CN110533453A - Based on the matched Products Show method, apparatus of user, computer equipment - Google Patents
Based on the matched Products Show method, apparatus of user, computer equipment Download PDFInfo
- Publication number
- CN110533453A CN110533453A CN201910662549.XA CN201910662549A CN110533453A CN 110533453 A CN110533453 A CN 110533453A CN 201910662549 A CN201910662549 A CN 201910662549A CN 110533453 A CN110533453 A CN 110533453A
- Authority
- CN
- China
- Prior art keywords
- client
- customers
- information
- value
- target
- 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
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000012549 training Methods 0.000 claims abstract description 75
- 238000012216 screening Methods 0.000 claims abstract description 34
- 230000009466 transformation Effects 0.000 claims description 21
- 230000006870 function Effects 0.000 claims description 20
- 238000010276 construction Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 15
- 238000005516 engineering process Methods 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 14
- 238000003064 k means clustering Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009182 swimming Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- 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
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0255—Targeted advertisements based on user history
-
- 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
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0269—Targeted advertisements based on user profile or attribute
- G06Q30/0271—Personalized advertisement
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses be based on the matched Products Show method, apparatus of user, computer equipment.Method includes: to be clustered to obtain customers' information to client included in client information table according to clustering rule, client's Matching Model is constructed according to customers' information and full connection hidden layer, client's Matching Model is trained according to training parameter, target customers corresponding with client to be recommended are obtained according to client's Matching Model after training, the product information in client information table is screened to obtain target product and be pushed according to target product screening rule and target customers.The present invention is based on prediction model technologies, it can be in the case where user buy our company's product, the user group to match with the user is obtained based on the matched mode of user, and it further obtains target product corresponding with the user group and recommends to the client, with more wide applicability, good technical effect is achieved in actual application.
Description
Technical field
The present invention relates to field of computer technology, more particularly to it is a kind of based on the matched Products Show method, apparatus of user,
Computer equipment and storage medium.
Background technique
Enterprise is to promote to itself product, the specifying information of product can be pushed to user, current use is more wide
General technology are as follows: the client is calculated does not score product by the scored score data of product of score in predicting model and client
Prediction scoring, and by prediction scoring never scoring product in selected section product push to user.However in the prior art
Products Show method be only capable of recommending corresponding target product to user in the case where user has purchased portioned product, work as user
Then without normal direction, it recommends target product in the case where purchase our company's product, therefore existing Products Show method is in reality
There are limitations in use process, are unable to satisfy the actual demand of enterprise.Thus, existing technical method is produced to user
There is use when recommending by limitation in product.
Summary of the invention
The embodiment of the invention provides one kind based on the matched Products Show method, apparatus of user, computer equipment and to deposit
Storage media, it is intended to solve the problems, such as to have use in art methods when carrying out Products Show to user by limitation.
In a first aspect, the embodiment of the invention provides one kind to be based on the matched Products Show method of user comprising:
If receiving the client information table that administrator is inputted, according to preset clustering rule to institute in the client information table
The client for including is clustered to obtain customers' information, wherein further includes what each client bought in the client information table
Product information;
According to customers' information and preset full client of the connection hidden layer building comprising input node and output node
With model;
Constructed client's Matching Model is trained according to preset training parameter and is matched with the client after being trained
Model;
If receiving client to be recommended passes through the customer information to be recommended that user terminal is inputted, by the client to be recommended
Client's Matching Model after information input training, and obtain in client's Matching Model the matching degree between the client to be recommended
Highest customers are as target customers;
According to preset target product screening rule and the target customers to the product information in the client information table
It is screened to obtain target product and push to the user terminal.
Second aspect, the embodiment of the invention provides one kind to be based on the matched Products Show device of user comprising:
Cluster cell, if the client information table inputted for receiving administrator, according to preset clustering rule to described
Client included in client information table is clustered to obtain customers' information, wherein further includes in the client information table
The product information of each client's purchase;
Client's Matching Model construction unit, for including defeated according to customers' information and preset full connection hidden layer building
Client's Matching Model of ingress and output node;
Client's Matching Model training unit, for being instructed according to preset training parameter to constructed client's Matching Model
Practice with client's Matching Model after being trained;
Target customers' matching unit, if passing through the visitor to be recommended that user terminal is inputted for receiving client to be recommended
Family information by client's Matching Model after the customer information input training to be recommended, and obtains in client's Matching Model
The highest customers of matching degree are as target customers between the client to be recommended;
Product information screening unit is used for according to preset target product screening rule and the target customers to the visitor
Product information in the information table of family is screened to obtain target product and push to the user terminal.
The third aspect, the embodiment of the present invention provide a kind of computer equipment again comprising memory, processor and storage
On the memory and the computer program that can run on the processor, the processor execute the computer program
Based on the matched Products Show method of user described in the above-mentioned first aspect of Shi Shixian.
Fourth aspect, the embodiment of the invention also provides a kind of computer readable storage mediums, wherein the computer can
It reads storage medium and is stored with computer program, it is above-mentioned that the computer program when being executed by a processor executes the processor
Based on the matched Products Show method of user described in first aspect.
The embodiment of the invention provides one kind based on the matched Products Show method, apparatus of user, computer equipment and to deposit
Storage media.Client included in client information table is clustered according to clustering rule to obtain customers' information, according to client
Group's information and full connection hidden layer construct client's Matching Model, are trained according to training parameter to client's Matching Model, according to instruction
Client's Matching Model after white silk obtains target customers corresponding with client to be recommended, according to target product screening rule and target
Customers are screened to obtain target product and be pushed to the product information in client information table.It, can by the above method
In the case where user does not buy our company's product, the user group to match with the user is obtained based on the matched mode of user,
And further obtain target product corresponding with the user group and recommend to the client, there is more wide applicability, in reality
Good technical effect is achieved in application process.
Detailed description of the invention
Technical solution in order to illustrate the embodiments of the present invention more clearly, below will be to needed in embodiment description
Attached drawing is briefly described, it should be apparent that, drawings in the following description are some embodiments of the invention, general for this field
For logical technical staff, without creative efforts, it is also possible to obtain other drawings based on these drawings.
Fig. 1 is the flow diagram provided in an embodiment of the present invention based on the matched Products Show method of user;
Fig. 2 is the application scenarios schematic diagram provided in an embodiment of the present invention based on the matched Products Show method of user;
Fig. 3 is the sub-process schematic diagram provided in an embodiment of the present invention based on the matched Products Show method of user;
Fig. 4 is another sub-process schematic diagram provided in an embodiment of the present invention based on the matched Products Show method of user;
Fig. 5 is another sub-process schematic diagram provided in an embodiment of the present invention based on the matched Products Show method of user;
Fig. 6 is another sub-process schematic diagram provided in an embodiment of the present invention based on the matched Products Show method of user;
Fig. 7 is another flow diagram provided in an embodiment of the present invention based on the matched Products Show method of user;
Fig. 8 is the schematic block diagram provided in an embodiment of the present invention based on the matched Products Show device of user;
Fig. 9 is the schematic block diagram of computer equipment provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Based on this hair
Embodiment in bright, every other implementation obtained by those of ordinary skill in the art without making creative efforts
Example, shall fall within the protection scope of the present invention.
It should be appreciated that ought use in this specification and in the appended claims, term " includes " and "comprising" instruction
Described feature, entirety, step, operation, the presence of element and/or component, but one or more of the other feature, whole is not precluded
Body, step, operation, the presence or addition of element, component and/or its set.
It is also understood that mesh of the term used in this description of the invention merely for the sake of description specific embodiment
And be not intended to limit the present invention.As description of the invention and it is used in the attached claims, unless on
Other situations are hereafter clearly indicated, otherwise " one " of singular, "one" and "the" are intended to include plural form.
It will be further appreciated that the term "and/or" used in description of the invention and the appended claims is
Refer to any combination and all possible combinations of one or more of associated item listed, and including these combinations.
Referring to Fig. 1, Fig. 1 is the process signal provided in an embodiment of the present invention based on the matched Products Show method of user
Figure;Fig. 2 is the application scenarios schematic diagram provided in an embodiment of the present invention based on the matched Products Show method of user.It should be based on use
The matched Products Show method in family is applied in management server 10, and this method passes through the application that is installed in management server 10
Software is executed, and user terminal 20 is by establishing the transmission being connected to the network to realize data information with management server 10.Pipe
Reason server 10 is for executing based on the matched Products Show method of user so that target product is pushed to client to be recommended
Enterprise terminal, user terminal 20 is the terminal device for sending data information to management server 10, such as desktop
Brain, laptop, tablet computer or mobile phone etc..Only illustrated in Fig. 2 management server 10 and a user terminal 20 into
Row information transmission, in practical applications, the management server 10 can also carry out information transmission with more user terminals 20.
As shown in Figure 1, the method comprising the steps of S110~S150.
If S110, the client information table that administrator is inputted is received, according to preset clustering rule to the customer information
Client included in table is clustered to obtain customers' information.
If receiving the client information table that administrator is inputted, according to preset clustering rule to institute in the client information table
The client for including is clustered to obtain customers' information, wherein in client information table comprising multiple clients customer information with
And the product information of each client's purchase.Administrator is the user of management server, and clustering rule is for all
Client carries out the Rule Information of clustering, after being clustered based on customer information to client can will similar client segmentation to one
Customers' information comprising multiple customers can be obtained in customers after clustering to all clients.
For example, certain a client information table is as shown in table 1.
Customer name | Age | Occupation | Hobby | Product information |
Client 1 | 28 | Lawyer | Running | Product 1, product 3 |
Client 2 | 36 | Civil servant | Tourism | Product 2 |
Client 3 | 22 | Student | Swimming | Product 1 |
Client 4 | 31 | Self-employed worker | Running | Product 3, product 4 |
Table 1
In one embodiment, as shown in figure 3, step S110 includes sub-step S111 and S112.
S111, according to the characteristic variable transformation rule in the clustering rule by visitor included in the client information table
Family information is converted to characteristic variable.
Client included in the client information table is believed according to the characteristic variable transformation rule in the clustering rule
Breath is converted to characteristic variable.Specifically, including the information such as customer name, age, occupation, hobby, feature in customer information
Variable transformation rule is the Rule Information for the customer information of each client to be converted to characteristic variable, every in customer information
One information can be converted to corresponding vector value by characteristic variable transformation rule and is indicated, then can be by each customer information
Corresponding conversion is the feature vector of a multidimensional, that is to say characteristic variable.Each client correspondence possesses phase in client information table
The customer information answered can then obtain the customer information of all clients in client information table, pass through characteristic variable transformation rule
It obtains the customer information of all clients and is converted to corresponding characteristic variable, characteristic variable transformation rule is for believing client
Breath is converted to the rule of characteristic variable, in customer information each single item information with a vector value in characteristic variable transformation rule
It is corresponding.
For example, several information for including in the customer information of first place client in table 1 are respectively: name " client 1 ", age
" 28 ", professional " lawyer ", hobby " running ", the vector value of this information of age is this in characteristic variable transformation rule
The numerical value of information, it is 5 that this professional information, which be " lawyer " corresponding vector value, hobby this be " running " correspondence
Vector value be " 12 ", then be converted to corresponding characteristic variable be D={ 28,5,12 }.
S112, according in the clustering rule clustering algorithm and the characteristic variable client is clustered to be wrapped
Customers' information containing multiple customers.
According in the clustering rule clustering algorithm and the characteristic variable client is clustered to obtain comprising more
Customers' information of a customers, the characteristic variable in customers' information comprising multiple customers and each customers' mass center
Value presets final required monoid specifically, clustering algorithm is K-means clustering algorithm in K-means clustering algorithm
Quantity k, specifically, the corresponding characteristic variable of each customer information is indicated using a point coordinate based on three-dimensional system of coordinate,
All three-dimensional coordinate points are clustered according to the numerical value of k, obtain the three-dimensional coordinate of k customers and each customers' mass center
Value, the D coordinates value of mass center that is to say the characteristic variable value of mass center.
For example, K=15 is arranged in K-means clustering algorithm, then 15 customers and each customers are finally obtained
The characteristic variable value of mass center, mass center are the central point of the customers, and the characteristic variable value of mass center is to own in the customers
The mean value of characteristic variable.
S120, the visitor comprising input node and output node is constructed according to customers' information and preset full connection hidden layer
Family Matching Model.
According to customers' information and preset full client of the connection hidden layer building comprising input node and output node
With model.Customers' information is by the obtained information comprising multiple customers of clustering algorithm, each customers Zhong Bao
Customer information and the customers containing multiple clients correspond to the characteristic information of mass center.It include multiple inputs in client's Matching Model
Node and multiple output nodes, wherein the corresponding input node of each of client characteristics variable dimension, each client
The corresponding output node of group, each output node one output node value of corresponding output, output node value can indicate a certain
Matching degree between customers corresponding to client and the output node.Due to passing through the spy of the obtained each client of transformation rule
Number of dimensions included in sign variable is all the same, therefore can pass through the corresponding input for generating identical quantity of the dimension of characteristic variable
Node passes through the corresponding output node for generating identical quantity of the quantity of customers in customers' information.Full connection hidden layer is use
In the middle layer contacted input node and output node, complete connect includes several feature units in hidden layer, each
Feature unit is associated with all input nodes and all output nodes, and feature unit can be used for reflecting every in customer information
The quantity of one relationship between information and customers, feature unit can be set according to customers' information, preferred feature
The quantity of unit may be set to the 1/3-1/2 of customers' quantity.
For example, the quantity of customers is 15 in customers' information, then the quantity of feature unit can be set as 7.
In one embodiment, as shown in figure 4, step S120 includes sub-step S121, S122, S123 and S124.
S121, the input node that client's Matching Model is constructed according to the dimension of client characteristics variable in customers' information.
According to the input node of dimension building client's Matching Model of client characteristics variable in customers' information, specifically,
Each of client characteristics variable dimension corresponds to an input node, therefore can generate phase by the way that the dimension of characteristic variable is corresponding
With the input node of quantity, the dimension and input node that characteristic variable is included are corresponded.
S122, the output node that client's Matching Model is constructed according to the quantity of customers in customers' information.
According to the output node of customers' information architecture client's Matching Model, specifically, each customers is one corresponding
Output node, therefore the corresponding output node for generating identical quantity of quantity of customers in customers' information, customers can be passed through
Customers and output node in information correspond.
S123, according to it is preset it is full connection hidden layer in feature unit and constructed input node, with input node value work
The first formula group of input node to feature unit is constructed as output valve for input value, feature unit value.
According to the preset complete multiple feature units connected in hidden layer and constructed multiple input nodes, with input node value
The first formula group of input node to feature unit is constructed as output valve as input value, feature unit value, wherein first is public
Formula group includes formula of all input nodes to all feature units.Input node is in client's Matching Model for a certain
The node that the characteristic variable of a client is inputted, the specific value of input node are input node value, that is to say the client
The vector value of respective dimensions in characteristic variable, due to a dimension in each input node character pair variable, all inputs
The i.e. corresponding client characteristics variable of nodal value, feature unit value are the calculated value of the feature unit in full connection hidden layer.
For example, the input node value of a certain input node is x1, the feature unit value of a certain feature unit is y1, then this is defeated
The formula of ingress to this feature unit is y1=a × x1+b;Wherein, a and b is the parameter in the formula, the parameter value in formula
For the number generated at random.
S124, according to it is preset it is full connection hidden layer in feature unit and constructed output node, with feature unit value work
The second formula group for input value, output node value as output valve construction feature unit to output node.
According to the multiple feature units and constructed multiple output nodes in preset full connection hidden layer, with feature unit value
The second formula group as input value, output node value as output valve construction feature unit to output node, wherein second is public
Formula group includes formula of all feature units to all output nodes.Output node is in client's Matching Model for client
The node that matching degree between respective client group is exported, the specific value of output node are output node value, output
Nodal value can indicate that the matching degree between customers corresponding to a certain client and the output node, feature unit value are to connect entirely
Connect the calculated value of the feature unit in hidden layer.
For example, the feature unit value of a certain feature unit is y1, the output node value of a certain output node is z1, then the spy
The formula for levying unit to the output node is z1=c × y1+d;Wherein, c and d is the parameter in the formula, the parameter value in formula
For the number generated at random.
S130, constructed client's Matching Model is trained according to preset training parameter with the client after being trained
Matching Model.
Constructed client's Matching Model is trained according to preset training parameter and is matched with the client after being trained
Model.For the accuracy for improving client's Matching Model, constructed client's Matching Model can be carried out by preset training parameter
Training, can be obtained the client's Matching Model trained.Characteristic variable and each visitor in training parameter including multiple clients
The matching degree at family and corresponding target customers, target customers are that matching degree is highest between the client in customers' information
One customers.
In one embodiment, as shown in figure 5, step S130 includes sub-step S131, S132, S133 and S134.
S131, the characteristic variable for obtaining a client in preset training parameter, according to respective formula in the first formula group and
The feature unit value of all feature units in full connection hidden layer is calculated in input node value.
The characteristic variable for obtaining a client in preset training parameter is saved according to respective formula in the first formula group and input
The feature unit value of all feature units in full connection hidden layer is calculated in point value.Due to every dimension in the characteristic variable of client
An input node is corresponded to, therefore using vector value included in characteristic variable as the input node of corresponding input node
The characteristic variable of the user can be inputted client's Matching Model by value.Each feature unit is and institute in client's Matching Model
There are input node and all output nodes to be associated, that is to say through respective formula and input node value in the first formula group i.e.
The feature unit value of all feature units in full connection hidden layer can be calculated.The calculated result of respective formula in first formula group
In a certain feature unit include multiple calculated values, then take the average value of multiple calculated values as the feature unit of this feature unit
Value.
S132, the output section that corresponding output node is calculated according to respective formula and feature unit value in the second formula group
Point value.
The output node value of corresponding output node is calculated according to respective formula and feature unit value in the second formula group,
The customers as included in customers' information and output node correspond, and pass through respective formula and spy in the second formula group
Output node value can be calculated in sign cell value, and output node value can indicate visitor corresponding to a certain client and the output node
Matching degree between the group of family, that is to say the matching degree being calculated between existing customer and each customers.Since client matches
The parameter of formula in model is also adjusted update without training, therefore the output node value being currently calculated is only capable of using
It is trained in client's Matching Model, and can not be as final result to match client with corresponding customers.
S133, the matching degree for obtaining existing customer and corresponding target customers and target corresponding with the target customers are defeated
Egress value is calculated according to matching degree and target output node value of the mean square error function to target customers to be worked as
The square mean error amount of preceding client.
The matching degree and target corresponding with target customers output for obtaining existing customer and corresponding target customers save
Point value, the matching degree of each client and corresponding target customers in training parameter, target customers be in customers' information with
The highest customers of matching degree between the client, then can obtain the client for the matching degree of target customers, and obtain
An output node value corresponding with the target customers is as target output node value in all output node values.According to square
Error function calculates the matching degree and target output node value of target customers, to obtain the mean square error of existing customer
Value, mean square error function (MSE function) is a kind of calculating letter for reflecting difference degree between estimator and the amount of being estimated
Number can be calculated the matching degree of target customers and client's Matching Model in training parameter by mean square error function and calculate
The obtained error amount between target output node value, the square mean error amount the big, shows client's target in client's training parameter
The error between target output node value that the matching degree and client's Matching Model of customers are calculated is bigger, square mean error amount
It is smaller, show client's target that the matching degree of target customers and client's Matching Model are calculated in client's training parameter
Error between output node value is smaller.
S134, judged according to preset mean square error threshold value whether square mean error amount is less than mean square error threshold value, root
It is judged that result and the parameter regulation coefficient are adjusted update to the parameter of formula in client's Matching Model.
Judged according to preset mean square error threshold value whether square mean error amount is less than mean square error threshold value, according to judgement
As a result and the parameter regulation coefficient is adjusted update to the parameter of formula in client's Matching Model, specifically, parameter adjusts
It include parameter adjustment direction and parameter adjustment magnitude in coefficient, parameter adjustment direction includes positive adjustment and negative sense adjustment, parameter
Adjustment amplitude is the specific range value being adjusted, and carries out what primary adjustment updated to the parameter of formula in client's Matching Model
If specific steps include: that judging result is square mean error amount no more than preset mean square error threshold value, according to parameter adjustment direction
In positive adjustment and the range value in parameter adjustment magnitude update is adjusted to the parameter of formula in client's Matching Model;If
Judging result is that square mean error amount is greater than preset mean square error threshold value, then according to the reversed adjustment and parameter in parameter adjustment direction
Range value in adjustment amplitude is adjusted update to the parameter of formula in client's Matching Model.It completes in client's Matching Model
After the parameter of formula carries out primary adjustment update, the characteristic variable of next client and the client couple in training parameter can be obtained
It answers the matching degree of target customers to be iterated training to client's Matching Model, passes through the feature of clients multiple in training parameter
Variable and each client correspond to after the matching degree of target customers is iterated training to client's Matching Model, can be obtained with
Trained client's Matching Model.
For example, the range value in parameter adjustment magnitude is 2%, judging result is that square mean error amount is not more than preset mean square error
Poor threshold value, then this adjustment need to carry out positive adjustment, the parameter original numerical value basis of this adjustment formula in client's Matching Model
On multiplied by 1.02 obtain new parameter value.
If S140, the customer information to be recommended that client to be recommended is inputted by user terminal is received, by described wait push away
Client's Matching Model after recommending customer information input training, and obtain in client's Matching Model between the client to be recommended
The highest customers of matching degree are as target customers.
If receiving client to be recommended passes through the customer information to be recommended that user terminal is inputted, by the client to be recommended
Client's Matching Model after information input training, and obtain in client's Matching Model the matching degree between the client to be recommended
Highest customers are as target customers.Wherein, client to be recommended is the user of user terminal, customer information to be recommended
In include client to be recommended characteristic variable.
In one embodiment, as shown in fig. 6, step S140 includes sub-step S141, S142 and S143.
S141, input node value of the characteristic variable of client to be recommended as respective formula in the first formula group, root are obtained
The feature unit of all feature units in full connection hidden layer is calculated according to respective formula in the first formula group and input node value
Value.
Input node value of the characteristic variable of client to be recommended as respective formula in the first formula group is obtained, according to first
The feature unit value of all feature units in full connection hidden layer is calculated in respective formula and input node value in formula group.Due to
Every dimension corresponds to an input node in the characteristic variable of client to be recommended, therefore by vector included in characteristic variable
It is worth the input node value as corresponding input node, the characteristic variable of the user can be inputted client's Matching Model.Client
It is associated, that is to say through the first public affairs with all input nodes and all output nodes with each feature unit in model
The feature unit value of all feature units in full connection hidden layer can be calculated in respective formula and input node value in formula group.The
A certain feature unit includes multiple calculated values in the calculated result of respective formula in one formula group, then takes being averaged for multiple calculated values
It is worth the feature unit value as this feature unit.
S142, the output section that all output nodes are calculated according to respective formula and feature unit value in the second formula group
Point value.
The output node value of all output nodes is calculated according to respective formula and feature unit value in the second formula group,
It is updated due to having carried out adjustment into the parameter for crossing all formula in the second formula group in client's Matching Model after training, because
The output node value of all output nodes can be calculated by all formula in the second formula group and feature unit value for this, defeated
Egress value can indicate the matching degree between customers corresponding to a certain client and the output node, that is to say be calculated to
Recommend the matching degree between client and each customers.
S143, the customers according to corresponding to the output node value acquisition highest output node value being calculated are as mesh
Mark customers.
According to customers corresponding to the output node value acquisition highest output node value being calculated as target visitor
Family group.Each output node corresponds to a customers in client's Matching Model, obtain the highest customers of output node value and
Target customers corresponding with client to be recommended can be obtained.
S150, according to preset target product screening rule and the target customers to the product in the client information table
Information is screened to obtain target product and push to the user terminal.
According to preset target product screening rule and the target customers to the product information in the client information table
It is screened to obtain target product and push to the user terminal.Specifically, comprising each client in client information table
The product information that customer information and each client are bought is produced comprising at least one in the corresponding product information of each client
Product include multiple clients in target customers, can be in the Customer Acquisition client information table according to included in target customers
Target product information corresponding with the target customers, and target product information is screened according to target product screening rule
To obtain target product, and target product is pushed into user terminal to complete the recommendation for carrying out product to client to be recommended.
In one embodiment, as shown in fig. 7, step S150 includes sub-step S151 and S152.
S151, target product information corresponding with the target customers in the product information of client information table is obtained.
Obtain target product information corresponding with the target customers in the product information of client information table.Specifically,
The product information that customer information and the client in client information table comprising each client are bought, then obtain target customer
Group in each client corresponding product information in client information table, can be obtained in product information with the target customers couple
The target product information answered.
S152, it is screened according to target product screening rule pair and the target product information to obtain target product simultaneously
Push to the user terminal.
It is screened according to target product screening rule pair with the target product information to obtain target product and push
To the user terminal.Specifically, the number occurred to each product in target product information is counted to obtain product
Number statistical result, and obtained product number statistical result is screened according to target product screening rule, final
User terminal is pushed to the target product recommended to client to be recommended, and by target product.
For example, the product quantity in settable target product recommendation rules is three, then product number statistical result is obtained
First three product of middle quantity ranking is recommended as target product to client to be recommended.
Based in the matched Products Show method of user provided by the embodiment of the present invention, according to clustering rule to client
Client included in information table is clustered to obtain customers' information, constructs client according to customers' information and full connection hidden layer
Matching Model is trained client's Matching Model according to training parameter, according to after training client's Matching Model obtain with to
Recommend the corresponding target customers of client, according to target product screening rule and target customers to the product in client information table
Information is screened to obtain target product and be pushed.By the above method, the feelings of our company's product can not be bought in user
Under condition, the user group to match with the user is obtained based on the matched mode of user, and is further obtained corresponding with the user group
Target product recommend to the client, there is more wide applicability, achieve good technology in actual application
Effect.
The embodiment of the present invention also provides one kind based on the matched Products Show device of user, should be based on the matched product of user
Recommendation apparatus is for executing the aforementioned any embodiment based on the matched Products Show method of user.Specifically, referring to Fig. 8,
Fig. 8 is the schematic block diagram provided in an embodiment of the present invention based on the matched Products Show device of user.It should be matched based on user
Products Show device can be configured in management server 10.
As shown in figure 8, including cluster cell 110, client's Matching Model structure based on the matched Products Show device 100 of user
Build unit 120, client's Matching Model training unit 130, target customers' matching unit 140 and product information screening unit 150.
Cluster cell 110, if the client information table inputted for receiving administrator, according to preset clustering rule to institute
Client included in client information table is stated to be clustered to obtain customers' information.
If receiving the client information table that administrator is inputted, according to preset clustering rule to institute in the client information table
The client for including is clustered to obtain customers' information, wherein in client information table comprising multiple clients customer information with
And the product information of each client's purchase.Administrator is the user of management server, and clustering rule is for all
Client carries out the Rule Information of clustering, after being clustered based on customer information to client can will similar client segmentation to one
Customers' information comprising multiple customers can be obtained in customers after clustering to all clients.
In other inventive embodiments, the cluster cell 110 includes subelement: characteristic variable converting unit 111 and client
Cluster cell 112.
Characteristic variable converting unit 111, for according to the characteristic variable transformation rule in the clustering rule by the visitor
Customer information included in the information table of family is converted to characteristic variable.
Client included in the client information table is believed according to the characteristic variable transformation rule in the clustering rule
Breath is converted to characteristic variable.Specifically, including the information such as customer name, age, occupation, hobby, feature in customer information
Variable transformation rule is the Rule Information for the customer information of each client to be converted to characteristic variable, every in customer information
One information can be converted to corresponding vector value by characteristic variable transformation rule and is indicated, then can be by each customer information
Corresponding conversion is the feature vector of a multidimensional, that is to say characteristic variable.Each client correspondence possesses phase in client information table
The customer information answered can then obtain the customer information of all clients in client information table, pass through characteristic variable transformation rule
It obtains the customer information of all clients and is converted to corresponding characteristic variable, characteristic variable transformation rule is for believing client
Breath is converted to the rule of characteristic variable, in customer information each single item information with a vector value in characteristic variable transformation rule
It is corresponding.
Customer clustering unit 112, for according in the clustering rule clustering algorithm and the characteristic variable to client
It is clustered to obtain customers' information comprising multiple customers.
According in the clustering rule clustering algorithm and the characteristic variable client is clustered to obtain comprising more
Customers' information of a customers, the characteristic variable in customers' information comprising multiple customers and each customers' mass center
Value presets final required monoid specifically, clustering algorithm is K-means clustering algorithm in K-means clustering algorithm
Quantity k, specifically, the corresponding characteristic variable of each customer information is indicated using a point coordinate based on three-dimensional system of coordinate,
All three-dimensional coordinate points are clustered according to the numerical value of k, obtain the three-dimensional coordinate of k customers and each customers' mass center
Value, the D coordinates value of mass center that is to say the characteristic variable value of mass center.
Client's Matching Model construction unit 120, for according to customers' information and preset full connection hidden layer building packet
Client's Matching Model containing input node and output node.
According to customers' information and preset full client of the connection hidden layer building comprising input node and output node
With model.Customers' information is by the obtained information comprising multiple customers of clustering algorithm, each customers Zhong Bao
Customer information and the customers containing multiple clients correspond to the characteristic information of mass center.It include multiple inputs in client's Matching Model
Node and multiple output nodes, wherein the corresponding input node of each of client characteristics variable dimension, each client
The corresponding output node of group, each output node one output node value of corresponding output, output node value can indicate a certain
Matching degree between customers corresponding to client and the output node.Due to passing through the spy of the obtained each client of transformation rule
Number of dimensions included in sign variable is all the same, therefore can pass through the corresponding input for generating identical quantity of the dimension of characteristic variable
Node passes through the corresponding output node for generating identical quantity of the quantity of customers in customers' information.Full connection hidden layer is use
In the middle layer contacted input node and output node, complete connect includes several feature units in hidden layer, each
Feature unit is associated with all input nodes and all output nodes, and feature unit can be used for reflecting every in customer information
The quantity of one relationship between information and customers, feature unit can be set according to customers' information, preferred feature
The quantity of unit may be set to the 1/3-1/2 of customers' quantity.
In other inventive embodiments, client's Matching Model construction unit 120 includes subelement: input node building is single
Member 121, output node construction unit 122, the first formula group construction unit 123 and the second formula group construction unit 124.
Input node construction unit 121, for constructing client according to the dimension of client characteristics variable in customers' information
Input node with model.
According to the input node of dimension building client's Matching Model of client characteristics variable in customers' information, specifically,
Each of client characteristics variable dimension corresponds to an input node, therefore can generate phase by the way that the dimension of characteristic variable is corresponding
With the input node of quantity, the dimension and input node that characteristic variable is included are corresponded.
Output node construction unit 122, for constructing client's Matching Model according to the quantity of customers in customers' information
Output node.
According to the output node of customers' information architecture client's Matching Model, specifically, each customers is one corresponding
Output node, therefore the corresponding output node for generating identical quantity of quantity of customers in customers' information, customers can be passed through
Customers and output node in information correspond.
First formula group construction unit 123, for according to feature unit in preset full connection hidden layer and constructed defeated
Ingress constructs input node to the of feature unit as input value, feature unit value as output valve using input node value
One formula group.
According to the preset complete multiple feature units connected in hidden layer and constructed multiple input nodes, with input node value
The first formula group of input node to feature unit is constructed as output valve as input value, feature unit value, wherein first is public
Formula group includes formula of all input nodes to all feature units.Input node is in client's Matching Model for a certain
The node that the characteristic variable of a client is inputted, the specific value of input node are input node value, that is to say the client
The vector value of respective dimensions in characteristic variable, due to a dimension in each input node character pair variable, all inputs
The i.e. corresponding client characteristics variable of nodal value, feature unit value are the calculated value of the feature unit in full connection hidden layer.
Second formula group construction unit 124, for according to feature unit in preset full connection hidden layer and constructed defeated
Egress, using feature unit value as input value, output node value as output valve construction feature unit to output node
Two formula groups.
According to the multiple feature units and constructed multiple output nodes in preset full connection hidden layer, with feature unit value
The second formula group as input value, output node value as output valve construction feature unit to output node, wherein second is public
Formula group includes formula of all feature units to all output nodes.Output node is in client's Matching Model for client
The node that matching degree between respective client group is exported, the specific value of output node are output node value, output
Nodal value can indicate that the matching degree between customers corresponding to a certain client and the output node, feature unit value are to connect entirely
Connect the calculated value of the feature unit in hidden layer.
Client's Matching Model training unit 130, for according to preset training parameter to constructed client's Matching Model into
Row training is with client's Matching Model after being trained.
Constructed client's Matching Model is trained according to preset training parameter and is matched with the client after being trained
Model.For the accuracy for improving client's Matching Model, constructed client's Matching Model can be carried out by preset training parameter
Training, can be obtained the client's Matching Model trained.Characteristic variable and each visitor in training parameter including multiple clients
The matching degree at family and corresponding target customers, target customers are that matching degree is highest between the client in customers' information
One customers.
In other inventive embodiments, client's Matching Model training unit 130 includes subelement: feature unit value is calculated
Unit 131, output node value acquiring unit 132, square mean error amount computing unit 133 and parameter adjustment unit 134.
Feature unit value computing unit 131, for obtaining the characteristic variable of a client in preset training parameter, according to
The feature unit value of all feature units in full connection hidden layer is calculated in respective formula and input node value in one formula group.
The characteristic variable for obtaining a client in preset training parameter is saved according to respective formula in the first formula group and input
The feature unit value of all feature units in full connection hidden layer is calculated in point value.Due to every dimension in the characteristic variable of client
An input node is corresponded to, therefore using vector value included in characteristic variable as the input node of corresponding input node
The characteristic variable of the user can be inputted client's Matching Model by value.Each feature unit is and institute in client's Matching Model
There are input node and all output nodes to be associated, that is to say through respective formula and input node value in the first formula group i.e.
The feature unit value of all feature units in full connection hidden layer can be calculated.The calculated result of respective formula in first formula group
In a certain feature unit include multiple calculated values, then take the average value of multiple calculated values as the feature unit of this feature unit
Value.
Output node value acquiring unit 132, for being calculated according to respective formula and feature unit value in the second formula group
To the output node value of corresponding output node.
The output node value of corresponding output node is calculated according to respective formula and feature unit value in the second formula group,
The customers as included in customers' information and output node correspond, and pass through respective formula and spy in the second formula group
Output node value can be calculated in sign cell value, and output node value can indicate visitor corresponding to a certain client and the output node
Matching degree between the group of family, that is to say the matching degree being calculated between existing customer and each customers.Since client matches
The parameter of formula in model is also adjusted update without training, therefore the output node value being currently calculated is only capable of using
It is trained in client's Matching Model, and can not be as final result to match client with corresponding customers.
Square mean error amount computing unit 133, for obtain existing customer and the matching degree of corresponding target customers and with this
The corresponding target output node value of target customers is exported according to matching degree and target of the mean square error function to target customers
Nodal value is calculated to obtain the square mean error amount of existing customer.
The matching degree and target corresponding with target customers output for obtaining existing customer and corresponding target customers save
Point value, the matching degree of each client and corresponding target customers in training parameter, target customers be in customers' information with
The highest customers of matching degree between the client, then can obtain the client for the matching degree of target customers, and obtain
An output node value corresponding with the target customers is as target output node value in all output node values.According to square
Error function calculates the matching degree and target output node value of target customers, to obtain the mean square error of existing customer
Value, mean square error function (MSE function) is a kind of calculating letter for reflecting difference degree between estimator and the amount of being estimated
Number can be calculated the matching degree of target customers and client's Matching Model in training parameter by mean square error function and calculate
The obtained error amount between target output node value, the square mean error amount the big, shows client's target in client's training parameter
The error between target output node value that the matching degree and client's Matching Model of customers are calculated is bigger, square mean error amount
It is smaller, show client's target that the matching degree of target customers and client's Matching Model are calculated in client's training parameter
Error between output node value is smaller.
Parameter adjustment unit 134, for whether being less than mean square error to square mean error amount according to preset mean square error threshold value
Threshold value is judged, is adjusted according to judging result and the parameter regulation coefficient to the parameter of formula in client's Matching Model
It updates.
Judged according to preset mean square error threshold value whether square mean error amount is less than mean square error threshold value, according to judgement
As a result and the parameter regulation coefficient is adjusted update to the parameter of formula in client's Matching Model, specifically, parameter adjusts
It include parameter adjustment direction and parameter adjustment magnitude in coefficient, parameter adjustment direction includes positive adjustment and negative sense adjustment, parameter
Adjustment amplitude is the specific range value being adjusted, and carries out what primary adjustment updated to the parameter of formula in client's Matching Model
If specific steps include: that judging result is square mean error amount no more than preset mean square error threshold value, according to parameter adjustment direction
In positive adjustment and the range value in parameter adjustment magnitude update is adjusted to the parameter of formula in client's Matching Model;If
Judging result is that square mean error amount is greater than preset mean square error threshold value, then according to the reversed adjustment and parameter in parameter adjustment direction
Range value in adjustment amplitude is adjusted update to the parameter of formula in client's Matching Model.It completes in client's Matching Model
After the parameter of formula carries out primary adjustment update, the characteristic variable of next client and the client couple in training parameter can be obtained
It answers the matching degree of target customers to be iterated training to client's Matching Model, passes through the feature of clients multiple in training parameter
Variable and each client correspond to after the matching degree of target customers is iterated training to client's Matching Model, can be obtained with
Trained client's Matching Model.
Target customers' matching unit 140, if for receive client to be recommended by user terminal inputted wait push away
Customer information is recommended, by client's Matching Model after the customer information input training to be recommended, and the client is obtained and matches mould
In type between the client to be recommended the highest customers of matching degree as target customers.
If receiving client to be recommended passes through the customer information to be recommended that user terminal is inputted, by the client to be recommended
Client's Matching Model after information input training, and obtain in client's Matching Model the matching degree between the client to be recommended
Highest customers are as target customers.Wherein, client to be recommended is the user of user terminal, customer information to be recommended
In include client to be recommended characteristic variable.
In other inventive embodiments, target customers' matching unit 140 includes subelement: the first computing unit 141,
Second computing unit 142 and target customers' acquiring unit 143.
First computing unit 141, for obtaining the characteristic variable of client to be recommended as respective formula in the first formula group
Input node value, all features in full connection hidden layer are calculated according to respective formula and input node value in the first formula group
The feature unit value of unit.
Input node value of the characteristic variable of client to be recommended as respective formula in the first formula group is obtained, according to first
The feature unit value of all feature units in full connection hidden layer is calculated in respective formula and input node value in formula group.Due to
Every dimension corresponds to an input node in the characteristic variable of client to be recommended, therefore by vector included in characteristic variable
It is worth the input node value as corresponding input node, the characteristic variable of the user can be inputted client's Matching Model.Client
It is associated, that is to say through the first public affairs with all input nodes and all output nodes with each feature unit in model
The feature unit value of all feature units in full connection hidden layer can be calculated in respective formula and input node value in formula group.The
A certain feature unit includes multiple calculated values in the calculated result of respective formula in one formula group, then takes being averaged for multiple calculated values
It is worth the feature unit value as this feature unit.
Second computing unit 142, it is all for being calculated according to respective formula and feature unit value in the second formula group
The output node value of output node.
The output node value of all output nodes is calculated according to respective formula and feature unit value in the second formula group,
It is updated due to having carried out adjustment into the parameter for crossing all formula in the second formula group in client's Matching Model after training, because
The output node value of all output nodes can be calculated by all formula in the second formula group and feature unit value for this, defeated
Egress value can indicate the matching degree between customers corresponding to a certain client and the output node, that is to say be calculated to
Recommend the matching degree between client and each customers.
Target customers' acquiring unit 143, for obtaining highest output node according to the output node value being calculated
The corresponding customers of value are as target customers.
According to customers corresponding to the output node value acquisition highest output node value being calculated as target visitor
Family group.Each output node corresponds to a customers in client's Matching Model, obtain the highest customers of output node value and
Target customers corresponding with client to be recommended can be obtained.
Product information screening unit 150 is used for according to preset target product screening rule and the target customers to institute
The product information in client information table is stated to be screened to obtain target product and push to the user terminal.
According to preset target product screening rule and the target customers to the product information in the client information table
It is screened to obtain target product and push to the user terminal.Specifically, comprising each client in client information table
The product information that customer information and each client are bought is produced comprising at least one in the corresponding product information of each client
Product include multiple clients in target customers, can be in the Customer Acquisition client information table according to included in target customers
Target product information corresponding with the target customers, and target product information is screened according to target product screening rule
To obtain target product, and target product is pushed into user terminal to complete the recommendation for carrying out product to client to be recommended.
In other inventive embodiments, the product information screening unit 150 includes subelement: target product acquisition of information list
Member 151 and target product acquiring unit 152.
Target product information acquisition unit 151, in the product information for obtaining client information table with the target customer
The corresponding target product information of group.
Obtain target product information corresponding with the target customers in the product information of client information table.Specifically,
The product information that customer information and the client in client information table comprising each client are bought, then obtain target customer
Group in each client corresponding product information in client information table, can be obtained in product information with the target customers couple
The target product information answered.
Target product acquiring unit 152, for being carried out according to target product screening rule pair and the target product information
Screening is to obtain target product and push to the user terminal.
It is screened according to target product screening rule pair with the target product information to obtain target product and push
To the user terminal.Specifically, the number occurred to each product in target product information is counted to obtain product
Number statistical result, and obtained product number statistical result is screened according to target product screening rule, final
User terminal is pushed to the target product recommended to client to be recommended, and by target product.
It is matched using above-mentioned based on user based on the matched Products Show device of user provided by the embodiment of the present invention
Products Show method, client included in client information table is clustered according to clustering rule to obtain customers' information,
Client's Matching Model is constructed according to customers' information and full connection hidden layer, client's Matching Model is instructed according to training parameter
Practice, target customers corresponding with client to be recommended are obtained according to client's Matching Model after training, are screened according to target product
Rule and target customers are screened to obtain target product and be pushed to the product information in client information table.By upper
Method is stated, can be obtained based on the matched mode of user and be matched with the user in the case where user does not buy our company's product
User group, and further obtain corresponding with user group target product and recommend to the client, have more extensive applicable
Property, good technical effect is achieved in actual application.
The above-mentioned form that can be implemented as computer program based on the matched Products Show device of user, the computer program
It can be run in computer equipment as shown in Figure 9.
Referring to Fig. 9, Fig. 9 is the schematic block diagram of computer equipment provided in an embodiment of the present invention.
Refering to Fig. 9, which includes processor 502, memory and the net connected by system bus 501
Network interface 505, wherein memory may include non-volatile memory medium 503 and built-in storage 504.
The non-volatile memory medium 503 can storage program area 5031 and computer program 5032.The computer program
5032 are performed, and processor 502 may make to execute based on the matched Products Show method of user.
The processor 502 supports the operation of entire computer equipment 500 for providing calculating and control ability.
The built-in storage 504 provides environment for the operation of the computer program 5032 in non-volatile memory medium 503, should
When computer program 5032 is executed by processor 502, processor 502 may make to execute based on the matched Products Show side of user
Method.
The network interface 505 is for carrying out network communication, such as the transmission of offer data information.Those skilled in the art can
To understand, structure shown in Fig. 9, only the block diagram of part-structure relevant to the present invention program, is not constituted to this hair
The restriction for the computer equipment 500 that bright scheme is applied thereon, specific computer equipment 500 may include than as shown in the figure
More or fewer components perhaps combine certain components or with different component layouts.
Wherein, the processor 502 is for running computer program 5032 stored in memory, to realize following function
Can: if the client information table that administrator is inputted is received, according to preset clustering rule to included in the client information table
Client clustered to obtain customers' information, wherein further include the product of each client purchase in the client information table
Information;Mould is matched comprising the client of input node and output node according to customers' information and preset full connection hidden layer building
Type;Constructed client's Matching Model is trained according to preset training parameter with client's Matching Model after being trained;
If receiving client to be recommended passes through the customer information to be recommended that user terminal is inputted, the customer information to be recommended is inputted
Client's Matching Model after training, and obtain in client's Matching Model the highest visitor of matching degree between the client to be recommended
Family group is used as target customers;According to preset target product screening rule and the target customers in the client information table
Product information screened to obtain target product and push to the user terminal.
In one embodiment, if processor 502 is executing the client information table for receiving administrator and being inputted, according to preset
When clustering rule clusters client included in the client information table to obtain the step of customers' information, execute such as
Lower operation: according to the characteristic variable transformation rule in the clustering rule by customer information included in the client information table
Be converted to characteristic variable;According in the clustering rule clustering algorithm and the characteristic variable client is clustered to obtain
Customers' information comprising multiple customers.
In one embodiment, processor 502 is being executed according to customers' information and preset full connection hidden layer building packet
It when the step of client's Matching Model containing input node and output node, performs the following operations: according to client in customers' information
The input node of dimension building client's Matching Model of characteristic variable;Client is constructed according to the quantity of customers in customers' information
The output node of Matching Model;According to the preset full feature unit connected in hidden layer and constructed input node, to input section
Point value constructs the first formula group of input node to feature unit as input value, feature unit value as output valve;According to pre-
The feature unit and constructed output node in full connection hidden layer are set, using feature unit value as input value, output node value
The second formula group as output valve construction feature unit to output node.
In one embodiment, processor 502 execute according to preset training parameter to constructed client's Matching Model into
When row training is with the step of client's Matching Model after being trained, performs the following operations: obtaining in preset training parameter one
All spies in full connection hidden layer are calculated according to respective formula and input node value in the first formula group in the characteristic variable of client
Levy the feature unit value of unit;Corresponding output node is calculated according to respective formula and feature unit value in the second formula group
Output node value;The matching degree and target corresponding with the target customers for obtaining existing customer and corresponding target customers export
Nodal value is calculated current to obtain according to matching degree and target output node value of the mean square error function to target customers
The square mean error amount of client;Sentenced according to preset mean square error threshold value to whether square mean error amount is less than mean square error threshold value
It is disconnected, update is adjusted to the parameter of formula in client's Matching Model according to judging result and the parameter regulation coefficient.
In one embodiment, if processor 502 execute receive client to be recommended by user terminal inputted to
Recommend customer information, by client's Matching Model after the customer information input training to be recommended, and obtains client's matching
In model between the client to be recommended when step of the highest customers of matching degree as target customers, following behaviour is executed
Make: input node value of the characteristic variable of client to be recommended as respective formula in the first formula group is obtained, according to the first formula
The feature unit value of all feature units in full connection hidden layer is calculated in respective formula and input node value in group;According to second
The output node value of all output nodes is calculated in respective formula and feature unit value in formula group;According to what is be calculated
Output node value obtains customers corresponding to highest output node value as target customers.
In one embodiment, processor 502 is being executed according to preset target product screening rule and the target customers
The step of product information in the client information table being screened to obtain target product and push to the user terminal
When, it performs the following operations: obtaining target product information corresponding with the target customers in the product information of client information table;
It is screened according to target product screening rule pair and the target product information to obtain target product and push to the use
Family terminal.
It will be understood by those skilled in the art that the embodiment of computer equipment shown in Fig. 9 is not constituted to computer
The restriction of equipment specific composition, in other embodiments, computer equipment may include components more more or fewer than diagram, or
Person combines certain components or different component layouts.For example, in some embodiments, computer equipment can only include depositing
Reservoir and processor, in such embodiments, the structure and function of memory and processor are consistent with embodiment illustrated in fig. 9,
Details are not described herein.
It should be appreciated that in embodiments of the present invention, processor 502 can be central processing unit (Central
Processing Unit, CPU), which can also be other general processors, digital signal processor (Digital
Signal Processor, DSP), specific integrated circuit (Application Specific Integrated Circuit,
ASIC), ready-made programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic
Device, discrete gate or transistor logic, discrete hardware components etc..Wherein, general processor can be microprocessor or
Person's processor is also possible to any conventional processor etc..
Computer readable storage medium is provided in another embodiment of the invention.The computer readable storage medium can be with
For non-volatile computer readable storage medium.The computer-readable recording medium storage has computer program, wherein calculating
If machine program performs the steps of when being executed by processor receives the client information table that administrator is inputted, according to preset poly-
Rule-like clusters client included in the client information table to obtain customers' information, wherein client's letter
It further include the product information of each client's purchase in breath table;Include according to customers' information and preset full connection hidden layer building
Client's Matching Model of input node and output node;Constructed client's Matching Model is instructed according to preset training parameter
Practice with client's Matching Model after being trained;If receiving client to be recommended passes through the client to be recommended that user terminal is inputted
Information, by client's Matching Model after the customer information to be recommended input training, and obtain in client's Matching Model with
The highest customers of matching degree are as target customers between the client to be recommended;According to preset target product screening rule and institute
Target customers are stated to screen the product information in the client information table to obtain target product and push to the use
Family terminal.
In one embodiment, if the client information table for receiving administrator and being inputted, according to preset clustering rule pair
Client included in the client information table is clustered the step of to obtain customers' information, comprising: according to the cluster
Customer information included in the client information table is converted to characteristic variable by the characteristic variable transformation rule in rule;According to
Clustering algorithm and the characteristic variable in the clustering rule cluster client to obtain the visitor comprising multiple customers
Family group's information.
In one embodiment, it is described according to customers' information and the building of preset full connection hidden layer comprising input node and
The step of client's Matching Model of output node, comprising: client is constructed according to the dimension of client characteristics variable in customers' information
The input node of Matching Model;According to the output node of quantity building client's Matching Model of customers in customers' information;Root
According to the preset full feature unit connected in hidden layer and constructed input node, using input node value as input value, feature list
Member value constructs the first formula group of input node to feature unit as output valve;According to the feature list in preset full connection hidden layer
First and constructed output node, using feature unit value as input value, output node value as output valve construction feature unit
To the second formula group of output node.
In one embodiment, described that constructed client's Matching Model is trained to obtain according to preset training parameter
The step of client's Matching Model after training, comprising: the characteristic variable for obtaining a client in preset training parameter, according to first
The feature unit value of all feature units in full connection hidden layer is calculated in respective formula and input node value in formula group;According to
The output node value of corresponding output node is calculated in respective formula and feature unit value in second formula group;Obtain existing customer
And the matching degree of corresponding target customers and target output node value corresponding with the target customers, according to mean square error function
Matching degree and target output node value to target customers are calculated to obtain the square mean error amount of existing customer;According to pre-
It sets mean square error threshold value and judges whether square mean error amount is less than mean square error threshold value, according to judging result and the parameter
Regulation coefficient is adjusted update to the parameter of formula in client's Matching Model.
In one embodiment, if the client to be recommended letter for receiving client to be recommended and being inputted by user terminal
Breath by client's Matching Model after the customer information input training to be recommended, and is obtained in client's Matching Model and is somebody's turn to do
The step of highest customers of matching degree are as target customers between client to be recommended, comprising: obtain the spy of client to be recommended
Input node value of the variable as respective formula in the first formula group is levied, according to respective formula and input node in the first formula group
The feature unit value of all feature units in full connection hidden layer is calculated in value;According to respective formula and feature in the second formula group
The output node value of all output nodes is calculated in cell value;Highest output is obtained according to the output node value being calculated
Customers corresponding to nodal value are as target customers.
In one embodiment, described that the client is believed according to preset target product screening rule and the target customers
Product information in breath table is screened the step of to obtain target product and push to the user terminal, comprising: obtains visitor
Target product information corresponding with the target customers in the product information of family information table;According to target product screening rule pair
It is screened with the target product information to obtain target product and push to the user terminal.
It is apparent to those skilled in the art that for convenience of description and succinctly, foregoing description is set
The specific work process of standby, device and unit, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
Those of ordinary skill in the art may be aware that unit described in conjunction with the examples disclosed in the embodiments of the present disclosure and algorithm
Step can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and software
Interchangeability generally describes each exemplary composition and step according to function in the above description.These functions are studied carefully
Unexpectedly the specific application and design constraint depending on technical solution are implemented in hardware or software.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
In several embodiments provided by the present invention, it should be understood that disclosed unit and method, it can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only logical function partition, there may be another division manner in actual implementation, can also will be with the same function
Unit set is at a unit, such as multiple units or components can be combined or can be integrated into another system or some
Feature can be ignored, or not execute.In addition, shown or discussed mutual coupling, direct-coupling or communication connection can
Be through some interfaces, the indirect coupling or communication connection of device or unit, be also possible to electricity, mechanical or other shapes
Formula connection.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.Some or all of unit therein can be selected to realize the embodiment of the present invention according to the actual needs
Purpose.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, is also possible to two or more units and is integrated in one unit.It is above-mentioned integrated
Unit both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, technical solution of the present invention substantially or
Person says that all or part of the part that contributes to existing technology or the technical solution can body in the form of software products
Reveal and, which is stored in a computer readable storage medium, including some instructions are used so that one
Platform computer equipment (can be personal computer, server or the network equipment etc.) executes described in each embodiment of the present invention
The all or part of the steps of method.And computer readable storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory
The various media that can store program code such as (ROM, Read-Only Memory), magnetic or disk.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can readily occur in various equivalent modifications or replace
It changes, these modifications or substitutions should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with right
It is required that protection scope subject to.
Claims (10)
1. one kind is based on the matched Products Show method of user characterized by comprising
If the client information table that administrator is inputted is received, according to preset clustering rule to included in the client information table
Client clustered to obtain customers' information, wherein further include the product of each client purchase in the client information table
Information;
Mould is matched comprising the client of input node and output node according to customers' information and preset full connection hidden layer building
Type;
Constructed client's Matching Model is trained according to preset training parameter with client's Matching Model after being trained;
If receiving client to be recommended passes through the customer information to be recommended that user terminal is inputted, by the customer information to be recommended
Client's Matching Model after input training, and obtain in client's Matching Model the matching degree highest between the client to be recommended
Customers as target customers;
The product information in the client information table is carried out according to preset target product screening rule and the target customers
Screening is to obtain target product and push to the user terminal.
2. according to claim 1 be based on the matched Products Show method of user, which is characterized in that described according to preset poly-
Rule-like clusters client included in the client information table to obtain customers' information, comprising:
Customer information included in the client information table is turned according to the characteristic variable transformation rule in the clustering rule
It is changed to characteristic variable;
According in the clustering rule clustering algorithm and the characteristic variable client is clustered to obtain comprising multiple visitors
Customers' information of family group.
3. according to claim 1 be based on the matched Products Show method of user, which is characterized in that described according to the visitor
Family group's information and preset full client Matching Model of the connection hidden layer building comprising input node and output node, comprising:
According to the input node of dimension building client's Matching Model of client characteristics variable in customers' information;
According to the output node of quantity building client's Matching Model of customers in customers' information;
According to it is preset it is full connection hidden layer in feature unit and constructed input node, using input node value as input value,
Feature unit value constructs the first formula group of input node to feature unit as output valve;
According to it is preset it is full connection hidden layer in feature unit and constructed output node, using feature unit value as input value,
Second formula group of the output node value as output valve construction feature unit to output node.
4. according to claim 1 be based on the matched Products Show method of user, which is characterized in that in the training parameter
Matching degree including parameter regulation coefficient, the characteristic variable of multiple clients and each client and corresponding target customers, described
Constructed client's Matching Model is trained according to preset training parameter with client's Matching Model after being trained, comprising:
The characteristic variable for obtaining a client in preset training parameter, according to respective formula and input node value in the first formula group
The feature unit value of all feature units in full connection hidden layer is calculated;
The output node value of corresponding output node is calculated according to respective formula and feature unit value in the second formula group;
The matching degree and target output node value corresponding with the target customers of existing customer and corresponding target customers are obtained,
It is calculated according to matching degree and target output node value of the mean square error function to target customers to obtain existing customer
Square mean error amount;
Judged according to preset mean square error threshold value whether square mean error amount is less than mean square error threshold value, according to judging result
And the parameter regulation coefficient is adjusted update to the parameter of formula in client's Matching Model.
5. according to claim 1 be based on the matched Products Show method of user, which is characterized in that the client to be recommended
It include the characteristic variable of client to be recommended in information, the client by after the customer information input training to be recommended matches mould
Type, and the highest customers of matching degree are obtained in client's Matching Model between the client to be recommended as target customer
Group, comprising:
Input node value of the characteristic variable of client to be recommended as respective formula in the first formula group is obtained, according to the first formula
The feature unit value of all feature units in full connection hidden layer is calculated in respective formula and input node value in group;
The output node value of all output nodes is calculated according to respective formula and feature unit value in the second formula group;
According to customers corresponding to the output node value acquisition highest output node value being calculated as target customers.
6. according to claim 1 be based on the matched Products Show method of user, which is characterized in that described according to preset mesh
Mark product screening rule and the target customers screen to obtain target the product information in the client information table
Product simultaneously pushes to the user terminal, comprising:
Obtain target product information corresponding with the target customers in the product information of client information table;
It is screened according to target product screening rule pair and the target product information to obtain target product and push to institute
State user terminal.
7. one kind is based on the matched Products Show device of user characterized by comprising
Cluster cell, if the client information table inputted for receiving administrator, according to preset clustering rule to the client
Client included in information table is clustered to obtain customers' information, wherein further includes each in the client information table
The product information of client's purchase;
Client's Matching Model construction unit, for including input section according to customers' information and preset full connection hidden layer building
Client's Matching Model of point and output node;
Client's Matching Model training unit, for according to preset training parameter to constructed client's Matching Model be trained with
Client's Matching Model after being trained;
Target customers' matching unit, if the client to be recommended letter inputted for receiving client to be recommended by user terminal
Breath by client's Matching Model after the customer information input training to be recommended, and is obtained in client's Matching Model and is somebody's turn to do
The highest customers of matching degree are as target customers between client to be recommended;
Product information screening unit, for being believed according to preset target product screening rule and the target customers the client
Product information in breath table is screened to obtain target product and push to the user terminal.
8. according to claim 7 be based on the matched Products Show device of user, which is characterized in that the cluster cell,
Include:
Characteristic variable converting unit, for according to the characteristic variable transformation rule in the clustering rule by the client information table
Included in customer information be converted to characteristic variable;
Customer clustering unit, for according in the clustering rule clustering algorithm and the characteristic variable client is clustered
To obtain customers' information comprising multiple customers.
9. a kind of computer equipment, including memory, processor and it is stored on the memory and can be on the processor
The computer program of operation, which is characterized in that the processor realizes such as claim 1 to 6 when executing the computer program
Any one of described in based on the matched Products Show method of user.
10. a kind of computer readable storage medium, which is characterized in that the computer-readable recording medium storage has computer journey
Sequence, the computer program make the processor execute such as base as claimed in any one of claims 1 to 6 when being executed by a processor
In the matched Products Show method of user.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910662549.XA CN110533453A (en) | 2019-07-22 | 2019-07-22 | Based on the matched Products Show method, apparatus of user, computer equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910662549.XA CN110533453A (en) | 2019-07-22 | 2019-07-22 | Based on the matched Products Show method, apparatus of user, computer equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110533453A true CN110533453A (en) | 2019-12-03 |
Family
ID=68661823
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910662549.XA Pending CN110533453A (en) | 2019-07-22 | 2019-07-22 | Based on the matched Products Show method, apparatus of user, computer equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110533453A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112085526A (en) * | 2020-09-04 | 2020-12-15 | 中国平安财产保险股份有限公司 | Information matching method and device based on user group, computer equipment and storage medium |
CN112330411A (en) * | 2020-11-17 | 2021-02-05 | 中国平安财产保险股份有限公司 | Group product recommendation method and device, computer equipment and storage medium |
CN112381598A (en) * | 2020-10-26 | 2021-02-19 | 泰康保险集团股份有限公司 | Product service information pushing method and device |
CN113706258A (en) * | 2021-08-31 | 2021-11-26 | 平安银行股份有限公司 | Product recommendation method, device, equipment and storage medium based on combined model |
CN116304374A (en) * | 2023-05-19 | 2023-06-23 | 云印技术(深圳)有限公司 | Customer matching method and system based on package data |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20100020365A (en) * | 2008-08-12 | 2010-02-22 | 건국대학교 산학협력단 | System for recommendation the goods and method thereof |
CN108228950A (en) * | 2016-12-22 | 2018-06-29 | 中国移动通信有限公司研究院 | A kind of information processing method and device |
CN109146193A (en) * | 2018-09-05 | 2019-01-04 | 平安科技(深圳)有限公司 | Product intelligent recommended method, device, computer equipment and storage medium |
WO2019037202A1 (en) * | 2017-08-24 | 2019-02-28 | 平安科技(深圳)有限公司 | Method and apparatus for recognising target customer, electronic device and medium |
CN109447728A (en) * | 2018-09-07 | 2019-03-08 | 平安科技(深圳)有限公司 | Financial product recommended method, device, computer equipment and storage medium |
WO2019062011A1 (en) * | 2017-09-28 | 2019-04-04 | 平安科技(深圳)有限公司 | Target customer group construction method, electronic device and storage medium |
US20190205965A1 (en) * | 2017-12-29 | 2019-07-04 | Samsung Electronics Co., Ltd. | Method and apparatus for recommending customer item based on visual information |
-
2019
- 2019-07-22 CN CN201910662549.XA patent/CN110533453A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20100020365A (en) * | 2008-08-12 | 2010-02-22 | 건국대학교 산학협력단 | System for recommendation the goods and method thereof |
CN108228950A (en) * | 2016-12-22 | 2018-06-29 | 中国移动通信有限公司研究院 | A kind of information processing method and device |
WO2019037202A1 (en) * | 2017-08-24 | 2019-02-28 | 平安科技(深圳)有限公司 | Method and apparatus for recognising target customer, electronic device and medium |
WO2019062011A1 (en) * | 2017-09-28 | 2019-04-04 | 平安科技(深圳)有限公司 | Target customer group construction method, electronic device and storage medium |
US20190205965A1 (en) * | 2017-12-29 | 2019-07-04 | Samsung Electronics Co., Ltd. | Method and apparatus for recommending customer item based on visual information |
CN109146193A (en) * | 2018-09-05 | 2019-01-04 | 平安科技(深圳)有限公司 | Product intelligent recommended method, device, computer equipment and storage medium |
CN109447728A (en) * | 2018-09-07 | 2019-03-08 | 平安科技(深圳)有限公司 | Financial product recommended method, device, computer equipment and storage medium |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112085526A (en) * | 2020-09-04 | 2020-12-15 | 中国平安财产保险股份有限公司 | Information matching method and device based on user group, computer equipment and storage medium |
CN112381598A (en) * | 2020-10-26 | 2021-02-19 | 泰康保险集团股份有限公司 | Product service information pushing method and device |
CN112381598B (en) * | 2020-10-26 | 2023-12-05 | 泰康保险集团股份有限公司 | Product service information pushing method and device |
CN112330411A (en) * | 2020-11-17 | 2021-02-05 | 中国平安财产保险股份有限公司 | Group product recommendation method and device, computer equipment and storage medium |
CN112330411B (en) * | 2020-11-17 | 2023-10-31 | 中国平安财产保险股份有限公司 | Group product recommendation method, group product recommendation device, computer equipment and storage medium |
CN113706258A (en) * | 2021-08-31 | 2021-11-26 | 平安银行股份有限公司 | Product recommendation method, device, equipment and storage medium based on combined model |
CN116304374A (en) * | 2023-05-19 | 2023-06-23 | 云印技术(深圳)有限公司 | Customer matching method and system based on package data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110533453A (en) | Based on the matched Products Show method, apparatus of user, computer equipment | |
Bickel et al. | A nonparametric view of network models and Newman–Girvan and other modularities | |
CN109146193A (en) | Product intelligent recommended method, device, computer equipment and storage medium | |
CN104521225B (en) | Utilize the calling mapped system and method for Bayes's mean regression (BMR) | |
Abreu et al. | Markov equilibria in a model of bargaining in networks | |
CN110909222B (en) | User portrait establishing method and device based on clustering, medium and electronic equipment | |
CN107358268A (en) | Method, apparatus, electronic equipment and computer-readable recording medium for data clusters packet | |
CN112764920A (en) | Edge application deployment method, device, equipment and storage medium | |
US20070198252A1 (en) | Optimum design management apparatus, optimum design calculation system, optimum design management method, and optimum design management program | |
Song et al. | Neural network-based reputation model in a distributed system | |
CN110413722A (en) | Address choice method, apparatus and non-transient storage medium | |
CN109167806B (en) | Uncertain QoS (quality of service) perception Web service selection method based on foreground theory | |
Cook et al. | Evaluating suppliers of complex systems: a multiple criteria approach | |
Kannan et al. | Competitive market structures: a subset selection analysis | |
Fleissig et al. | Dynamic asymptotically ideal models and finite approximation | |
Davami et al. | Improving the performance of mobile phone crowdsourcing applications | |
CN111708936B (en) | Web service recommendation system and method based on similarity propagation strategy | |
AGELL | Perceptual maps to aggregate information from decision makers | |
Kim et al. | A direct utility model for access costs and economies of scope | |
CN114116740A (en) | Method and device for determining contribution degree of participants in joint learning | |
CN110610479B (en) | Object scoring method and device | |
CN112765413A (en) | Graph data query recommendation method based on user characteristics | |
CN110807251A (en) | Network public opinion polarization method and system integrating individual heterogeneity and dynamic dependency | |
CN111400611A (en) | Service discovery method based on Web complex relationship network | |
CN104809166A (en) | Business requirement determination method and equipment |
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 |