CN110111131A - The determination method and device of false visitor's standing breath - Google Patents
The determination method and device of false visitor's standing breath Download PDFInfo
- Publication number
- CN110111131A CN110111131A CN201910276821.0A CN201910276821A CN110111131A CN 110111131 A CN110111131 A CN 110111131A CN 201910276821 A CN201910276821 A CN 201910276821A CN 110111131 A CN110111131 A CN 110111131A
- Authority
- CN
- China
- Prior art keywords
- visitor
- user
- feature
- channel
- objective standing
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
Abstract
The invention discloses a kind of determination method and devices of false objective standing breath.Wherein, method includes: to obtain objective standing breath, and provide feature according to visitor money information extraction visitor；It obtains and ceases associated user data with objective standing, user characteristics are extracted based on user data；And obtain and cease associated channel data with objective standing, channel feature is extracted based on channel data；According to visitor's money feature, user characteristics and channel feature, the falseness visitor that training constructs in advance, which provides, determines model, and the determining model of falseness visitor's money after train；When receiving new objective standing breath, the false grade for determining that model determines new objective standing breath is provided using the falseness visitor after training.The falseness visitor that this programme trains acquisition final according to the various dimensions characteristic information of visitor's money feature, user characteristics and channel feature provides determining model, so as to promote the accuracy of identification of false objective standing breath；And it avoids and at high cost, the low drawback of recognition efficiency is identified by manual identified bring.
Description
Technical field
The present invention relates to Internet technical fields, and in particular to a kind of determination method and device of false objective standing breath.
Background technique
Now, more and more trade companies tend to realize its product by the way of obtaining visitor's money (customer resources) information
Accurate dispensing, to promote its product sales volume and profit.With the rapid development of Internet technology, the canal of objective standing breath is obtained
Road is increasing.However, there are still many false objective standing breaths in the objective standing breath obtained by all kinds of means.False visitor's standing breath
Occur not only damaging the interests of the platform of businessman or businessman's docking, also will cause the harassing and wrecking to certain user, reduce user experience.
Currently, mostly using the mode of manual identified to the identification of false objective standing breath, i.e., passed through by special return visit personnel
Contact method in objective standing breath contacts user, determines whether objective standing breath is deceptive information according to contact result.However it adopts
With the mode of this kind of manual identified, operation cost can not only be substantially improved, be also unfavorable for the quick identification of false objective standing breath, know
Other inefficiency.For this drawback, the Chinese invention patent application that application publication number is CN109063433A is proposed, in telecommunications
Field can carry out the identification of fictitious users in the way of machine learning, however it is carrying out false use using machine learning
In the identification process at family, train learning model just with user data, and using the model judge user itself whether be
Fictitious users, so that the recognition accuracy of fictitious users is lower.
Summary of the invention
In view of the above problems, it proposes on the present invention overcomes the above problem or at least be partially solved in order to provide one kind
State the determination method and device of the objective standing breath of falseness of problem.
According to an aspect of the invention, there is provided a kind of determination method of false objective standing breath, comprising:
Objective standing breath is obtained, and feature is provided according to visitor's money information extraction visitor；
It obtains and ceases associated user data with the objective standing, user characteristics are extracted based on the user data；And it obtains
It takes and ceases associated channel data with the objective standing, extract channel feature based on the channel data；
According to visitor's money feature, the user characteristics and the channel feature, falseness visitor's money that training constructs in advance is true
Cover half type, and the falseness visitor after being trained provides and determines model；
When receiving new objective standing breath, is provided using the falseness visitor after training and determine that model determines the new objective standing
The false grade of breath.
Optionally, visitor's money feature includes at least one of following characteristics:
Creation time, user identifier, channel mark, estimated consumption commodity belonging to classification, estimated consumption time and
Consumer budget.
Optionally, the acquisition further comprises with the associated user data of the objective standing breath:
According to the user identifier for including in the objective standing breath, user data corresponding with the user identifier is obtained.
Optionally, the user characteristics include at least one of following characteristics:
User's history search characteristics, user's history transaction feature and user's Figure Characteristics.
Optionally, the user's history search characteristics include: the quotient for the target industry that user searches within a preset period of time
The searching times of product；
The user's history transaction feature includes: the number of the commodity for the target industry that user trades within a preset period of time
Mesh；
User's Figure Characteristics include: age, occupation, income and/or marital status；
Wherein, the target industry is the industry that the estimated commodity consumed are subordinate to.
Optionally, the acquisition further comprises with the associated channel data of the objective standing breath:
According to the channel mark for including in the objective standing breath, channel data corresponding with the channel mark are obtained.
Optionally, the channel feature includes at least one of following characteristics:
The channel in the total number and preset time period that the corresponding objective standing of the channel mark ceases in preset time period
Identify the number of corresponding effective objective standing breath.
Optionally, the method also includes: obtain the objective standing and cease corresponding standard falseness grade label；
It is described objective according to visitor's money feature, the user characteristics and the channel feature, the falseness that training constructs in advance
It provides and determines model, and the determining model of falseness visitor's money after train further comprises:
Corresponding standard is ceased according to visitor's money feature, the user characteristics, the channel feature and the objective standing
False grade label, the falseness visitor that training constructs in advance, which provides, determines model, and the determining model of falseness visitor's money after train.
Optionally, described that feature, the user characteristics and the channel feature are provided according to the visitor, what training constructed in advance
False visitor, which provides, determines model, and the determining model of falseness visitor's money after train further comprises:
By visitor's money feature, the user characteristics and the channel feature, inputs the falseness visitor's money constructed in advance and determine
Model obtains the reasoning results；
According to the comparison result of the reasoning results and the standard falseness grade label, calculating ratio is to error；
Model training is carried out using preset algorithm, to minimize the comparison error；
The determining model of falseness visitor's money corresponding when error is minimum value is compared by described, the falseness visitor after being determined as training
It provides and determines model.
Optionally, the preset algorithm is gradient descent algorithm；The falseness visitor constructed in advance, which provides, determines model for volume
Product neural network model.
According to another aspect of the present invention, a kind of determining device of false objective standing breath is provided, comprising:
Obtain module, be suitable for obtaining objective standing breath, obtain and cease associated user data with the objective standing, and obtain with
Visitor's standing ceases associated channel data；
Extraction module is suitable for providing feature according to visitor's money information extraction visitor, and it is special to extract user based on the user data
Sign, and, channel feature is extracted based on the channel data；
Training module is suitable for according to visitor's money feature, the user characteristics and the channel feature, training building in advance
Falseness visitor provide and determine model, and the determining model of falseness visitor's money after train；
Determining module, suitable for being provided using the falseness visitor after training and determining that model is determined when receiving new objective standing breath
The false grade of the new objective standing breath.
Optionally, visitor's money feature includes at least one of following characteristics:
Creation time, user identifier, channel mark, estimated consumption commodity belonging to classification, estimated consumption time and
Consumer budget.
Optionally, the module that obtains is further adapted for: according to the user identifier for including in the objective standing breath, obtain with
The corresponding user data of the user identifier.
Optionally, the user characteristics include at least one of following characteristics:
User's history search characteristics, user's history transaction feature and user's Figure Characteristics.
Optionally, the user's history search characteristics include: the quotient for the target industry that user searches within a preset period of time
The searching times of product；
The user's history transaction feature includes: the number of the commodity for the target industry that user trades within a preset period of time
Mesh；
User's Figure Characteristics include: age, occupation, income and/or marital status；
Wherein, the target industry is the industry that the estimated commodity consumed are subordinate to.
Optionally, the module that obtains is further adapted for: according to the channel mark for including in the objective standing breath, obtain with
The corresponding channel data of the channel mark.
Optionally, the channel feature includes at least one of following characteristics:
The channel in the total number and preset time period that the corresponding objective standing of the channel mark ceases in preset time period
Identify the number of corresponding effective objective standing breath.
Optionally, described device further include: label acquisition module is suitable for obtaining the objective standing and ceases corresponding standard falseness
Grade label；
The training module is further adapted for:
Corresponding standard is ceased according to visitor's money feature, the user characteristics, the channel feature and the objective standing
False grade label, the falseness visitor that training constructs in advance, which provides, determines model, and the determining model of falseness visitor's money after train.
Optionally, the training module is further adapted for:
By visitor's money feature, the user characteristics and the channel feature, inputs the falseness visitor's money constructed in advance and determine
Model obtains the reasoning results；
According to the comparison result of the reasoning results and the standard falseness grade label, calculating ratio is to error；
Model training is carried out using preset algorithm, to minimize the comparison error；
The determining model of falseness visitor's money corresponding when error is minimum value is compared by described, the falseness visitor after being determined as training
It provides and determines model.
Optionally, the preset algorithm is gradient descent algorithm；The falseness visitor constructed in advance, which provides, determines model for volume
Product neural network model.
According to a further aspect of the invention, a kind of calculating equipment is provided, comprising: processor, memory, communication interface
And communication bus, the processor, the memory and the communication interface complete mutual lead to by the communication bus
Letter；
For the memory for storing an at least executable instruction, it is as above that the executable instruction executes the processor
State the corresponding operation of determination method of the objective standing breath of the falseness.
According to a further aspect of the invention, a kind of computer storage medium is provided, is stored in the storage medium
An at least executable instruction, the executable instruction make processor execute the determination method that the objective standing of falseness as described above ceases
Corresponding operation.
A kind of determination method and device of the false objective standing breath provided according to the present invention, the objective standing breath of acquisition, and according to
Visitor money information extraction visitor provides feature；It obtains and ceases associated user data with objective standing, user characteristics are extracted based on user data；With
And obtain and cease associated channel data with objective standing, channel feature is extracted based on channel data；Feature, Yong Hute are provided according to visitor
Sign and channel feature, the falseness visitor that training constructs in advance, which provides, determines model, and the determining model of falseness visitor's money after train；When
When receiving new objective standing breath, the false grade for determining that model determines new objective standing breath is provided using the falseness visitor after training.
This programme is trained according to the various dimensions characteristic information of visitor's money feature, user characteristics and channel feature obtains final falseness visitor's money
Model is determined, so as to promote the accuracy of identification of false objective standing breath；And it avoids and is identified by manual identified bring
It is at high cost, the low drawback of recognition efficiency.
The above description is only an overview of the technical scheme of the present invention, in order to better understand the technical means of the present invention,
And it can be implemented in accordance with the contents of the specification, and in order to allow above and other objects of the present invention, feature and advantage can
It is clearer and more comprehensible, the followings are specific embodiments of the present invention.
Detailed description of the invention
By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field
Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
The process that Fig. 1 shows a kind of determination method of the false objective standing breath provided according to an embodiment of the present invention is shown
It is intended to；
Fig. 2 shows a kind of processes of the determination method of the false objective standing breath provided according to a further embodiment of the invention
Schematic diagram；
Fig. 3 shows a kind of function knot of the determining device of the false objective standing breath provided according to an embodiment of the present invention
Structure schematic diagram；
Fig. 4 shows a kind of structural schematic diagram of the calculating equipment provided according to an embodiment of the present invention.
Specific embodiment
Exemplary embodiments of the present disclosure are described in more detail below with reference to accompanying drawings.Although showing the disclosure in attached drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
It is limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure
It is fully disclosed to those skilled in the art.
The process that Fig. 1 shows a kind of determination method of the false objective standing breath provided according to an embodiment of the present invention is shown
It is intended to.As shown in Figure 1, this method comprises:
Step S110 obtains objective standing breath, and provides feature according to visitor money information extraction visitor.
In actual implementation process, the objective standing breath of multiple support channels offer can be periodically obtained, and uses corresponding feature
Extracting method extracts visitor's money feature from the objective standing breath got, which, which provides feature, can reflect that the core of objective standing breath is special
Point, such as visitor's money feature may include creation time, user identifier, channel mark, expect classification belonging to the commodity of consumption, is pre-
Count consumption time, and/or consumer budget etc..
Step S120, obtains and ceases associated user data with objective standing, extracts user characteristics based on user data, and obtain
It takes and ceases associated channel data with objective standing, extract channel feature based on channel data.
In the present embodiment, ceased according to objective standing to obtain associated user data and channel data.Wherein, correlation is being obtained
In the user data and channel data procedures of connection, specifically can according to objective standing cease in include user identifier and channel mark come
It quickly determines and ceases associated user and channel with objective standing, and then user is accurately obtained according to user identifier and channel mark
Data and channel data.
Further, according to the user data and channel data got using corresponding feature extracting method come from user
User characteristics are extracted in data, and channel feature is extracted from channel data.The user characteristics extracted can be largely anti-
Reflect user behavior feature, for example, user characteristics may include user consumer behavior feature or user essential attribute feature (such as
Gender, occupation) etc.；And the channel feature extracted can largely reflection channel confidence level, such as channel feature
It can be effective visitor's money information scales etc. of the channel.
Step S130, according to visitor's money feature, user characteristics and channel feature, the falseness visitor that training constructs in advance, which provides, determines mould
Type, and the falseness visitor after being trained provides and determines model.
The present embodiment is to carry out false visitor using the characteristic information of three dimensions to provide the training for determining model, that is, utilizes step
Visitor's money feature, user characteristics and the channel feature that rapid S110 and step S120 is obtained, to train the falseness visitor constructed in advance jointly
It provides and determines model.Wherein, the present embodiment, which provides false visitor, determines the specific training method of model without limitation, those skilled in the art
Member can voluntarily select corresponding machine learning algorithm, such as the falseness visitor constructed in advance to provide and determine that model can according to the actual situation
Think convolutional neural networks model etc..
Step S140 is provided using the false visitor after training when receiving new objective standing breath and is determined that model determines newly
The false grade of objective standing breath.
After the completion of false visitor provides and determines model training, when receiving new objective standing breath, falseness visitor can be utilized
Provide the false grade for determining that model determines new objective standing breath.Wherein, this implementation of the setting means of false grade of objective standing breath
It is not limited here, for example, the false grade of objective standing breath can for effectively, without demand, refuse connection, falseness etc..
It can be seen that this programme is based on objective standing breath, and associated user data and channel data are ceased with objective standing, extracted
Visitor's money feature, user characteristics and channel feature out, and believed according to the various dimensions feature of visitor's money feature, user characteristics and channel feature
It ceases to train the falseness visitor constructed in advance to provide and determine model, model is determined to obtain final falseness visitor and provide, eventually by instruction
Falseness visitor after white silk, which provides, determines model to carry out the determination of the false grade of new objective standing breath, so as to avoid by manually knowing
Not at high cost, the low drawback of recognition efficiency of bring identification, and the accuracy of identification of false objective standing breath with higher.
Fig. 2 shows a kind of processes of the determination method of the false objective standing breath provided according to a further embodiment of the invention
Schematic diagram.As shown in Fig. 2, this method comprises:
Step S210 obtains the objective standing breath for being labeled with standard falseness grade label in advance, and provides information extraction according to visitor
Visitor's money feature.
In the present embodiment false visitor provide determine model be trained by Supervised machine learning method, so, this
Specifically there is the objective standing breath obtained in step the objective standing of standard falseness grade label to cease.Wherein, it is different from existing
Technology only determines whether user is fictitious users, and the present embodiment is convenient for the accurate of the subsequent false grade to new objective standing breath
Identification has carried out fine-grained division to standard falseness grade in this step, such as standard falseness grade can be with are as follows: effectively, nothing
Demand refuses connection and falseness etc..Wherein, the mode manually marked can be used, carry out standard falseness grade label is ceased to objective standing
Mark, can also be used the mode for manually marking and being combined with automatic marking come to objective standing cease carry out standard falseness grade label
Mark, for example, if it is a certain visitor standing breath consumed, automatically for the visitor standing breath distribute " effective " standard falseness etc.
Grade label.
In actual implementation process, the objective standing breath obtained from each channel is normally stored in preset visitor's money information bank
In, then message identification can be provided according to visitor, being labeled in preset period of time section in advance is regularly obtained from visitor's money information bank
The objective standing breath of quasi- falseness grade label.For example, can be labeled in advance on the day before regularly obtaining the daily set time
The objective standing of standard falseness grade label ceases.
Further, feature is provided according to visitor money information extraction visitor.Wherein, visitor's money feature includes at least one in following characteristics
Kind: creation time, channel mark, expects classification belonging to the commodity of consumption, estimated consumption time and consumption at user identifier
Budget.
Below by taking marriage leave class visitor standing breath as an example, specifically to illustrate the extracted visitor's money feature of this step: where when creation
Between can be time that objective standing breath is obtained from channel, or by objective standing breath be recorded in preset visitor's money information bank when
Between；User identifier can be to uniquely determine the information of user, such as contact method (such as phone number, mailbox number etc. of user
Deng)；Channel mark is specifically as follows the channel title (such as portal website A, portal website B etc.) in visitor's standing breath source；In advance
Classification etc. can be clapped for local wedding photo classification, trip by counting classification belonging to the commodity of consumption；It is expected that consumption time can be
The specific wedding day etc. that user fills in；Consumption is default to be specifically as follows the wedding budget etc. that user specifically fills in.
Step S220, obtains and ceases associated user data with objective standing, extracts user characteristics based on user data.
In the specific implementation process, according to the user identifier for including in objective standing breath, use corresponding with user identifier is obtained
User data, and user characteristics are further extracted from user data.User characteristics in the application can be anti-from user's dimension
It reflects, classification belonging to the commodity of the estimated consumption in objective standing that user fills in breath, it is expected that consumption time and consumer budget
Authenticity.
Specifically, user characteristics include at least one of following characteristics: user's history search characteristics, user's history transaction
Feature and user's Figure Characteristics.Wherein, user's history search characteristics specifically include the mesh that user searches within a preset period of time
Mark the searching times of the commodity of industry；User's history transaction feature includes: the target industry that user trades within a preset period of time
Commodity number；User's Figure Characteristics include: age, occupation, income and/or marital status；Wherein, target industry is estimated
The industry that the commodity of consumption are subordinate to.
For example, Fig. 3 is that extracted part visitor provides feature.By taking visitor's money feature of visitor's money characteristic serial number 1 as an example, root
Search record of the user in nearly 3 months is obtained according to user identifier " 123456 ", and extracts and is subordinate to from search record
It records in the search of marriage industry (the target industry that " trip's bat " is subordinate to), and is recorded according to the search extracted, count the user
The searching times of the commodity of marriage leave industry are searched in nearly 3 months, and using the searching times as user's history search characteristics；Together
Reason, can obtain transaction record of the user in nearly 3 months, and the friendship for being under the jurisdiction of marriage industry is extracted from the transaction record
Easily record, according to the transaction record extracted, counts the number of the commodity for the marriage leave industry that the user trades in nearly 3 months,
And using the number of the commodity as user's history transaction feature；And user's Figure Characteristics specifically can be from preset user's representation data
It is obtained in library.
Table 1
In a kind of optional embodiment, preset time period described in this step can be according to specific target industry
Information carries out dynamic adjustment, so that promoting subsequent false visitor provides the precision of prediction for determining model.For example, if target industry is wedding
Industry is transferred, preset time period can be in nearly 3 months；Target industry is catering trade, and preset time period is nearly 1 month etc..
Step S230, obtains and ceases associated channel data with objective standing, extracts channel feature based on channel data.
It is the recognition accuracy for promoting fictitious users in the present embodiment, other than using user data and objective standing breath,
It also further obtains to have and be obtained with the associated channel data of objective standing breath specifically according to the channel mark for including in objective standing breath
Take channel data corresponding with channel mark.And channel feature is further extracted from channel data.What the present embodiment extracted
Channel feature can largely reflection channel confidence level, to further be conducive to the visitor's money for predicting to obtain from the channel
The false grade of information.
Specifically, the channel feature extracted includes at least one of following characteristics: channel mark pair in preset time period
The number of the corresponding effective objective standing breath of the channel mark in the total number and preset time period of the objective standing breath answered.With canal
For road CH-A, the total number for the objective standing breath that should be obtained from the channel in nearly 3 months can be obtained, and is further determined that from this
The standard falseness grade label of the objective standing breath obtained in channel is the number that the objective standing of " effective " ceases, thus by the total number
And the number of effectively objective standing breath is as channel feature.
In addition, channel feature can also include: to adopt further to promote the precision of prediction that subsequent false visitor provides determining model
The scoring to channel of collection.In actual implementation process, commenting for each channel can be obtained from preset forum, feedback of businessman etc.
Valence, and then the evaluation for obtaining channel is screened, the frequency for extracting keyword, and further being occurred according to keyword, really
Determine channel scoring.For example, if sharing 10 evaluations about channel CH-A in certain forum；Then for this 10 evaluations, mention respectively
The evaluation keyword (such as " too disappointing ", " a lot of vacations ", " very true ") etc. in every evaluation in relation to creditworthiness is taken, and
The frequency of occurrence of each keyword is counted (for example, " disappointing " corresponding 5 evaluation informations, " good " to correspond in 1 evaluation information
Deng), according to keyword score corresponding to each keyword (for example, " disappointing " corresponding -10 points, " good " 10 points of correspondence, " one
As " it is 0 point corresponding etc.) and each keyword frequency of occurrence, calculate the channel scoring (for example, if " disappointing " occur 5 times,
" good " to occur 1 time, general to occur 4 times, then the scoring of the channel is (- 10) * 5+10*1+0*4=-40).
Step S240 provides feature, user characteristics, channel feature according to visitor, and objective standing ceases corresponding standard falseness etc.
Grade label, the falseness visitor that training constructs in advance, which provides, determines model, and the determining model of falseness visitor's money after train.
False visitor be built in the present embodiment in advance provide and determine model, after model parameter to be initiated, by visitor's money feature,
User characteristics and channel feature input falseness visitor's money and determine that model is trained.It in the training process, can be by visitor's money feature, use
Family feature and channel feature obtain the reasoning results after inputting the determining model of falseness visitor's money constructed in advance；And result by inference
With the comparison result of standard falseness grade label, calculating ratio is to error；Model training finally is carried out using preset algorithm, with minimum
Change and compares error；Falseness visitor's money corresponding when error is minimum value will finally be compared and determine model, the void after being determined as training
False visitor, which provides, determines model.Wherein, the present embodiment to specific training algorithm without limitation, such as in trained reasoning process,
Two convolutional layers and a full articulamentum can be constructed, and pass through sofamax layers of output reasoning based on the mode of convolutional neural networks
As a result；And when minimizing to ratio error, gradient descent algorithm also can be used, to minimize to ratio error, thus training for promotion
Efficiency.
Step S250 is provided using the false visitor after training when receiving new objective standing breath and is determined that model determines newly
The false grade of objective standing breath.
Specifically, the falseness visitor that the present embodiment training obtains, which provides, determines model accuracy of identification with higher, so as to standard
Really determine the false grade of new objective standing breath.Wherein, the false grade of objective standing breath include effectively, without demand, refuse to join
System and falseness etc..
Optionally, it can further be ceased according to the false grade of subsequent new objective standing breath and the new objective standing rear
Consumption in continuous preset time period, the falseness visitor after input training, which provides, determines model, thus to falseness visitor's money after training
Determine that model is further trained, further to promote the precision of prediction that false visitor provides determining model.
It can be seen that this programme is based on objective standing breath, and associated user data and channel data are ceased with objective standing, extracted
Visitor's money feature, user characteristics and channel feature out, and believed according to the various dimensions feature of visitor's money feature, user characteristics and channel feature
It ceases to train the falseness visitor constructed in advance to provide and determine model, model is determined to obtain final falseness visitor and provide, eventually by instruction
Falseness visitor after white silk, which provides, determines model to carry out the determination of the false grade of new objective standing breath, so as to avoid by manually knowing
Not at high cost, the low drawback of recognition efficiency of bring identification, and the accuracy of identification of false objective standing breath with higher；And
And visitor's money of classification, estimated consumption time and consumer budget belonging to the commodity of estimated consumption is specifically extracted in the present embodiment
Feature and user's history search characteristics, user's history transaction feature and user's Figure Characteristics user characteristics and it is default when
Between in section in the total number of the corresponding objective standing breath of the channel mark and preset time period effective objective standing breath of channel number
Purpose channel feature is conducive to false visitor's money and determines that model carries out comprehensive judge to new objective standing breath, and then promotes void
False visitor provides the precision of prediction for determining model；And the present embodiment can get the fine-grained false grade of new objective standing breath, can
The concrete condition that objective standing breath is effectively grasped for businessman and platform, is conducive to promote businessman and platform interests.
Fig. 3 shows a kind of function knot of the determining device of the false objective standing breath provided according to an embodiment of the present invention
Structure schematic diagram.As shown in figure 3, the device includes: to obtain module 31,32 training module 33 of extraction module and determining module 34.
Module 31 is obtained, is suitable for obtaining objective standing breath, obtains and ceases associated user data with the objective standing, and obtain
Associated channel data are ceased with the objective standing；
Extraction module 32 is suitable for providing feature according to visitor's money information extraction visitor, extracts user based on the user data
Feature, and, channel feature is extracted based on the channel data；
Training module 33 is suitable for according to visitor's money feature, the user characteristics and the channel feature, the preparatory structure of training
The falseness visitor built, which provides, determines model, and the determining model of falseness visitor's money after train；
Determining module 34, suitable for being provided using the falseness visitor after training and determining that model is true when receiving new objective standing breath
The false grade of the fixed new objective standing breath.
Optionally, visitor's money feature includes at least one of following characteristics: creation time, user identifier, channel mark
Know, classification belonging to the commodity of estimated consumption, expect consumption time and consumer budget.
Optionally, the module that obtains is further adapted for: according to the user identifier for including in the objective standing breath, obtain with
The corresponding user data of the user identifier.
Optionally, the user characteristics include at least one of following characteristics:
User's history search characteristics, user's history transaction feature and user's Figure Characteristics.
Optionally, the user's history search characteristics include: the quotient for the target industry that user searches within a preset period of time
The searching times of product；
The user's history transaction feature includes: the number of the commodity for the target industry that user trades within a preset period of time
Mesh；
User's Figure Characteristics include: age, occupation, income and/or marital status；
Wherein, the target industry is the industry that the estimated commodity consumed are subordinate to.
Optionally, module 31 is obtained to be further adapted for: according to the channel mark for including in the objective standing breath, acquisition and institute
State the corresponding channel data of channel mark.
Optionally, the channel feature includes at least one of following characteristics:
The channel in the total number and preset time period that the corresponding objective standing of the channel mark ceases in preset time period
Identify the number of corresponding effective objective standing breath.
Optionally, device further include: label acquisition module (not shown) is suitable for obtaining the objective standing breath and corresponds to
Standard falseness grade label；
Training module 33 is further adapted for:
Corresponding standard is ceased according to visitor's money feature, the user characteristics, the channel feature and the objective standing
False grade label, the falseness visitor that training constructs in advance, which provides, determines model, and the determining model of falseness visitor's money after train.
Optionally, training module 33 is further adapted for: visitor's money feature, the user characteristics and the channel is special
Sign inputs the falseness visitor's money constructed in advance and determines model, obtains the reasoning results；
According to the comparison result of the reasoning results and the standard falseness grade label, calculating ratio is to error；
Model training is carried out using preset algorithm, to minimize the comparison error；
The determining model of falseness visitor's money corresponding when error is minimum value is compared by described, the falseness visitor after being determined as training
It provides and determines model.
Optionally, the preset algorithm is gradient descent algorithm；The falseness visitor constructed in advance, which provides, determines model for volume
Product neural network model.
Wherein, the specific implementation process of each module can refer in embodiment of the method shown in Fig. 1 and/or Fig. 2 in the present embodiment
The description of corresponding portion, this will not be repeated here for the present embodiment.
It can be seen that this programme is based on objective standing breath, and associated user data and channel data are ceased with objective standing, extracted
Visitor's money feature, user characteristics and channel feature out, and believed according to the various dimensions feature of visitor's money feature, user characteristics and channel feature
It ceases to train the falseness visitor constructed in advance to provide and determine model, model is determined to obtain final falseness visitor and provide, eventually by instruction
Falseness visitor after white silk, which provides, determines model to carry out the determination of the false grade of new objective standing breath, so as to avoid by manually knowing
Not at high cost, the low drawback of recognition efficiency of bring identification, and the accuracy of identification of false objective standing breath with higher.
A kind of nonvolatile computer storage media is provided according to an embodiment of the present invention, and the computer storage is situated between
Matter is stored with an at least executable instruction, which can be performed the falseness visitor in above-mentioned any means embodiment
The determination method of standing breath.
Fig. 4 shows a kind of structural schematic diagram of the calculating equipment provided according to an embodiment of the present invention, present invention tool
Body embodiment does not limit the specific implementation for calculating equipment.
As shown in figure 4, the calculating equipment may include: processor (processor) 402, communication interface
(Communications Interface) 404, memory (memory) 406 and communication bus 408.
Wherein:
Processor 402, communication interface 404 and memory 406 complete mutual communication by communication bus 408.
Communication interface 404, for being communicated with the network element of other equipment such as client or other servers etc..
Processor 402, for executing program 410, the determination method that can specifically execute above-mentioned false objective standing breath is implemented
Correlation step in example.
Specifically, program 410 may include program code, which includes computer operation instruction.
Processor 402 may be central processor CPU or specific integrated circuit ASIC (Application
Specific Integrated Circuit), or be arranged to implement the integrated electricity of one or more of the embodiment of the present invention
Road.The one or more processors that equipment includes are calculated, can be same type of processor, such as one or more CPU；It can also
To be different types of processor, such as one or more CPU and one or more ASIC.
Memory 406, for storing program 410.Memory 406 may include high speed RAM memory, it is also possible to further include
Nonvolatile memory (non-volatile memory), for example, at least a magnetic disk storage.
Program 410 specifically can be used for so that processor 402 executes following operation:
Objective standing breath is obtained, and feature is provided according to visitor's money information extraction visitor；
It obtains and ceases associated user data with the objective standing, user characteristics are extracted based on the user data；And it obtains
It takes and ceases associated channel data with the objective standing, extract channel feature based on the channel data；
According to visitor's money feature, the user characteristics and the channel feature, falseness visitor's money that training constructs in advance is true
Cover half type, and the falseness visitor after being trained provides and determines model；
When receiving new objective standing breath, is provided using the falseness visitor after training and determine that model determines the new objective standing
The false grade of breath.
In a kind of optional embodiment, visitor's money feature includes at least one of following characteristics:
Creation time, user identifier, channel mark, estimated consumption commodity belonging to classification, estimated consumption time and
Consumer budget.
In a kind of optional embodiment, program 410 specifically can be used for so that processor 402 executes following operation:
According to the user identifier for including in the objective standing breath, user data corresponding with the user identifier is obtained.
In a kind of optional embodiment, the user characteristics include at least one of following characteristics:
User's history search characteristics, user's history transaction feature and user's Figure Characteristics.
In a kind of optional embodiment, the user's history search characteristics include: that user searches within a preset period of time
The searching times of the commodity of the target industry of rope；
The user's history transaction feature includes: the number of the commodity for the target industry that user trades within a preset period of time
Mesh；
User's Figure Characteristics include: age, occupation, income and/or marital status；
Wherein, the target industry is the industry that the estimated commodity consumed are subordinate to.
In a kind of optional embodiment, program 410 specifically can be used for so that processor 402 executes following operation:
According to the channel mark for including in the objective standing breath, channel data corresponding with the channel mark are obtained.
In a kind of optional embodiment, the channel feature includes at least one of following characteristics:
The channel in the total number and preset time period that the corresponding objective standing of the channel mark ceases in preset time period
Identify the number of corresponding effective objective standing breath.
In a kind of optional embodiment, program 410 specifically can be used for so that processor 402 executes following operation:
It obtains the objective standing and ceases corresponding standard falseness grade label；
It is described objective according to visitor's money feature, the user characteristics and the channel feature, the falseness that training constructs in advance
It provides and determines model, and the determining model of falseness visitor's money after train further comprises:
Corresponding standard is ceased according to visitor's money feature, the user characteristics, the channel feature and the objective standing
False grade label, the falseness visitor that training constructs in advance, which provides, determines model, and the determining model of falseness visitor's money after train.
In a kind of optional embodiment, program 410 specifically can be used for so that processor 402 executes following operation:
By visitor's money feature, the user characteristics and the channel feature, inputs the falseness visitor's money constructed in advance and determine
Model obtains the reasoning results；
According to the comparison result of the reasoning results and the standard falseness grade label, calculating ratio is to error；
Model training is carried out using preset algorithm, to minimize the comparison error；
The determining model of falseness visitor's money corresponding when error is minimum value is compared by described, the falseness visitor after being determined as training
It provides and determines model.
In a kind of optional embodiment, the preset algorithm is gradient descent algorithm；The falseness constructed in advance
Visitor, which provides, determines that model is convolutional neural networks model.
Algorithm and display are not inherently related to any particular computer, virtual system, or other device provided herein.
Various general-purpose systems can also be used together with teachings based herein.As described above, it constructs required by this kind of system
Structure be obvious.In addition, the present invention is also not directed to any particular programming language.It should be understood that can use various
Programming language realizes summary of the invention described herein, and the description done above to language-specific is to disclose this hair
Bright preferred forms.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this specification.
Similarly, it should be understood that in order to simplify the disclosure and help to understand one or more of the various inventive aspects,
Above in the description of exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes
In example, figure or descriptions thereof.However, the disclosed method should not be interpreted as reflecting the following intention: i.e. required to protect
Shield the present invention claims features more more than feature expressly recited in each claim.More precisely, as following
Claims reflect as, inventive aspect is all features less than single embodiment disclosed above.Therefore,
Thus the claims for following specific embodiment are expressly incorporated in the specific embodiment, wherein each claim itself
All as a separate embodiment of the present invention.
Those skilled in the art will understand that can be carried out adaptively to the module in the equipment in embodiment
Change and they are arranged in one or more devices different from this embodiment.It can be the module or list in embodiment
Member or component are combined into a module or unit or component, and furthermore they can be divided into multiple submodule or subelement or
Sub-component.Other than such feature and/or at least some of process or unit exclude each other, it can use any
Combination is to all features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed
All process or units of what method or apparatus are combined.Unless expressly stated otherwise, this specification is (including adjoint power
Benefit require, abstract and attached drawing) disclosed in each feature can carry out generation with an alternative feature that provides the same, equivalent, or similar purpose
It replaces.
In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments
In included certain features rather than other feature, but the combination of the feature of different embodiments mean it is of the invention
Within the scope of and form different embodiments.For example, in detail in the claims, embodiment claimed it is one of any
Can in any combination mode come using.
Various component embodiments of the invention can be implemented in hardware, or to run on one or more processors
Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice
Microprocessor or digital signal processor (DSP) realize the determining device of false objective standing breath according to embodiments of the present invention
In some or all components some or all functions.The present invention is also implemented as described herein for executing
Some or all device or device programs (for example, computer program and computer program product) of method.In this way
Realization program of the invention can store on a computer-readable medium, or can have the shape of one or more signal
Formula.Such signal can be downloaded from an internet website to obtain, and perhaps be provided on the carrier signal or with any other shape
Formula provides.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability
Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real
It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch
To embody.The use of word first, second, and third does not indicate any sequence.These words can be explained and be run after fame
Claim.
Claims (10)
1. a kind of determination method of false objective standing breath, comprising:
Objective standing breath is obtained, and feature is provided according to visitor's money information extraction visitor；
It obtains and ceases associated user data with the objective standing, user characteristics are extracted based on the user data；And obtain with
Visitor's standing ceases associated channel data, extracts channel feature based on the channel data；
According to visitor's money feature, the user characteristics and the channel feature, the falseness visitor that training constructs in advance, which provides, determines mould
Type, and the falseness visitor after being trained provides and determines model；
When receiving new objective standing breath, is provided using the falseness visitor after training and determine that model determines the new objective standing breath
False grade.
2. according to the method described in claim 1, wherein, visitor's money feature includes at least one of following characteristics:
Creation time, channel mark, expects classification belonging to the commodity of consumption, estimated consumption time and consumption at user identifier
Budget.
3. according to the method described in claim 2, wherein, the acquisition and the associated user data of the objective standing breath are further
Include:
According to the user identifier for including in the objective standing breath, user data corresponding with the user identifier is obtained.
4. according to the method described in claim 3, wherein, the user characteristics include at least one of following characteristics:
User's history search characteristics, user's history transaction feature and user's Figure Characteristics.
5. according to the method described in claim 4, wherein, the user's history search characteristics include: user in preset time period
The searching times of the commodity of the target industry of interior search；
The user's history transaction feature includes: the number of the commodity for the target industry that user trades within a preset period of time；
User's Figure Characteristics include: age, occupation, income and/or marital status；
Wherein, the target industry is the industry that the estimated commodity consumed are subordinate to.
6. according to the method described in claim 2, wherein, the acquisition and the associated channel data of the objective standing breath are further
Include:
According to the channel mark for including in the objective standing breath, channel data corresponding with the channel mark are obtained.
7. according to the method described in claim 6, wherein, the channel feature includes at least one of following characteristics:
The channel mark in the total number and preset time period that the corresponding objective standing of the channel mark ceases in preset time period
The number of corresponding effective objective standing breath.
8. a kind of determining device of false objective standing breath, comprising:
Obtain module, be suitable for obtaining objective standing breath, obtain and cease associated user data with the objective standing, and obtain with it is described
Objective standing ceases associated channel data；
Extraction module is suitable for providing feature according to visitor's money information extraction visitor, extracts user characteristics based on the user data, with
And channel feature is extracted based on the channel data；
Training module is suitable for according to visitor's money feature, the user characteristics and the channel feature, the void that training constructs in advance
False visitor, which provides, determines model, and the determining model of falseness visitor's money after train；
Determining module, described in when receiving new objective standing breath, being provided using the falseness visitor after training and determining that model determines
The false grade of new objective standing breath.
9. a kind of calculating equipment, comprising: processor, memory, communication interface and communication bus, the processor, the storage
Device and the communication interface complete mutual communication by the communication bus；
The memory executes the processor as right is wanted for storing an at least executable instruction, the executable instruction
Ask the corresponding operation of determination method of the objective standing breath of falseness described in any one of 1-7.
10. a kind of computer storage medium, an at least executable instruction, the executable instruction are stored in the storage medium
Processor is set to execute the corresponding operation of determination method such as false objective standing breath of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910276821.0A CN110111131A (en) | 2019-04-08 | 2019-04-08 | The determination method and device of false visitor's standing breath |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910276821.0A CN110111131A (en) | 2019-04-08 | 2019-04-08 | The determination method and device of false visitor's standing breath |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110111131A true CN110111131A (en) | 2019-08-09 |
Family
ID=67485318
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910276821.0A Pending CN110111131A (en) | 2019-04-08 | 2019-04-08 | The determination method and device of false visitor's standing breath |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110111131A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110782342A (en) * | 2019-10-29 | 2020-02-11 | 北京明略软件系统有限公司 | Method and device for verifying correctness of new channel feature engineering based on binary classification model |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106355420A (en) * | 2016-08-30 | 2017-01-25 | 江苏车置宝信息科技股份有限公司 | Customer data quality identification and automatic order distribution system |
| CN107292634A (en) * | 2017-06-23 | 2017-10-24 | 浙江极赢信息技术有限公司 | A kind of user's appraisal procedure and system towards telemarketing |
| CN108470253A (en) * | 2018-04-02 | 2018-08-31 | 腾讯科技（深圳）有限公司 | A kind of user identification method, device and storage device |
| CN109063433A (en) * | 2018-07-09 | 2018-12-21 | 中国联合网络通信集团有限公司 | Recognition methods, device and the readable storage medium storing program for executing of fictitious users |
| CN109583651A (en) * | 2018-12-03 | 2019-04-05 | 焦点科技股份有限公司 | A kind of method and apparatus for insuring electric business platform user attrition prediction |
| CN109767327A (en) * | 2018-12-20 | 2019-05-17 | 平安科技（深圳）有限公司 | Customer information acquisition and its application method based on anti money washing |
-
2019
- 2019-04-08 CN CN201910276821.0A patent/CN110111131A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106355420A (en) * | 2016-08-30 | 2017-01-25 | 江苏车置宝信息科技股份有限公司 | Customer data quality identification and automatic order distribution system |
| CN107292634A (en) * | 2017-06-23 | 2017-10-24 | 浙江极赢信息技术有限公司 | A kind of user's appraisal procedure and system towards telemarketing |
| CN108470253A (en) * | 2018-04-02 | 2018-08-31 | 腾讯科技（深圳）有限公司 | A kind of user identification method, device and storage device |
| CN109063433A (en) * | 2018-07-09 | 2018-12-21 | 中国联合网络通信集团有限公司 | Recognition methods, device and the readable storage medium storing program for executing of fictitious users |
| CN109583651A (en) * | 2018-12-03 | 2019-04-05 | 焦点科技股份有限公司 | A kind of method and apparatus for insuring electric business platform user attrition prediction |
| CN109767327A (en) * | 2018-12-20 | 2019-05-17 | 平安科技（深圳）有限公司 | Customer information acquisition and its application method based on anti money washing |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110782342A (en) * | 2019-10-29 | 2020-02-11 | 北京明略软件系统有限公司 | Method and device for verifying correctness of new channel feature engineering based on binary classification model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108363821A (en) | A kind of information-pushing method, device, terminal device and storage medium | |
| CN107590688A (en) | The recognition methods of target customer and terminal device | |
| CN109299994A (en) | Recommended method, device, equipment and readable storage medium storing program for executing | |
| CN110163647A (en) | A kind of data processing method and device | |
| CN105931068A (en) | Cardholder consumption figure generation method and device | |
| CN110263235A (en) | Information pushes object updating method, device and computer equipment | |
| CN107153656A (en) | A kind of information search method and device | |
| CN106844407A (en) | Label network production method and system based on data set correlation | |
| CN110334356A (en) | Article matter method for determination of amount, article screening technique and corresponding device | |
| CN110009430A (en) | Cheating user's detection method, electronic equipment and computer readable storage medium | |
| CN112116168B (en) | User behavior prediction method and device and electronic equipment | |
| CA3059929A1 (en) | Text searching method, apparatus, and non-transitory computer-readable storage medium | |
| CN102402717A (en) | Data analysis facility and method | |
| CN108363730A (en) | A kind of content recommendation method, system and terminal device | |
| CN107832444A (en) | Event based on search daily record finds method and device | |
| CN114240495A (en) | Method and device for predicting business opportunity conversion probability and computer readable storage medium | |
| CN105989066A (en) | Information processing method and device | |
| CN104077288B (en) | Web page contents recommend method and web page contents recommendation apparatus | |
| CN109978624A (en) | Information processing method, electronic equipment and computer readable storage medium | |
| CN111325255B (en) | Specific crowd delineating method and device, electronic equipment and storage medium | |
| CN110111131A (en) | The determination method and device of false visitor's standing breath | |
| CN110717095B (en) | Service item pushing method and device | |
| CN109033078B (en) | The recognition methods of sentence classification and device, storage medium, processor | |
| CN107944589A (en) | The Forecasting Methodology and prediction meanss of ad click rate | |
| CN110389963A (en) | The recognition methods of channel effect, device, equipment and storage medium based on big data |
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 |
Application publication date: 20190809 |
|
| RJ01 | Rejection of invention patent application after publication |