CN108182587A - A kind of electric business platform brush single act detection method and system - Google Patents
A kind of electric business platform brush single act detection method and system Download PDFInfo
- Publication number
- CN108182587A CN108182587A CN201810084277.5A CN201810084277A CN108182587A CN 108182587 A CN108182587 A CN 108182587A CN 201810084277 A CN201810084277 A CN 201810084277A CN 108182587 A CN108182587 A CN 108182587A
- Authority
- CN
- China
- Prior art keywords
- brush
- rate
- value
- probability
- commodity
- 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
-
- 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0609—Buyer or seller confidence or verification
-
- 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0633—Lists, e.g. purchase orders, compilation or processing
- G06Q30/0635—Processing of requisition or of purchase orders
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Economics (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)
- Entrepreneurship & Innovation (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of electric business platform brush single act detection method and system.This method includes the initial data for obtaining multiple commodity, multiple commodity include the single commodity of brush and do not brush single commodity, feature rate value is calculated according to initial data, feature rate value includes silent conversion ratio, order consulting rate, traffic transformation rate, conclusion of the business conversion ratio, order payment rate, collection rate, time difference of receiving, brush hand accounting, shop residence time；Feature rate value is normalized and be converted into the form to match with support vector cassification method, obtains characteristic value；The characteristic value of the single commodity of brush is handled using support vector cassification method and does not brush the characteristic value of single commodity, obtains optimal training pattern；The characteristic value of target data is calculated, obtains object feature value；The brush list probability of end article is calculated according to optimal training pattern and object feature value.Optimal training pattern is obtained by calculating feature rate value, the brush list probability of end article is calculated, is supplied to the intuitive reference data of user.
Description
Technical field
The present invention relates to electric business field, more particularly to a kind of electric business platform brush single act detection method and system.
Background technology
With the fast development of e-commerce industry, shopping at network is increasingly becoming a kind of new life style, electric business industry
Competition also grow in intensity, under the driving of interests, electric business behavior has been increasingly becoming " underlying rule " of electric business platform, due to electric business
Just under development, various constrained qualifications are all not perfect for business, so the brush single act of electric business platform is than more serious.
For brush single act, the counter-brush single system in Jingdone district is counted from multiple dimensions such as order, commodity, user, logistics,
The different characteristic value under each dimension is calculated respectively, can more precisely identify the malicious act of brush simple correlation.
Single auditing system is brushed in detection from the background for Taobao, in terms of mainly being examined including machine with manual examination and verification two, for wherein machine
The careful order judged that is difficult to obtains final result into artificial investigation, and hotel owner can appeal, and can enter after complaint and manually judge rank
Section, by checking that the content of commodity evaluation, Bidder Information are judged.
Detection method brush list detection of the prior art is only limitted to self supervise and examine mechanism of electric business platform interior, single for brush
The single shop of doubtful brush in testing result, major electric business platform are also only that drop power or positive closing are carried out to its part, and
Non- external disclosure, user can not obtain an intuitive data and check the single situation of brush, principle based on benefit and work as
Preceding electric business current situation, electric business platform can not fundamentally prevent to brush single phenomenon.
Invention content
The object of the present invention is to provide a kind of brush list phenomenon electric business platform brushes that can fundamentally prevent in electric business platform
Single act detection method and system.
To achieve the above object, the present invention provides following schemes：
A kind of electric business platform brush single act detection method, the detection method include：
The initial data of multiple commodity is obtained, the multiple commodity include the single commodity of brush and do not brush single commodity, described original
Data include visitor's number, consulting number, payment number, order numbers, collection number, number of clicks, buyer ID, lower single time, confirm and receive
Time, payment time, shop residence time, exchange hour, IP address information.
Feature rate value is calculated according to the initial data, the feature rate value includes silent conversion ratio, order consulting rate, stream
Measure conversion ratio, conclusion of the business conversion ratio, order payment rate, collection rate, time difference of receiving, brush hand accounting, shop residence time.It is described quiet
Silent conversion ratio is the order numbers without consulting and the quotient of visitor's number, the order consulting rate for the consulting number with
The quotient of visitor's number, the traffic transformation rate are the payment number and the quotient of the click volume, and the conclusion of the business conversion ratio is institute
The quotient of payment number and visitor's number is stated, the order payment rate is the payment number and the quotient of the order numbers, the collection
Quotient of the rate for the collection number and visitor's number, the time difference of receiving receive time and the payment time for the confirmation
Difference, the brush hand accounting be IP repetitive rates.
The feature rate value is normalized and be converted into the form to match with support vector cassification method, is obtained
Obtain characteristic value.
The characteristic value of the single commodity of the brush and the characteristic value for not brushing single commodity are handled using support vector cassification method,
Obtain optimal training pattern.
The target data of end article is obtained, calculates the characteristic value of the target data, obtains object feature value.
The brush list probability of the end article is calculated according to the optimal training pattern and the object feature value.
Optionally, it is described that single quotient described is not brushed according to the characteristic values of the single commodity of brush and using support vector cassification method
The characteristic value of product obtains optimal training pattern and specifically includes：
It is more according to the characteristic value of the single commodity of the brush and the characteristic value for not brushing single commodity using K folding cross validation algorithms
Penalty factor and kernel function in the secondary verification support vector cassification method, obtain optimal penalty factor and optimal kernel function.
The optimal training pattern is obtained according to the optimal penalty factor and the optimal kernel function.
Optionally, the brush list that the end article is calculated according to the optimal training pattern and the object feature value
Probability specifically includes：
The classification results of the end article are calculated according to the optimal training pattern and the object feature value, are divided
Class is as a result, the classification results include brush list and do not brush list.
Conditional probability of multiple object feature values in the classification results is calculated respectively, obtains characteristic condition probability.
The probability that different classifications result occurs is calculated respectively, obtains class probability.
According to the brush list probability of end article described in the class probability and the characteristic condition probability calculation.
Optionally, the method for obtaining the initial data of multiple commodity specifically includes：Literature query, businessman's investigation and webpage are climbed
Any one in worm.
To achieve these goals, the present invention also provides following technical solutions：
A kind of electric business platform brush single act detecting system, the detecting system include：
Initial data acquisition module, for obtaining the initial data of multiple commodity, the multiple commodity include the single commodity of brush
Single commodity are not brushed, and the initial data includes visitor's number, consulting number, payment number, order numbers, collection number, number of clicks, buyer
ID, lower single time, confirm receive time, payment time, shop residence time, exchange hour, IP address information.
Feature rate value computing module is connect with the initial data acquisition module, and the feature rate value computing module is used for root
Calculate feature rate value according to the initial data, the feature rate value include silent conversion ratio, order consulting rate, traffic transformation rate,
Conclusion of the business conversion ratio, order payment rate, collection rate, time difference of receiving, brush hand accounting, shop residence time.
Feature rate value modular converter is connect with the feature rate value computing module, and the feature rate value modular converter is used for will
The feature rate value is normalized and is converted into the form to match with support vector cassification method, obtains characteristic value.
Training module is connect with the feature rate value modular converter, and the training module is used for using support vector cassification
The characteristic value and the characteristic value for not brushing single commodity of the single commodity of the method processing brush, obtain optimal training pattern.
Object feature value acquisition module is connect with the training module, for obtaining the target data of end article, is calculated
The characteristic value of the target data obtains object feature value.
The single probability evaluation entity of brush is connect respectively with the object feature value acquisition module and the training module, for root
The brush list probability of the end article is calculated according to the optimal training pattern and the object feature value.
Optionally, the single probability evaluation entity of the brush specifically includes：
Classification results computing unit, for calculating the target according to the optimal training pattern and the object feature value
The classification results of commodity, obtain classification results, and the classification results include brush list and do not brush list.
Conditional probability computing unit is connect with the classification results computing unit, and the conditional probability computing unit is used to divide
Conditional probability of multiple object feature values in the classification results is not calculated, obtains characteristic condition probability.
Class probability computing unit is connect with the conditional probability computing unit, and the class probability computing unit is used to divide
Not Ji Suan different classifications result occur probability, obtain class probability.
The single probability calculation unit of brush is connect with the class probability computing unit, and the single probability calculation unit of brush is used for root
According to the brush list probability of end article described in the class probability and the characteristic condition probability calculation.
According to specific embodiment provided by the invention, the invention discloses following technique effects：The present invention provides one kind
Electric business platform brush single act detection method and system, by obtaining the initial data of multiple commodity, the initial data includes visiting
Objective number, consulting number, payment number, order numbers, collection number, number of clicks, buyer ID, lower single time, when confirming time of receiving, payment
Between, the shop residence time, exchange hour, IP address information；Feature rate value, the feature rate value are calculated according to the initial data
Including silent conversion ratio, order consulting rate, traffic transformation rate, conclusion of the business conversion ratio, order payment rate, collection rate, the time difference of receiving,
Brush hand accounting, shop residence time, then the brush list probability of the end article is obtained by calculating optimal training pattern, it is directly logical
Data are crossed to user about commodity brush one-state intuitively to experience.
Description of the drawings
It in order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to institute in embodiment
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the present invention
Example, for those of ordinary skill in the art, without having to pay creative labor, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is the flow chart of electric business platform brush single act detection method provided by the invention；
Fig. 2 calculates the end article to be provided by the invention according to the optimal training pattern and the object feature value
The single probability of brush flow chart；
Fig. 3 is the structure chart of electric business platform brush single act detecting system provided by the invention.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other without making creative work
Embodiment shall fall within the protection scope of the present invention.
The object of the present invention is to provide a kind of brush list phenomenon electric business platform brushes that can fundamentally prevent in electric business platform
Single act detection method and system.
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, it is below in conjunction with the accompanying drawings and specific real
Applying mode, the present invention is described in further detail.
A kind of flow chart of electric business platform brush single act detection method as shown in Figure 1, the detection method include：
Step 100：The initial data of multiple commodity is obtained, the multiple commodity include the single commodity of brush and do not brush single commodity,
The initial data include visitor's number, consulting number, payment number, order numbers, collection number, number of clicks, buyer ID, lower single time,
Confirm receive time, payment time, shop residence time, exchange hour, IP address information.
Step 200：Feature rate value is calculated according to the initial data, the feature rate value includes silent conversion ratio, order
When consulting rate, traffic transformation rate, conclusion of the business conversion ratio, order payment rate, collection rate, time difference of receiving, brush hand accounting, shop stop
Between, the shop residence time represents that brush hand in order to save the time, completes more brush single tasks, the corresponding residence time can compare
It is shorter.
The silence conversion ratio is the order numbers without consulting and the quotient of visitor's number, the order consulting rate
For the quotient of the consulting number and visitor's number, the traffic transformation rate is the quotient of the payment number and the click volume, described
Quotient of the conclusion of the business conversion ratio for the payment number and visitor's number, the order payment rate are the payment number and the order numbers
Quotient, quotient of the collection rate for collection number and visitor's number, the time difference of receiving receives the time for the confirmation
With the difference of the payment time, the brush hand accounting is IP repetitive rates.
The silence conversion ratio represents that buyer not by seeking advice from the details of customer service commodity, directly places an order, can reflect
The integral level in shop.The order consulting rate represents the accounting of lower single number in customer service reception number, comprehensive consideration customer service
Service skill and ability.The conclusion of the business conversion ratio and the traffic transformation rate represent that web hit is converted into purchasing power or profit
The ratio of profit.The order payment rate represents that buyer can be cancelled an order and arrearage due to personal, comes compared to the single commodity of brush
It says, the order payment rate of Normal Goods can be than relatively low.The collection rate represents that if buyer likes certain shop that can collect the shop, to obtain
Brush single act can be reflected by obtaining the prompting of new or discount information, the collection rate on shop.The time difference of receiving represents that brush hand is
The time of receiving can be shortened as possible per single exchange hour by saving.The IP repetitive rates can reflect the identity of brush hand.
Step 300：The feature rate value is normalized and is converted into match with support vector cassification method
Form, obtain characteristic value, normalized method so that the processing procedure of algorithm is more convenient, while accelerates training process
Convergence.
Step 400：The characteristic value of the single commodity of brush is handled using support vector cassification method and described does not brush single commodity
Characteristic value, obtain optimal training pattern.
Step 500：The target data of end article is obtained, calculates the characteristic value of the target data, obtains target signature
Value.
Step 600：The brush that the end article is calculated according to the optimal training pattern and the object feature value is singly general
Rate.
The step 400：List according to the characteristic values of the single commodity of brush and described is not brushed using support vector cassification method
The characteristic value of commodity, obtain optimal training pattern the specific steps are：
It is more according to the characteristic value of the single commodity of the brush and the characteristic value for not brushing single commodity using K folding cross validation algorithms
Penalty factor and kernel function in the secondary verification support vector cassification method, obtain optimal penalty factor and optimal kernel function.
The optimal training pattern is obtained according to the optimal penalty factor and the optimal kernel function.
The step 600 as shown in Figure 2：The mesh is calculated according to the optimal training pattern and the object feature value
The flow chart of the single probability of brush of commodity is marked, is specifically included：
Step 601：The classification knot of the end article is calculated according to the optimal training pattern and the object feature value
Fruit, obtains classification results, and the classification results include brush list and do not brush list.
Step 602：Conditional probability of multiple object feature values in the classification results is calculated respectively, obtains characteristic condition
Probability.
Step 603：The probability that different classifications result occurs is calculated respectively, obtains class probability.
Step 604：According to the brush list probability of end article described in the class probability and the characteristic condition probability calculation.
The method for obtaining the initial data of multiple commodity specifically includes：In literature query, businessman's investigation and spiders
Any one, businessman's investigation is mainly that sales situation and the shop of its commodity are acquired to the electric business in Taobao and Jingdone district platform
The information such as credit worthiness；Web crawlers is then to utilize collection of the crawler technology from the commodity basic information page into row information；Document is looked into
It askes and the single detection relevant information of brush is mainly collected by paper book under the electronic theories such as middle National IP Network library and line.
To achieve these goals, the present invention also provides following technical solutions：
A kind of electric business platform brush single act detecting system, the detecting system include：
Initial data acquisition module 1, for obtaining the initial data of multiple commodity, the multiple commodity include the single commodity of brush
Single commodity are not brushed, and the initial data includes visitor's number, consulting number, payment number, order numbers, collection number, number of clicks, buyer
ID, lower single time, confirm receive time, payment time, shop residence time, exchange hour, IP address information.
Feature rate value computing module 2 is connect with the initial data acquisition module 1, and the feature rate value computing module 2 is used
In calculating feature rate value according to the initial data, the feature rate value includes silent conversion ratio, order consulting rate, traffic transformation
Rate, conclusion of the business conversion ratio, order payment rate, collection rate, time difference of receiving, brush hand accounting, shop residence time.
Feature rate value modular converter 3 is connect with the feature rate value computing module 2, and the feature rate value modular converter 3 is used
In being normalized and being converted into the form to match with support vector cassification method by the feature rate value, feature is obtained
Value.
Training module 4 is connect with the feature rate value modular converter 3, and the training module 4 is used for using support vector machines
The characteristic value and the characteristic value for not brushing single commodity of the single commodity of the classification processing brush, obtain optimal training pattern.
Object feature value acquisition module 5 for obtaining the target data of end article, calculates the feature of the target data
Value obtains object feature value.
The single probability evaluation entity 6 of brush is connect respectively with the object feature value acquisition module 5 and the training module 4, is used
In the brush list probability that the end article is calculated according to the optimal training pattern and the object feature value.
Optionally, the single probability evaluation entity 6 of the brush specifically includes：
Classification results computing unit, for calculating the target according to the optimal training pattern and the object feature value
The classification results of commodity, obtain classification results, and the classification results include brush list and do not brush list.
Conditional probability computing unit is connect with the classification results computing unit, and the conditional probability computing unit is used to divide
Conditional probability of multiple object feature values in the classification results is not calculated, obtains characteristic condition probability.
Class probability computing unit is connect with the conditional probability computing unit, and the class probability computing unit is used to divide
Not Ji Suan different classifications result occur probability, obtain class probability.
The single probability calculation unit of brush is connect with the class probability computing unit, and the single probability calculation unit of brush is used for root
According to the brush list probability of end article described in the class probability and the characteristic condition probability calculation.
The advantages of using support vector cassification method：
(1) the brush list probability accuracy highest obtained, and speed；The accuracy rate of BP neural network algorithm is placed in the middle, but
Speed is most slow, computationally intensive；NB Algorithm is fastest but accuracy rate is minimum, because naive Bayesian requirement is every special
Between sign independently of each other, but there is certain contact between brush forms data characteristic item of the present invention, so NB Algorithm is not
It is suitble to require herein, from the point of view of comprehensive three kinds of algorithms, the sorting algorithm used herein is more suitable.
(2) mode being association of activity and inertia is employed.When user detection commodity have existed relevant information in the database, then
For static detection, the data in database need to be only converted in test sample input algorithm and analyzed, obtain the single probability of brush
And be stored in database, it can save the time when being detected so as to next time for same commodity.When the commodity of user's detection exist
When relevant information being not present in database, crawled in webpage into Mobile state first with crawler technology according to network address input by user
Hold, obtain the commodity up-to-date information in user's selection period.
(3) there is preferable integrality and closed loop, can the shop excessively high to the single probability of brush sound a warning information, not only
It can be based on existing shop merchandise news in database to be detected, and can realize that dynamic updates, ensure that brush is single general
The availability and accuracy of rate testing result；After final testing result is provided, user also is able to check the detailed of selected commodity
Thin test data and data of the same trade, it is simple and clear, increase the convincingness of result.
(4) user to multiple commodity can brush the detection of single probability simultaneously, and final system can show selected
The result of calculation of multiple commodity so that user carries out the comparison of similar commodity brush single act testing result, and can be shown simultaneously
Show the data information of multiple commodity.
Each embodiment is described by the way of progressive in this specification, the highlights of each of the examples are with other
The difference of embodiment, just to refer each other for identical similar portion between each embodiment.For system disclosed in embodiment
For, since it is corresponded to the methods disclosed in the examples, so description is fairly simple, related part is said referring to method part
It is bright.
Specific case used herein is expounded the principle of the present invention and embodiment, and above example is said
The bright method and its core concept for being merely used to help understand the present invention；Meanwhile for those of ordinary skill in the art, foundation
The thought of the present invention, in specific embodiments and applications there will be changes.In conclusion the content of the present specification is not
It is interpreted as limitation of the present invention.
Claims (6)
1. a kind of electric business platform brush single act detection method, which is characterized in that the detection method includes：
The initial data of multiple commodity is obtained, the multiple commodity include the single commodity of brush and do not brush single commodity, the initial data
Including visitor's number, consulting number, payment number, order numbers, collection number, number of clicks, buyer ID, lower single time, confirm the time of receiving,
Payment time, shop residence time, exchange hour, IP address information；
Feature rate value is calculated according to the initial data, the feature rate value includes silent conversion ratio, order consulting rate, flow and turns
Rate, conclusion of the business conversion ratio, order payment rate, collection rate, time difference of receiving, brush hand accounting, shop residence time；Described silent turn
Rate is the order numbers without consulting and the quotient of visitor's number, the order consulting rate for the consulting number with it is described
The quotient of visitor's number, the traffic transformation rate are the payment number and the quotient of the click volume, and the conclusion of the business conversion ratio is paid to be described
Amount of money and the quotient of visitor's number, the order payment rate are the payment number and the quotient of the order numbers, and the collection rate is
The quotient of the collection number and visitor's number, the time difference of receiving receive time and the payment time for the confirmation
Difference, the brush hand accounting are IP repetitive rates；
The feature rate value is normalized and be converted into the form to match with support vector cassification method, is obtained special
Value indicative；
The characteristic value of the single commodity of the brush and the characteristic value for not brushing single commodity are instructed using support vector cassification method
Practice, obtain optimal training pattern；
The target data of end article is obtained, calculates the characteristic value of the target data, obtains object feature value；
The brush list probability of the end article is calculated according to the optimal training pattern and the object feature value.
2. a kind of electric business platform brush single act detection method according to claim 1, which is characterized in that described using support
Characteristic value and the characteristic value of not brushing single commodity of the vector machine classification according to the single commodity of the brush, obtain optimal training pattern
It specifically includes：
It is repeatedly tested according to the characteristic value and the characteristic value for not brushing single commodity of the single commodity of the brush using K folding cross validation algorithms
The penalty factor and kernel function in the support vector cassification method are demonstrate,proved, obtains optimal penalty factor and optimal kernel function；
The optimal training pattern is obtained according to the optimal penalty factor and the optimal kernel function.
3. a kind of electric business platform brush single act detection method according to claim 1, which is characterized in that described in the basis
The brush list probability that optimal training pattern and the object feature value calculate the end article specifically includes：
The classification results of the end article are calculated according to the optimal training pattern and the object feature value, obtain classification knot
Fruit, the classification results include brush list and do not brush list；
Conditional probability of multiple object feature values in the classification results is calculated respectively, obtains characteristic condition probability；
The probability that different classifications result occurs is calculated respectively, obtains class probability；
According to the brush list probability of end article described in the class probability and the characteristic condition probability calculation.
4. a kind of electric business platform brush single act detection method according to claim 1, which is characterized in that obtain multiple commodity
The method of initial data specifically include：Literature query, businessman investigation and spiders in any one.
5. a kind of electric business platform brush single act detecting system, which is characterized in that the detecting system includes：
Initial data acquisition module, for obtaining the initial data of multiple commodity, the multiple commodity are including the single commodity of brush and not
The single commodity of brush, the initial data include visitor's number, consulting number, payment number, order numbers, collection number, number of clicks, buyer ID,
The lower list time will confirm receive time, payment time, shop residence time, exchange hour, IP address information；
Feature rate value computing module is connect with the initial data acquisition module, and the feature rate value computing module is used for according to institute
It states initial data and calculates feature rate value, the feature rate value includes silent conversion ratio, order consulting rate, traffic transformation rate, strikes a bargain
Conversion ratio, order payment rate, collection rate, time difference of receiving, brush hand accounting, shop residence time；
Feature rate value modular converter is connect with the feature rate value computing module, and the feature rate value modular converter is used for by described in
Feature rate value is normalized and is converted into the form to match with support vector cassification method, obtains characteristic value；
Training module is connect with the feature rate value modular converter, and the training module is used for using at support vector cassification method
The characteristic value of the single commodity of the brush and the characteristic value for not brushing single commodity are managed, obtains optimal training pattern；
Object feature value acquisition module is connect with the training module, for obtaining the target data of end article, described in calculating
The characteristic value of target data obtains object feature value；
The single probability evaluation entity of brush is connect respectively with the object feature value acquisition module and the training module, for according to institute
It states optimal training pattern and the object feature value calculates the brush list probability of the end article.
A kind of 6. electric business platform brush single act detecting system according to claim 5, which is characterized in that the single probability of brush
Computing module specifically includes：
Classification results computing unit, for calculating the end article according to the optimal training pattern and the object feature value
Classification results, obtain classification results, it is single and do not brush list that the classification results include brush；
Conditional probability computing unit is connect with the classification results computing unit, and the conditional probability computing unit is based on respectively
Conditional probability of multiple object feature values in the classification results is calculated, obtains characteristic condition probability；
Class probability computing unit is connect with the conditional probability computing unit, and the class probability computing unit is based on respectively
The probability that different classifications result occurs is calculated, obtains class probability；
The single probability calculation unit of brush is connect with the class probability computing unit, and the single probability calculation unit of brush is used for according to institute
State the brush list probability of end article described in class probability and the characteristic condition probability calculation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810084277.5A CN108182587A (en) | 2018-01-29 | 2018-01-29 | A kind of electric business platform brush single act detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810084277.5A CN108182587A (en) | 2018-01-29 | 2018-01-29 | A kind of electric business platform brush single act detection method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108182587A true CN108182587A (en) | 2018-06-19 |
Family
ID=62551549
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810084277.5A Pending CN108182587A (en) | 2018-01-29 | 2018-01-29 | A kind of electric business platform brush single act detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108182587A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109064189A (en) * | 2018-07-13 | 2018-12-21 | 北京亚鸿世纪科技发展有限公司 | Brush list detecting and alarm device based on the detection of intensive block |
| CN109377301A (en) * | 2018-08-27 | 2019-02-22 | 中国民航信息网络股份有限公司 | A kind of Feature Extraction Method based on Airline reservation behavioral data |
| CN111553726A (en) * | 2020-04-22 | 2020-08-18 | 上海海事大学 | HMM-based (hidden Markov model) -based system and method for predicting bill swiping |
| CN112734508A (en) * | 2021-03-24 | 2021-04-30 | 于淼 | E-commerce transaction data analysis system based on cloud platform |
| CN113051433A (en) * | 2019-12-27 | 2021-06-29 | 华为技术有限公司 | Data processing method and device |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105069626A (en) * | 2015-07-23 | 2015-11-18 | 北京京东尚科信息技术有限公司 | Detection method and detection system for shopping abnormity |
| CN106059847A (en) * | 2016-04-08 | 2016-10-26 | 久远谦长（北京）技术服务有限公司 | User ordering behavior detection method and device |
| CN106096974A (en) * | 2016-06-02 | 2016-11-09 | 中国联合网络通信集团有限公司 | A kind of anti-cheat method for shopping at network and system |
| CN106204157A (en) * | 2016-07-24 | 2016-12-07 | 广东聚联电子商务股份有限公司 | Behavior processing method evaluated by a kind of brush list based on big data collection and analysis |
| 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 |
| CN106557955A (en) * | 2016-11-29 | 2017-04-05 | 流量海科技成都有限公司 | Net about car exception order recognition methodss and system |
| CN107146089A (en) * | 2017-03-29 | 2017-09-08 | 北京三快在线科技有限公司 | The single recognition methods of one kind brush and device, electronic equipment |
| CN107305665A (en) * | 2016-04-19 | 2017-10-31 | 李辉 | It is a kind of to differentiate wash sale, prevent the single method and device of brush |
| CN107464169A (en) * | 2017-08-10 | 2017-12-12 | 北京小度信息科技有限公司 | Information output method and device |
-
2018
- 2018-01-29 CN CN201810084277.5A patent/CN108182587A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105069626A (en) * | 2015-07-23 | 2015-11-18 | 北京京东尚科信息技术有限公司 | Detection method and detection system for shopping abnormity |
| CN106059847A (en) * | 2016-04-08 | 2016-10-26 | 久远谦长（北京）技术服务有限公司 | User ordering behavior detection method and device |
| CN107305665A (en) * | 2016-04-19 | 2017-10-31 | 李辉 | It is a kind of to differentiate wash sale, prevent the single method and device of brush |
| CN106096974A (en) * | 2016-06-02 | 2016-11-09 | 中国联合网络通信集团有限公司 | A kind of anti-cheat method for shopping at network and system |
| CN106204157A (en) * | 2016-07-24 | 2016-12-07 | 广东聚联电子商务股份有限公司 | Behavior processing method evaluated by a kind of brush list based on big data collection and analysis |
| 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 |
| CN106557955A (en) * | 2016-11-29 | 2017-04-05 | 流量海科技成都有限公司 | Net about car exception order recognition methodss and system |
| CN107146089A (en) * | 2017-03-29 | 2017-09-08 | 北京三快在线科技有限公司 | The single recognition methods of one kind brush and device, electronic equipment |
| CN107464169A (en) * | 2017-08-10 | 2017-12-12 | 北京小度信息科技有限公司 | Information output method and device |
Non-Patent Citations (5)
| Title |
|---|
| 乐元: "基于C2C电子商务市场虚假销量识别的若干技术研究", 《中国优秀博硕士学位论文全文数据库(硕士)经济与管理科学辑》 * |
| 康海燕 等: "CTCIS2017+79面向用户的电商平台刷单行为智能检测方法", 《WWW.JOCA.CN/CN/ABSTRACT/ABSTRACT21281.SHTML》 * |
| 康海燕 等: "面向用户的电商平台刷单行为智能检测方法", 《计算机应用》 * |
| 忆臻: "带你理解朴素贝叶斯分类算法", 《HTTPS://ZHUANLAN.ZHIHU.COM/P/26262151》 * |
| 谢军 等: "基于网购评论分析的电商平台刷单行为推测方法", 《吉林工商学院学报》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109064189A (en) * | 2018-07-13 | 2018-12-21 | 北京亚鸿世纪科技发展有限公司 | Brush list detecting and alarm device based on the detection of intensive block |
| CN109377301A (en) * | 2018-08-27 | 2019-02-22 | 中国民航信息网络股份有限公司 | A kind of Feature Extraction Method based on Airline reservation behavioral data |
| CN113051433A (en) * | 2019-12-27 | 2021-06-29 | 华为技术有限公司 | Data processing method and device |
| CN111553726A (en) * | 2020-04-22 | 2020-08-18 | 上海海事大学 | HMM-based (hidden Markov model) -based system and method for predicting bill swiping |
| CN112734508A (en) * | 2021-03-24 | 2021-04-30 | 于淼 | E-commerce transaction data analysis system based on cloud platform |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108182587A (en) | A kind of electric business platform brush single act detection method and system | |
| CN108573432B (en) | Transaction supervision system and method for electronic commerce | |
| US8615458B2 (en) | Industry size of wallet | |
| US20090070289A1 (en) | Methods, Systems, and Computer Program Products for Estimating Accuracy of Linking of Customer Relationships | |
| US20160321679A1 (en) | Device and membership identity matching | |
| KR101265976B1 (en) | A industrial technology market analysis system and based on the quantitative information, and method thereof | |
| CA2722119A1 (en) | Payment portfolio optimization | |
| CN102081774A (en) | Card-raising identification method and system | |
| CN105074748A (en) | Profiling auction assets and/or participants to predict auction outcome | |
| Gupta et al. | Audit selection strategy for improving tax compliance: application of data mining techniques | |
| Maranzato et al. | Fraud detection in reputation systems in e-markets using logistic regression | |
| CN102004999A (en) | Behaviour revenue model based collusion group identification method in electronic commerce network | |
| CN110009365B (en) | User group detection method, device and equipment for abnormally transferring electronic assets | |
| Maranzato et al. | Fraud detection in reputation systems in e-markets using logistic regression and stepwise optimization | |
| CN110009449A (en) | Business trust appraisal procedure, server and storage medium | |
| CN109767270A (en) | The old information recommendation method of housing support and system are deposited based on artificial intelligence | |
| JP4755911B2 (en) | Scoring system and scoring method for calculating fraud score using credit card | |
| CN112819476A (en) | Risk identification method and device, nonvolatile storage medium and processor | |
| CN105631732A (en) | Method and device for determining user authority | |
| CN117236996B (en) | User behavior prediction method and system based on big data analysis | |
| CN112330373A (en) | User behavior analysis method and device and computer readable storage medium | |
| Mzoughia et al. | An improved customer lifetime value model based on Markov chain | |
| US20140164057A1 (en) | Systems and methods for determining consumer purchasing behavior | |
| KR101409273B1 (en) | Method and apparatus for calculating to measure influential power of users | |
| CN113592505B (en) | System for realizing suspicious transaction scene model identification processing based on combination construction |
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 |