existing single rule is easy to be found, and then can be practised fraud
The problem of being utilized, causing recognition accuracy low and easy failure；Wherein, user includes driver and lower single user, for ease of retouching
It states, driver is labeled as driver, then respective markers are passenger to lower single user.
S2, according to the attribute information and the behavioural information, the cheating of the order is obtained by prediction model of practising fraud
Probability；
Specifically, the attribute information of the order obtained in input step S1 in trained cheating prediction model and with this
The behavioural information of the relevant user of order, after being computed, can export the order whether be practise fraud order probability, to be follow-up
Judge the order whether for cheating order data supporting is provided.Wherein, the cheating probability of order is higher, then what the order was practised fraud can
Energy property is bigger.
S3, the cheating probability of the order is compared with first threshold, if the cheating probability of the order is more than institute
First threshold is stated, then judges the order for order of practising fraud.
Specifically, the cheating probability for the order to be identified being calculated in step S2 is compared with first threshold,
If the cheating probability of the order to be identified is more than first threshold, assert that the order to be identified is cheating order；Wherein, first
The choice relation of threshold value selects high threshold value to the identification of cheating order, the identification for order of practising fraud can be more accurate, but phase
The cheating quantity on order that can be identified answered can reduce, and rational determining first threshold can be specifically carried out according to actual implementation condition
Numerical values recited.
Anti- cheating recognition methods and system provided in an embodiment of the present invention based on order attributes and user behavior, synthesis are examined
Consider the attribute information of order and the behavioural information with the relevant user of the order, and the order is identified by practising fraud prediction model
Whether it is cheating order, accuracy is high, and then can effectively improve the accuracy of identification order, while being also avoided that and being practised fraud
User utilizes, to extend the use effect of cheating prediction model.
On the basis of the above embodiments, further, in step s3, further include：If the cheating of the order to be identified
Probability is less than or equal to first threshold, then assert that the order to be identified is normal order.In the present embodiment, the state of the order
It is divided into two kinds of cheating order and normal order, passes through the cheating probability and the first threshold of the order that cheating prediction model is calculated
The comparison of value, to judge the order for cheating order or normal order.
Further, on the basis of the above embodiments, after judging the order for cheating order, so that it may with according to work
Disadvantage order come judge the user whether be cheating user, i.e.,：
Increase the cheating order numbers of the corresponding user of the cheating order；By the cheating order numbers compared with second threshold
Compared with if the cheating order numbers judge the user for the user that practises fraud more than second threshold.Specifically, increase and the cheating
The cheating order numbers of the relevant user of order, can such as take single to add one mode to increase the cheating order numbers of user, when
When the cheating order numbers of the user are more than second threshold, then the user is classified as cheating user.Wherein, the selection of second threshold is straight
The identification for being related to cheating user is connect, that is, selects high threshold value, the identification of the user that practises fraud can be more accurate, but accordingly can be with
The cheating number of users of identification can reduce, and the numerical value that rational determining second threshold can be specifically carried out according to actual implementation condition is big
It is small.
Particularly, in the present embodiment, user includes driver and passenger, and specifically, cheating driver sentences with the passenger's that practises fraud
It is fixed, can be according to the corresponding cheating order numbers of a driver or passenger, and judged after threshold value comparison corresponding with setting,
As it is considered that the cheating order numbers of certain passenger are more than n, (wherein, threshold value is set as n), then the passenger is cheating passenger；Can such as it recognize
For certain driver cheating order numbers be more than m (wherein, threshold value is set as m), then the driver be cheating driver；The value of m and n can basis
Specific implementation condition freely adjusts.
In addition, anti-cheating recognition methods and system provided in an embodiment of the present invention based on order attributes and user behavior,
Can be used for it is all exist " order ", " passenger ", " driver " user O2O platforms, such as：Passenger, driver's brush singly gain the benefit of platform by cheating
Patch.Especially by consider order attribute information and driver, passenger the order behavioural information, based on cheating prediction mould
Type come identify the order whether be cheating order, improve identification order accuracy, while also avoid the driver to be practised fraud and
Passenger utilizes, to extend the use effect of cheating prediction model.
On the basis of the above embodiments, further, before step S2, the method further includes cheating prediction model
Establish process, specifically as shown in Fig. 2, including：
P1, order sample set is obtained, at least one of described order sample set order sample includes the category of order
Property information, the user are in the behavioural information of the order sample and the historical behavior information of the user；Wherein, user includes department
Machine and passenger；
Specifically, order sample set is first obtained, which includes several order samples, at least one to order
Single sample includes the attribute information of order, the user in the behavioural information of the order sample and the historical behavior letter of the user
Breath；The order sample that part passenger uses for the first time only includes the attribute information of order, driver and passenger believes in the behavior of the order
The historical behavior information of breath and driver；The order sample that part driver uses for the first time only includes the attribute information of order, Si Jihe
Passenger is in the behavioural information of the order and the historical behavior information of passenger.
P2, it trains using machine learning algorithm according to the order sample set and obtains the cheating prediction model.
Specifically, the order sample set in step P1 is analyzed, using machine learning algorithm, mould is predicted to cheating
The parameter of type optimizes, and cheating prediction model is obtained by training.Wherein, under line training when, select machine learning algorithm into
The continuous iteration of row and Optimized model, the cheating prediction model that the accuracy until obtaining identification cheating order is met the requirements are real
It is now simple, should be readily appreciated that and realize, calculating speed quickly and storage resource is low.Wherein, machine learning algorithm can determine for iteration
Plan tree, or logistic regression, or to be other as NB Algorithm or average list rely on estimation scheduling algorithm.
Wherein, cheating prediction model has considered the attribute information of order, driver and passenger believes in the behavior of the order
It ceases, the historical behavior information of driver and passenger, avoids the department that existing single rule is easy to be found, and then can be practised fraud
The problem of machine and passenger utilize, cause recognition accuracy low and easy failure.
On the basis of the above embodiments, further, it according to the order sample set, is calculated using machine learning
Method optimizes the parameter of the cheating prediction model, before training obtains the cheating prediction model, further includes：It is based on
Judgment rule identifies the order classification of each order sample of the order sample set, the order classification be positive sample or
Negative sample.
Specifically, the order classification in the order sample set can be divided into positive sample and negative sample two major classes；Wherein,
Positive sample refers to that normal order, negative sample refer to cheating order, and positive sample and negative sample are confirmed according to judgment rule
's.Particularly, cheating order includes driver's cheating order and passenger's cheating order；It accidentally injures, judge driver or multiplies in order to prevent
Whether visitor practises fraud, and corresponding threshold value can be arranged to judge, specifically according to the corresponding cheating order numbers of a driver or passenger
Threshold value can be accordingly arranged according to the accuracy rate that required negative sample identifies.
On the basis of the above embodiments, further, it after the training obtains the cheating prediction model, also wraps
It includes：The parameter of the cheating prediction model is optimized, until the cheating prediction model identification cheating order after optimization
Accuracy reach third threshold value；Wherein, the accuracy of the identification cheating order includes the recall rate of the negative sample identification
With the accuracy rate of negative sample identification.Wherein, the two is defined as follows：
Negative sample recall rate=a/ (a+b)；Negative sample accuracy rate=a/ (a+c)
Wherein, a：Former negative sample, is predicted as negative sample；b：Former negative sample, is predicted as positive sample；c：Former positive sample, prediction
For negative sample.
Particularly, the recall rate of negative sample identification and accuracy rate the two index ordinary circumstances of negative sample identification are negative senses
Relationship, the accuracy rate of the more high then negative sample identification of the negative sample recall rate of same model can be lower, specifically can be according to business
Needs select to need the index be biased to.Such as, it is desirable that more as possible recalls negative sample, identification cheating order, then negative sample
The accuracy rate of this identification can be down slightly；The order it is required that high-precision identification is practised fraud, then negative sample identification is recalled
Rate can be down slightly.
Anti- cheating recognition methods provided by the present application based on order attributes and user behavior has considered to be identified order
Single attribute information, driver, passenger establish cheating and predict in the behavioural information of list and the historical behavior information of driver and passenger
Model, substantially increases the accuracy of identification order, while the driver and passenger to be practised fraud also being avoided to utilize.
On the basis of the above embodiments, further, the historical behavior information of the user includes the one of following information
Kind is a variety of：The history cheating of the passenger, the registion time of the passenger, the passenger cell-phone number/equipment whether
It is the order of this day of black list user, the history cheating of the driver, the registion time of the driver and the driver
Number；The attribute information of the order includes the one or more of following information：The bill time of the passenger, the hair of the passenger
Single region, the driver complete the speed of the order, the order apart from expense, when the driver completes the order
Time；The behavioural information includes the one or more of following information：The travel speed of the driver, the traveling of the driver
Apart from the position of terminal when the position and the passenger of track, the passenger apart from order starting point are paid.
Wherein, the bill region of the passenger refers to passenger its region when sending out order；The order away from
It is referred to from expense ratio：Driver is from passenger place is connect to the amount of money for being sent to the distance between passenger place with passenger and paying the order
Ratio, can judge that the order belongs to cheating order or normal order based on experience value.
On the basis of the above embodiments, further, the embodiment of the present invention additionally provide it is a kind of based on order attributes and
The anti-cheating identifying system of user behavior, including：
First acquisition unit, the behavior letter of attribute information and the relevant user of the order for obtaining the order
Breath；Second acquisition unit, for according to the attribute information and the behavioural information, being ordered described in prediction model acquisition by practising fraud
Single cheating probability；First judging unit, for the cheating probability of the order to compare with first threshold, if described order
Single cheating probability is more than the first threshold, then judges the order for order of practising fraud.Wherein, user includes driver and multiplies
Visitor.
Specifically, the anti-cheating identifying system provided by the present application based on order attributes and user behavior passes through first
One acquiring unit come obtain order attribute information and driver, passenger the order behavioural information, secondly by second obtain
Unit order calculates the order as cheating order according to the attribute information and behavioural information of order in conjunction with cheating prediction model
Probability, by the order be that practise fraud probability and the first threshold of order does ratio finally by the first judging unit, ordered if obtaining
Single cheating probability is more than first threshold, then judges the order for order of practising fraud；Correspondingly, if the cheating probability for obtaining order is small
In equal to first threshold, then judge the order be normal order.Wherein, knowledge of the choice relation of first threshold to cheating order
Not, that is, high threshold value is selected, the identification for order of practising fraud can cheating quantity on order meeting that is more accurate, but can identifying accordingly
It reduces, the rational numerical values recited for determining first threshold can be specifically carried out according to actual implementation condition.
On the basis of the above embodiments, further, the system also includes：
Recording unit, for it is described judge that the order is to practise fraud order after, the corresponding use of the increase cheating order
The cheating order numbers at family；
Second judging unit, for the cheating order numbers to compare with second threshold, if the cheating order numbers are big
In second threshold, then judge the user for the user that practises fraud.
Specifically, after judging the order for cheating order, the corresponding use of the cheating order is increased by recording unit
The cheating order numbers at family, and the cheating order numbers of the user after increase are compared with second threshold, if the cheating order numbers
More than second threshold, then judge the user for the user that practises fraud.
User includes driver and passenger, if the order is cheating, order further can be each according to driver and passenger respectively
Self-corresponding threshold value is further judged whether the driver is cheating driver and whether the passenger is cheating passenger.Wherein,
Discrimination standard is if more than default cheating threshold value, then is the user that practises fraud；If it is just common less than or equal to default cheating threshold value
Family.
On the basis of the above embodiments, further, the system also includes：
Third acquiring unit, for being obtained before the cheating probability for obtaining the order by prediction model of practising fraud
Order sample set, at least one of described order sample set order sample include the attribute information of order, the user
In the behavioural information of the order sample and the historical behavior information of the user；Modeling unit, for according to the order sample
Set, using machine learning algorithm, training obtains the cheating prediction model.
It needs first to establish cheating prediction model before calculating the cheating probability for obtaining order in computing unit, specifically, first
Order sample set is obtained by third acquiring unit, the information of order sample set is analyzed secondly by modeling unit, is used
Machine learning algorithm, training obtain cheating prediction model.
On the basis of the above embodiments, further, the system also includes：Recognition unit, in the basis
The order sample set before training obtains the cheating prediction model, is known using machine learning algorithm based on judgment rule
The order classification of each order sample of the not described order sample set, the order classification are positive sample or negative sample.
Specifically, before training obtains cheating prediction model, the every of order sample set is identified by recognition unit
The order classification of one order is done with the accuracy rate of recall rate and negative sample identification for subsequent calculating negative sample identification
Prepare.
On the basis of the above embodiments, further, the system also includes：Optimize unit, in the training
After obtaining the cheating prediction model, the parameter of the cheating prediction model is optimized, until the work after optimization
The accuracy of disadvantage prediction model identification cheating order reaches third threshold value；Wherein, the accuracy of the identification cheating order includes
The accuracy rate of the recall rate and negative sample identification of the negative sample identification.
User includes passenger and driver, specifically, the historical behavior information of the user include following information one kind or
It is a variety of：The history cheating of the passenger, the registion time of the passenger, the passenger cell-phone number/equipment whether be black
Name single user, the history cheating of the driver, the registion time of the driver and the driver are within a preset period of time
Order numbers；The attribute information includes the one or more of following information：The bill time of the passenger, the bill of the passenger
Region, the driver complete the speed of the order, the order apart from expense, when the driver completes the order
Time；The behavioural information includes the one or more of following information：The travel speed of the driver, the driver traveling rail
Apart from the position of terminal when the position and the passenger of mark, the passenger apart from order starting point are paid.
Finally it should be noted that：The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, it will be understood by those of ordinary skill in the art that：It still may be used
With technical scheme described in the above embodiments is modified or equivalent replacement of some of the technical features；
And these modifications or replacements, various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (17)
1. a kind of anti-cheating recognition methods based on order attributes and user behavior, which is characterized in that including：
Obtain the behavioural information of the attribute information and the relevant user of the order of the order；
According to the attribute information and the behavioural information, the cheating probability of the order is obtained by prediction model of practising fraud；
The cheating probability of the order is compared with first threshold, if the cheating probability of the order is more than first threshold
Value then judges the order for order of practising fraud.
2. according to the method described in claim 1, it is characterized in that, if the cheating probability of the order is less than or equal to described the
One threshold value then judges the order for normal order.
3. according to the method described in claim 1, it is characterized in that, it is described judge the order for cheating order after, it is described
Method further includes：
Increase the cheating order numbers of the corresponding user of the cheating order；
The cheating order numbers are compared with second threshold, if the cheating order numbers be more than second threshold, judge described in
User is cheating user.
4. according to the method described in claim 1, it is characterized in that, obtaining the order by prediction model of practising fraud described
It practises fraud before probability, the method further includes：
Obtain order sample set, at least one of described order sample set order sample include order attribute information,
The user is in the behavioural information of the order sample and the historical behavior information of the user；
According to the order sample set, using machine learning algorithm, training obtains the cheating prediction model.
5. according to the method described in claim 4, it is characterized in that, described according to the order sample set, using machine
Learning algorithm trains before obtaining the cheating prediction model, further includes：
Identify that the order classification of each order sample of the order sample set, the order classification are based on judgment rule
Positive sample or negative sample.
6. according to the method described in claim 5, it is characterized in that, after the training obtains the cheating prediction model,
Further include：
The parameter of the cheating prediction model is optimized, until the cheating prediction model identification cheating order after optimization
Accuracy reach third threshold value；Wherein, the accuracy of the identification cheating order includes the recall rate of the negative sample identification
With the accuracy rate of negative sample identification.
7. method according to any one of claims 1 to 6, which is characterized in that the user includes passenger and driver.
8. the method according to the description of claim 7 is characterized in that the historical behavior information of the user includes following information
It is one or more：The history cheating of the passenger, the registion time of the passenger, the passenger cell-phone number/equipment be
No is black list user, the history cheating of the driver, the registion time of the driver and the driver in preset time
Order numbers in section.
9. the method according to the description of claim 7 is characterized in that the attribute information includes one kind or more of following information
Kind：The bill time of the passenger, the bill region of the passenger, the driver complete the speed of the order, the order
Apart from expense, when the driver completes time of the order；
The behavioural information includes the one or more of following information：The travel speed of the driver, the driver traveling rail
Apart from the position of terminal when the position and the passenger of mark, the passenger apart from order starting point are paid.
10. a kind of anti-cheating identifying system based on order attributes and user behavior, which is characterized in that including：
First acquisition unit, the behavioural information of attribute information and the relevant user of the order for obtaining the order；
Second acquisition unit is used for according to the attribute information and the behavioural information, by practising fraud described in prediction model acquisition
The cheating probability of order；
First judging unit, for the cheating probability of the order to compare with first threshold, if the cheating of the order
Probability is more than the first threshold, then judges the order for order of practising fraud.
11. system according to claim 10, which is characterized in that the system also includes：
Recording unit, for it is described judge that the order is cheating order after, increase is described to practise fraud the corresponding user's of order
Cheating order numbers；
Second judging unit, for the cheating order numbers to compare with second threshold, if the cheating order numbers are more than the
Two threshold values then judge the user for the user that practises fraud.
12. system according to claim 10, which is characterized in that the system also includes：
Third acquiring unit, for obtaining order before the cheating probability for obtaining the order by prediction model of practising fraud
Sample set, at least one of described order sample set order sample include the attribute information of order, the user at this
The behavioural information of order sample and the historical behavior information of the user；
Modeling unit, for according to the order sample set, using machine learning algorithm, training to obtain the cheating prediction mould
Type.
13. system according to claim 12, which is characterized in that the system also includes：
Recognition unit, for, according to the order sample set, using machine learning algorithm, training to obtain the cheating described
Before prediction model, the order classification of each order sample of the order sample set is identified based on judgment rule, it is described
Order classification is positive sample or negative sample.
14. system according to claim 13, which is characterized in that the system also includes：
Optimize unit, after obtaining the cheating prediction model in the training, to the parameter of the cheating prediction model
It optimizes, until the accuracy of the cheating prediction model identification cheating order after optimization reaches third threshold value；Wherein, institute
The accuracy for stating identification cheating order includes the accuracy rate of the recall rate and negative sample identification of the negative sample identification.
15. according to claim 10-14 any one of them systems, which is characterized in that the user includes driver and passenger.
16. system according to claim 15, which is characterized in that the historical behavior information of the user includes following information
It is one or more：The history cheating of the passenger, the registion time of the passenger, the passenger cell-phone number/equipment
Whether it is black list user, the history cheating of the driver, the registion time of the driver and the driver when default
Between order numbers in section.
17. according to claim 15 any one of them system, which is characterized in that the attribute information includes the one of following information
Kind is a variety of：The bill time of the passenger, the bill region of the passenger, the driver complete the speed of the order, institute
State order apart from expense, when the driver completes time of the order；
The behavioural information includes the one or more of following information：The travel speed of the driver, the driver traveling rail
Apart from the position of terminal when the position and the passenger of mark, the passenger apart from order starting point are paid.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710132470.7A CN108596632A (en) | 2017-03-07 | 2017-03-07 | A kind of anti-cheating recognition methods and system based on order attributes and user behavior |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710132470.7A CN108596632A (en) | 2017-03-07 | 2017-03-07 | A kind of anti-cheating recognition methods and system based on order attributes and user behavior |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108596632A true CN108596632A (en) | 2018-09-28 |
Family
ID=63632903
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710132470.7A Pending CN108596632A (en) | 2017-03-07 | 2017-03-07 | A kind of anti-cheating recognition methods and system based on order attributes and user behavior |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108596632A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110659954A (en) * | 2019-08-29 | 2020-01-07 | 北京三快在线科技有限公司 | Cheating identification method and device, electronic equipment and readable storage medium |
| CN111105284A (en) * | 2018-10-29 | 2020-05-05 | 北京嘀嘀无限科技发展有限公司 | Order processing method and device, M-layer order processing model, electronic equipment and storage medium |
| CN111126629A (en) * | 2019-12-25 | 2020-05-08 | 上海携程国际旅行社有限公司 | Model generation method, system, device and medium for identifying brushing behavior |
| CN111262854A (en) * | 2020-01-15 | 2020-06-09 | 卓望数码技术（深圳）有限公司 | Internet anti-cheating behavior method, device, equipment and readable storage medium |
| CN111861496A (en) * | 2019-04-11 | 2020-10-30 | 天津五八到家科技有限公司 | Driver management and control and order auditing method, equipment and storage medium |
| CN112184334A (en) * | 2020-10-27 | 2021-01-05 | 北京嘀嘀无限科技发展有限公司 | Method, apparatus, device and medium for determining problem users |
| CN112837119A (en) * | 2021-01-28 | 2021-05-25 | 天津五八到家货运服务有限公司 | Abnormal order identification method and device, electronic equipment and storage medium |
| CN113723970A (en) * | 2021-08-25 | 2021-11-30 | 深圳依时货拉拉科技有限公司 | Order list pushing method and device, storage medium and computer equipment |
| CN114078016A (en) * | 2020-08-12 | 2022-02-22 | 腾讯科技（深圳）有限公司 | Anti-cheating behavior identification method and device, electronic equipment and storage medium |
| CN114119037A (en) * | 2022-01-24 | 2022-03-01 | 深圳尚米网络技术有限公司 | Marketing anti-cheating system based on big data |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105389722A (en) * | 2015-11-20 | 2016-03-09 | 小米科技有限责任公司 | Malicious order identification method and device |
| CN105894359A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | Order pushing method, device and system |
| CN105894360A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | Cheated order identification method, device and system |
| CN106384273A (en) * | 2016-10-08 | 2017-02-08 | 江苏通付盾科技有限公司 | Malicious order scalping detection system and method |
-
2017
- 2017-03-07 CN CN201710132470.7A patent/CN108596632A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105389722A (en) * | 2015-11-20 | 2016-03-09 | 小米科技有限责任公司 | Malicious order identification method and device |
| CN105894359A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | Order pushing method, device and system |
| CN105894360A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | Cheated order identification method, device and system |
| CN106384273A (en) * | 2016-10-08 | 2017-02-08 | 江苏通付盾科技有限公司 | Malicious order scalping detection system and method |
Non-Patent Citations (1)
| Title |
|---|
| 赵乃真: "《电子商务技术与应用》", 31 January 2010 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111105284A (en) * | 2018-10-29 | 2020-05-05 | 北京嘀嘀无限科技发展有限公司 | Order processing method and device, M-layer order processing model, electronic equipment and storage medium |
| CN111105284B (en) * | 2018-10-29 | 2024-04-19 | 北京嘀嘀无限科技发展有限公司 | Order processing method and device, M-layer order processing model, electronic equipment and storage medium |
| CN111861496A (en) * | 2019-04-11 | 2020-10-30 | 天津五八到家科技有限公司 | Driver management and control and order auditing method, equipment and storage medium |
| CN110659954B (en) * | 2019-08-29 | 2022-06-17 | 北京三快在线科技有限公司 | Cheating identification method and device, electronic equipment and readable storage medium |
| CN110659954A (en) * | 2019-08-29 | 2020-01-07 | 北京三快在线科技有限公司 | Cheating identification method and device, electronic equipment and readable storage medium |
| CN111126629B (en) * | 2019-12-25 | 2023-09-19 | 上海携程国际旅行社有限公司 | Model generation method, brush list identification method, system, equipment and medium |
| CN111126629A (en) * | 2019-12-25 | 2020-05-08 | 上海携程国际旅行社有限公司 | Model generation method, system, device and medium for identifying brushing behavior |
| CN111262854A (en) * | 2020-01-15 | 2020-06-09 | 卓望数码技术（深圳）有限公司 | Internet anti-cheating behavior method, device, equipment and readable storage medium |
| CN114078016A (en) * | 2020-08-12 | 2022-02-22 | 腾讯科技（深圳）有限公司 | Anti-cheating behavior identification method and device, electronic equipment and storage medium |
| CN114078016B (en) * | 2020-08-12 | 2023-10-10 | 腾讯科技（深圳）有限公司 | Anti-cheating behavior identification method and device, electronic equipment and storage medium |
| CN112184334A (en) * | 2020-10-27 | 2021-01-05 | 北京嘀嘀无限科技发展有限公司 | Method, apparatus, device and medium for determining problem users |
| CN112837119A (en) * | 2021-01-28 | 2021-05-25 | 天津五八到家货运服务有限公司 | Abnormal order identification method and device, electronic equipment and storage medium |
| CN113723970A (en) * | 2021-08-25 | 2021-11-30 | 深圳依时货拉拉科技有限公司 | Order list pushing method and device, storage medium and computer equipment |
| CN113723970B (en) * | 2021-08-25 | 2024-02-02 | 深圳依时货拉拉科技有限公司 | Order list pushing method, order list pushing device, storage medium and computer equipment |
| CN114119037A (en) * | 2022-01-24 | 2022-03-01 | 深圳尚米网络技术有限公司 | Marketing anti-cheating system based on big data |
| CN114119037B (en) * | 2022-01-24 | 2022-05-17 | 深圳尚米网络技术有限公司 | Marketing anti-cheating system based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108596632A (en) | A kind of anti-cheating recognition methods and system based on order attributes and user behavior | |
| CN110070391B (en) | Data processing method and device, computer readable medium and electronic equipment | |
| US8527317B2 (en) | Service level agreement work prioritization system | |
| US20170262770A1 (en) | Cascaded boosted predictive models | |
| CN108846547A (en) | A kind of Enterprise Credit Risk Evaluation method of dynamic adjustment | |
| CN107330731A (en) | It is a kind of to recognize that advertisement position clicks on abnormal method and apparatus | |
| CN108280542A (en) | A kind of optimization method, medium and the equipment of user's portrait model | |
| CN105719045A (en) | Retention risk determiner | |
| CN104866922B (en) | A kind of off-grid prediction technique of user and device | |
| CN108734296A (en) | Optimize method, apparatus, electronic equipment and the medium of the training data of supervised learning | |
| CN116342073B (en) | Book printing digital information management system and method thereof | |
| US11734766B1 (en) | Telematics devices and ridesharing | |
| CN111192090A (en) | Seat allocation method and device for flight, storage medium and electronic equipment | |
| CN104484558A (en) | Method and system for automatically generating analysis reports of biological information projects | |
| CN106408325A (en) | User consumption behavior prediction analysis method based on user payment information and system | |
| CN111340606A (en) | Full-process income auditing method and device | |
| CN107122786A (en) | A kind of mass-rent learning method and device | |
| CN110874744A (en) | Data anomaly detection method and device | |
| CN109684309A (en) | A kind of quality of data evaluating method and device, computer equipment and storage medium | |
| CN112907356A (en) | Overdue collection method, device and system and computer readable storage medium | |
| CN106713267A (en) | Network security assessment method and system | |
| CN111932106A (en) | Effective and practical cloud manufacturing task and service resource matching method | |
| CN112187653A (en) | Network flow judgment method and system | |
| CN112134898A (en) | Network flow judgment method and system | |
| CN107886217A (en) | A kind of labor turnover Risk Forecast Method and device based on clustering algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180928 |