CN107492192A - Refuse classification integrates the changer for commodity exchange - Google Patents
Refuse classification integrates the changer for commodity exchange Download PDFInfo
- Publication number
- CN107492192A CN107492192A CN201710681263.7A CN201710681263A CN107492192A CN 107492192 A CN107492192 A CN 107492192A CN 201710681263 A CN201710681263 A CN 201710681263A CN 107492192 A CN107492192 A CN 107492192A
- Authority
- CN
- China
- Prior art keywords
- user
- formula
- model
- integration
- commodity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07F—COIN-FREED OR LIKE APPARATUS
- G07F7/00—Mechanisms actuated by objects other than coins to free or to actuate vending, hiring, coin or paper currency dispensing or refunding apparatus
- G07F7/06—Mechanisms actuated by objects other than coins to free or to actuate vending, hiring, coin or paper currency dispensing or refunding apparatus by returnable containers, i.e. reverse vending systems in which a user is rewarded for returning a container that serves as a token of value, e.g. bottles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Operations Research (AREA)
- Life Sciences & Earth Sciences (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to commodity exchange machine.Refuse classification integrates the changer for commodity exchange, and the mechanism and taking mouth of storage commodity are provided with changer, and changer is provided with controller and coupled user affirming device and radio transmitting device, and user affirming device confirms user profile；User integral is launched by refuse classification and obtained, and the first model, the second model and the 3rd model are established respectively based on LS SVN, and the rewards and punishments integration and final integration deposit user integral account that user obtains are used for commodity exchange.It is rational in infrastructure that the refuse classification, which integrates the advantages of changer for commodity exchange, can obtain commodity using refuse classification integration and e-payment simultaneously, and the mode for obtaining integration is reasonable, can improve the enthusiasm that user launches for refuse classification.
Description
Technical field
The present invention relates to commodity exchange machine, and the integration obtained is launched particular by refuse classification and carries out converting for commodity exchange
Change planes.
Background technology
Refuse classification is the reform to refuse collection disposal traditional approach, is a kind of science effectively disposed to rubbish
Management method.The situation that people deteriorate in face of growing Municipal Garbage Yield and environmental aspect, how by refuse classification management,
Realize that waste resources utilize to greatest extent, reduce refuse disposal amount, improve living environment quality, be that our times various countries are common
One of pressing issues of concern.By periodically promoting meaning of the citizen for Waste sorting recycle to citizen's Free distribution refuse bag
Know, thus in the market has some refuse bag delivery systems, such system binding refuse bag information and user profile, and rubbish is carried out
Tracking, and preferable user can be realized to Waste sorting recycle and carry out integrating system reward, preferably realize Waste sorting recycle
Meaning.Existing refuse classification integrates the amount of money that simply simply rubbish is weighed for the changer of commodity exchange and switchs to integrate
Commodity exchange is carried out, for the correct measure launched without reward of refuse classification, program is obtained without more rational integration, because
The enthusiasm that this user carries out refuse classification dispensing is not high, and existing changer can only support accumulated point exchanging function not support in addition
E-payment, in the case where integration is finished, user can not obtain commodity.
The content of the invention
The purpose of the present invention is that open one kind can be integrated for commodity exchange and e-payment using refuse classification simultaneously
The refuse classification of acquisition commodity integrates the changer for commodity exchange, and refuse classification integration is reasonable in design, can improve
The enthusiasm that user launches for refuse classification.
The present invention is achieved through the following technical solutions above-mentioned purpose：Refuse classification integrates the exchange for commodity exchange
Machine, the interior mechanism and taking mouth for being provided with storage commodity of changer, changer is provided with controller and coupled user confirms
Device and radio transmitting device, user affirming device confirm user profile；Changer is provided with electronic fare payment system, is wirelessly transferred
User profile and payment information are reached server by device, and server carries out confirmation user by radio transmitting device and controller
Integration, pass through accumulated point exchanging commodity and the operation of completion e-payment；User integral is launched by refuse classification and obtained, and is based on
LS-SVN establishes the first model, the second model and the 3rd model respectively, and the first model is the number of user's garbage throwing and correct
Rate and the model for obtaining rewards and punishments integration, the second model are number, accuracy and the rewards and punishments integration of user's garbage throwing with being weighed
The model of value, the 3rd model are market prices corresponding to weights, garbage weight, rubbish type and the different rubbish of user with obtaining most
The model integrated eventually, the rewards and punishments integration and final integration deposit user integral account that user obtains are used for commodity exchange.
Preferably, changer is provided with operating display, operating display is used for user and operates and show user and business
Product information, it is easy to use.
Preferably, user affirming device is scanner or Quick Response Code, it is easy to use.
Preferably, user affirming device includes Quick Response Code and the mobile phone of user, user passes through the user profile that confirms phone
Commodity exchange, user integral inquiry and e-payment are carried out by cell phone software afterwards, it is easy to use.
Preferably, establishing the first model using least square method SVMs LS-SVM, comprise the following steps successively：
(1) for given data vector { xj,yj}N, j=1,2 ..., N, input vector collection is combined into That is x ∈ R2；Output vector setThat is formula 1：y∈R1, yj=f (xj)+ej, f (xj) it is match value；ejIt is the inclined of fitting codomain actual value
Difference；(2) Nonlinear Mapping is definedThe data of nonlinear correlation in step 1 from former space reflection to feature spaceFormer nonlinear model is converted into the linear model of feature space, the i.e. modular form 2 of LS-SVM：ω, b are the parameters for needing to be recognized in model；(3) utilize and deconstruct principle of minimization risk, it is right
The parameter for needing to recognize in formula 2 is handled to obtain formula 3：γ is regularization parameter；RempTo lose letter
Number；(4) the rewards and punishments integration forecast model based on LS-SVN selects quadratic power of the loss function for error in object functionLetter
Several expression formulas 4：Formula 5：(5) by glug
Bright day function solves formula 4, according to KKT conditions, asks Lagrangian local derviation to obtain formula 6：
；(6) according to Mercer conditions, kernel function is defined, selects Gauss radial direction machine kernel functions formula 7 herein： K(x,xi)=
exp{-||xj-xv||}2/σ2, in formula 7σ is core width；(7) formula 7 is substituted into formula 6, it is excellent
Change problem is converted into linear equation Solve problems, obtains formula 8：
；xjIt is training number of days input data vector；xvIt is the input data vector for predicting number of days；f(xv) it is output vector collection
Close；Formula 8 is that the rewards and punishments established based on LS-SVM integrate forecast model；(8) multiple data are normalized,·2Represent the 2- norms of vector；(9) it needs to be determined that parameter directed quantity machine regularization
Parameter C, and RBF nuclear parameters σ2, the value of parameter is determined using cross-validation method, the performance table on collecting is being verified according to model
It is existing, it is determined that suitable parameter.
Preferably, according to the behavioural habits of user's garbage throwing, obtain a weights and obtained as user's garbage throwing
One important indicator of integration.
Preferably, establishing the second model using least square method SVMs LS-SVM, comprise the following steps successively：
(1) for given data vector { xj,yj}N, j=1,2 ..., N, input vector collection is combined into That is x ∈ R2；Output vector
Collection is combined intoThat is formula 1： y∈R1, yj=f (xj)+ej, f (xj) it is match value；ejIt is fitting codomain actual value
Deviation；(2) Nonlinear Mapping is definedThe data of nonlinear correlation in step 1 from former space reflection to feature spaceFormer nonlinear model is converted into the linear model of feature space, the i.e. modular form 2 of LS-SVM：ω, b are the parameters for needing to be recognized in model；(3) utilize and deconstruct principle of minimization risk, it is right
The parameter for needing to recognize in formula 2 is handled to obtain formula 3：γ is regularization parameter；RempTo lose letter
Number；(4) the rewards and punishments integration forecast model based on LS-SVN selects quadratic power of the loss function for error in object functionLetter
Several expression formulas 4：Formula 5：(5) by glug
Bright day function solves formula 4, according to KKT conditions, asks Lagrangian local derviation to obtain formula 6：
；(6) according to Mercer conditions, kernel function is defined, selects Gauss radial direction machine kernel functions formula 7 herein： K(x,xi)=
exp{-||xj-xv||}2/σ2, in formula 7σ is core width；(7) formula 7 is substituted into formula 6, it is excellent
Change problem is converted into linear equation Solve problems, obtains formula 8：
；xjIt is training number of days input data vector；xvIt is the input data vector for predicting number of days；f(xv) it is output vector collection
Close；Formula 8 is that the rewards and punishments established based on LS-SVM integrate forecast model；(8) multiple data are normalized,||·||2Represent the 2- norms of vector；(9) vector machine regularization in LS-SVM is determined
Parameter C, and RBF nuclear parameters σ2Parameter.
Preferably, establishing the 3rd model using least square method SVMs LS-SVM, comprise the following steps successively：
(1) for given data vector { xj,yj}N, j=1,2 ..., N, input vector collection is combined into That is x ∈ R2；Output vector collection is combined intoThat is formula 1：y∈R1, yj=f (xj)+ej, f (xj) it is match value；ejIt is the deviation for being fitted codomain actual value；
(2) Nonlinear Mapping is definedThe data of nonlinear correlation in step 1 from former space reflection to feature spaceFormer nonlinear model is converted into the linear model of feature space, the i.e. modular form 2 of LS-SVM：ω, b are the parameters for needing to be recognized in model；(3) utilize and deconstruct principle of minimization risk, it is right
The parameter for needing to recognize in formula 2 is handled to obtain formula 3：γ is regularization parameter；RempTo lose letter
Number；(4) the rewards and punishments integration forecast model based on LS-SVN selects quadratic power of the loss function for error in object functionLetter
Several expression formulas 4：Formula 5：(5) by glug
Bright day function solves formula 4, according to KKT conditions, asks Lagrangian local derviation to obtain formula 6：(6) according to Mercer conditions, kernel function is defined, selects Gauss radial direction machine core letters herein
Numerical expression 7：K(x,xi)=exp-| | xj-xv||}2/σ2, in formula 7σ is core width；(7) by formula
7 substitute into formula 6, and optimization problem is converted into linear equation Solve problems, obtains formula 8：
xjIt is training number of days input data vector；xvIt is the input data vector for predicting number of days；f(xv) it is output vector set；Formula 8 is
Rewards and punishments to be established based on LS-SVM integrate forecast model；(8) multiple data are normalized,·2Represent the 2- norms of vector；(9) determine that vector machine regularization is joined in LS-SVM
Number C, and RBF nuclear parameters σ 2 parameter.
It is as follows the step of row commodity exchange preferably, commodity include refuse bag and other commodity：(1) user's checking is believed
Cease, the database authentication on server, which passes through, to be entered in next step；(2) judge that user exchanges refuse bag or other commodity, if
What user exchanged is refuse bag, calls database data to continue to judge the user got record whether the calendar month has, such as
Fruit does not get record, sends instructions under server, and the user can get refuse bag, gets record if crossing, entrance is next
Step, if what is exchanged is commodity, into next step；(3) server calls database data searches the existing integration of the user, sentences
Whether disconnected integration reaches exchange integration, if existing integration reaches exchange integration, server operation database deducts associated quad
And the integration after deducting is preserved, if existing integration is inadequate, into next step；(4) judge whether user uses e-payment, such as
Fruit is without using can not get refuse bag and other commodity；If used, judge whether to deduct the remaining integration of user again, if
Agreeing to deduct remaining integration, then deduction user is remaining is integrated and pays remaining expense, if disagreed, deduction is remaining to be integrated,
Direct electron pays corresponding expense, and user can get refuse bag or other commodity, easy to use.
Preferably, cell phone software is wechat public number or refuse classification integrates APP, it is easy to use.
The refuse classification for employing above-mentioned technical proposal integrates changer for commodity exchange, and server is by wirelessly passing
Defeated device and controller carry out confirmation user integral, by accumulated point exchanging commodity and the operation of completion e-payment, and changer was both
Refuse classification can be used to launch the integration obtained and carry out commodity exchange, commodity directly can also be bought by e-payment,
The mode that integration can be used to be combined with e-payment obtains commodity.Establish the first model, the second model respectively based on LS-SVN
With the 3rd model, the rewards and punishments for being accustomed to obtaining the first model according to user behavior integrate, are integrated according to rewards and punishments and obtained by the second model
To weights, finally the market price according to corresponding to weights, the weight of garbage throwing, rubbish type, different rubbish is obtained by the 3rd model
It must integrate, due to the continuous change of data, the integration of acquisition also and differs, and the rewards and punishments integration of the better acquisition of behavioural habits can more
More, the weights of acquisition can be more, and obtaining integration also can be more, improve the enthusiasm of user's refuse classification dispensing.In summary,
The refuse classification integrate the advantages of changer for commodity exchange be it is rational in infrastructure, can simultaneously using refuse classification integration and
E-payment obtains commodity, and the mode for obtaining integration is reasonable, can improve the enthusiasm that user launches for refuse classification.
Brief description of the drawings
Fig. 1 is the structural representation of the present invention.
Embodiment
With reference to embodiment, the invention will be further described.
Refuse classification integrates the changer for commodity exchange, and the mechanism and taking mouth 3 of storage commodity are provided with changer,
Changer front panel is pinned by mechanical lock 2, changes and fill-ins opens mechanical lock 2.Changer is provided with controller
And coupled user affirming device, operating display 1, electronic fare payment system and radio transmitting device, user affirming device
Confirm user profile.Operating display is used for user and operates and show that user and merchandise news, changer are provided with e-payment
System.User profile and payment information are reached server by radio transmitting device, and server passes through radio transmitting device and control
Device carries out confirmation user integral, by the operation such as accumulated point exchanging commodity and completion e-payment.
User affirming device can be the combination of scanner, Quick Response Code 5 and user mobile phone, and the scanning mouth 6 of scanner scans
The Quick Response Code of user can confirm that user profile, and user is by the Quick Response Code 5 on mobile phone scanning changer it can be identified that user
Information.User by cell phone software after the user profile that confirms phone by carrying out commodity exchange, user integral inquiry and electronics branch
Pay.
User integral is launched by refuse classification and obtained, and establishes the first model, the second model and the respectively based on LS-SVN
Three models, the first model are the model of the number and accuracy of user's garbage throwing with obtaining rewards and punishments integration, and the second model is to use
Model of the number, accuracy and rewards and punishments integration of family garbage throwing with obtaining weights, the 3rd model is the weights of user, rubbish weight
Market price corresponding to amount, rubbish type and different rubbish and the model of the final integration obtained, rewards and punishments integration that user obtains and most
Integration deposit user integral account is used for commodity exchange eventually.
First model is established using least square method SVMs LS-SVM, comprised the following steps successively：(1) for
Fixed data vector { xj,yj}N, j=1,2 ..., N, input vector collection is combined into That is x ∈ R2；Output vector setThat is formula 1：y∈R1, yj=f (xj)+ej, f
(xj) it is match value；ejIt is the deviation for being fitted codomain actual value；(2) Nonlinear Mapping is definedNon-linear phase in step 1
The data of pass are from former space reflection to feature spaceFormer nonlinear model is converted into feature space
The modular form 2 of linear model, i.e. LS-SVM：ω, b are the parameters for needing to be recognized in model；
(3) using principle of minimization risk is deconstructed, the parameter for needing to recognize in formula 2 is handled to obtain formula 3：γ is regularization parameter；RempFor loss function；(4) the rewards and punishments integration forecast model based on LS-SVN exists
Quadratic power of the loss function for error is selected in object functionThe expression formula 4 of function：Formula 5：(5) formula 4 is solved by Lagrangian, it is bright to glug according to KKT conditions
Day function asks local derviation to obtain formula 6：；(6) according to Mercer conditions, core letter is defined
Number, Gauss radial direction machine kernel functions formula 7 is selected herein： K(x,xi)=exp-| | xj-xv||}2/σ2, in formula 7σ is core width；(7) formula 7 is substituted into formula 6, optimization problem is converted into linear equation solution
Problem, obtain formula 8：
；xjIt is training number of days input data vector；xvIt is the input data vector for predicting number of days；f(xv) it is output vector collection
Close；Formula 8 is that the rewards and punishments established based on LS-SVM integrate forecast model；(8) fixed inputoutput data includes multiple sides
The content in face, the unit of data, the order of magnitude are all not quite similar, if directly carrying out computing with initial data, are likely to occur
The data changed in a big way flood the situation of the data of smaller range change, make the reduction of prediction result accuracy, and Gauss
Radial direction machine kernel function is related to the inner product of vector in calculating, the value of data crosses conference and causes dyscalculia, so as to influence whole model
Computational efficiency, based on more than analyze, multiple data are normalized,|
|·||2Represent the 2- norms of vector；(9) it needs to be determined that parameter directed quantity machine regularization parameter C, and RBF nuclear parameters σ2, use
Cross-validation method determines the value of parameter, according to performance of the model on checking collection, it is determined that suitable parameter.
The rewards and punishments integration that dispensing obtains every time after user can be so known with the big data in database, due to
The continuous change of data, the integration obtained every time also and differ, and the rewards and punishments integration of the better acquisition of behavioural habits can be more.
According to the behavioural habits of user's garbage throwing, one that a weights obtain integration as user's garbage throwing is obtained
Important indicator.
Second model is established using least square method SVMs LS-SVM, comprised the following steps successively：(1) for
Fixed data vector { xj,yj}N, j=1,2 ..., N, input vector collection is combined into That is x ∈ R2；Output vector setThat is formula 1：y∈R1, yj=f (xj)+ej, f (xj)
It is match value；ejIt is the deviation for being fitted codomain actual value；(2) Nonlinear Mapping is definedNonlinear correlation in step 1
Data are from former space reflection to feature spaceFormer nonlinear model is converted into the line of feature space
Property model, the i.e. modular form 2 of LS-SVM：ω, b are the parameters for needing to be recognized in model；(3)
Using principle of minimization risk is deconstructed, the parameter for needing to recognize in formula 2 is handled to obtain formula 3：γ
It is regularization parameter；RempFor loss function；(4) the rewards and punishments integration forecast model based on LS-SVN selects damage in object function
Lose the quadratic power that function is errorThe expression formula 4 of function：Formula 5：(5) formula 4 is solved by Lagrangian, it is bright to glug according to KKT conditions
Day function asks local derviation to obtain formula 6：；(6) according to Mercer conditions, core letter is defined
Number, Gauss radial direction machine kernel functions formula 7 is selected herein： K(x,xi)=exp-| | xj-xv||}2/σ2, in formula 7σ is core width；(7) formula 7 is substituted into formula 6, optimization problem is converted into linear equation solution
Problem, obtain formula 8：；xjIt is training number of days input data vector；xvIt is pre- observation
Several input data vectors；f(xv) it is output vector set；Formula 8 is that the rewards and punishments established based on LS-SVM integrate forecast model；
(8) fixed inputoutput data includes the content of many aspects, and the unit of data, the order of magnitude are all not quite similar, if directly
Connect and carry out computing with initial data, then the data for being likely to occur changing in a big way flood the feelings of the data of smaller range change
Condition, make the reduction of prediction result accuracy, and be related to the inner product of vector, the value mistake of data in the calculating of Gauss radial direction machines kernel function
Conference causes dyscalculia, so as to influence the computational efficiency of whole model, is analyzed based on more than, multiple data are normalized
Processing,||·||2Represent the 2- norms of vector；(9) it needs to be determined that parameter it is oriented
Amount machine regularization parameter C, and RBF nuclear parameters σ2, the value of parameter is determined using cross-validation method, according to model on checking collection
Performance, it is determined that suitable parameter.
The weights of each user can be so predicted with the big data in database, due to the continuous change of data,
The weights of acquisition also and differ, and the weights of the better acquisition of behavioural habits can be more.
According to above-mentioned model, the 3rd model is established using least square method SVMs LS-SVM, successively
Comprise the following steps：(1) for given data vector { xj,yj}N, j=1,2 ..., N, input vector collection is combined into That is x ∈ R2；Export to
Quantity set is combined intoThat is formula 1：y∈R1, yj=f (xj)+ej, f (xj) it is match value；ejIt is that fitting codomain is actual
The deviation of value；(2) Nonlinear Mapping is definedData nonlinear correlation in step 1 are empty from former space reflection to feature
BetweenFormer nonlinear model is converted into the linear model of feature space, the i.e. modular form 2 of LS-SVM：ω, b are the parameters for needing to be recognized in model；(3) utilize and deconstruct principle of minimization risk, it is right
The parameter for needing to recognize in formula 2 is handled to obtain formula 3：γ is regularization parameter；RempFor loss
Function；(4) the rewards and punishments integration forecast model based on LS-SVN selects quadratic power of the loss function for error in object function
The expression formula 4 of function：Formula 5： (5)
Formula 4 is solved by Lagrangian, according to KKT conditions, asks Lagrangian local derviation to obtain formula 6：(6) according to Mercer conditions, kernel function is defined, selects Gauss radial direction machines herein
Kernel function formula 7：K(x,xi)=exp-| | xj-xv||}2/σ2, in formula 7σ is core width；
(7) formula 7 is substituted into formula 6, optimization problem is converted into linear equation Solve problems, obtains formula 8：xjIt is training number of days input data vector；xvIt is the input number for predicting number of days
According to vector；f(xv) it is output vector set；Formula 8 is that the rewards and punishments established based on LS-SVM integrate forecast model；(8) have determined that
Inputoutput data include the contents of many aspects, the unit of data, the order of magnitude are all not quite similar, if directly using original number
According to computing is carried out, then the data for being likely to occur changing in a big way flood the situation of the data of smaller range change, make prediction
As a result accuracy reduces, and is related to the inner product of vector in the calculating of Gauss radial direction machines kernel function, and the value of data crosses conference and causes meter
Difficulty is calculated, so as to influence the computational efficiency of whole model, is analyzed based on more than, multiple data is normalized,||·||2Represent the 2- norms of vector；(9) vector machine regularization in LS-SVM is determined
Parameter C, and RBF nuclear parameters σ2Parameter.
Big data precognition of the rule and three model cans of foundation formulated more than in database is every
The integration that individual user's garbage throwing obtains, due to the continuous change of data, the integration of acquisition also and differs, and behavioural habits are better
The weights of acquisition can be more, and the higher acquisition integration of market price of rubbish also can be more.It is accustomed to obtaining the first model according to user behavior
Rewards and punishments integration, integrated according to rewards and punishments and weights obtained by the second model, finally according to weights, the weight of garbage throwing, rubbish
Market price corresponding to type, different rubbish is integrated by the 3rd model, and the integration that the 3rd model obtains is that user obtains integration
Main Means, the first model obtain rewards and punishments integration be it is a kind of to user behavior custom rewards and punishments form.
Following table is the tables of data that different user garbage throwing obtains rewards and punishments integration, weights and integration.
The step of user integral rewards and punishments is：Inspector opens the input situation that intelligent dustbin checks each refuse bag, patrols
The Quick Response Code that inspection person is scanned with scanner on each refuse bag obtains user profile, if the refuse classification that user launches is correct,
Inspector uploads onto the server data and obtains certain integration；If the refuse classification mistake that user launches, inspector's handle
Data are uploaded onto the server and deduct certain integration.
The flow that integration obtains specifically includes：
Step 1：User is with the Quick Response Code card for binding user profile in advance, by Quick Response Code card sweeping close to intelligent dustbin
At code mouth；
Step 2：The Quick Response Code on card is scanned by code reader and obtains user's card number, and by installing in intelligent dustbin
Wireless communication terminal uploads onto the server the user profile of acquisition；
Step 3：Server by the database lookup of the server user profile, judges user with tying up user's card number
Whether fixed intelligent dustbin matches, if mismatched, server is by wireless communication terminal to sending instructions under intelligent dustbin
Without any operation, intelligent dustbin does not open the door；If it does, enter in next step；
Step 5：The rubbish species for selecting to launch according to user, presses corresponding button door and checks card, and judges the rubbish launched
It is recyclable rubbish or non-recyclable rubbish, if not recyclable rubbish, user obtains certain integration；If it can return
Rubbish, the junk data that server parsing user launches are received, and a definite integral is obtained by respective algorithms.
The flow of accumulated point exchanging specifically includes：
Step 1：User is with the Quick Response Code card for binding user profile in advance, by Quick Response Code card close to commodity exchange machine barcode scanning
At mouth 6；
Step 2：The Quick Response Code scanned by code reader on card obtains user's card number, and by wireless in commodity exchange machine
Transmitting device is specifically that wireless router 4 uploads onto the server the user profile of acquisition；
Step 3：User's card number by the database lookup user profile, is judged user and binding by background server
Whether commodity exchange machine matches, if mismatch server by wireless router 4 to send instructions under commodity exchange machine without
Any operation, if it does, entering in next step；
Step 4：It is to exchange refuse bag or commodity to judge user, if what user exchanged is refuse bag, server calls
Database data continues to judge the user got record whether the calendar month has, if not getting record, server leads to
Cross wireless router and the rotation of command operating Android system controlled motor is issued to commodity exchange machine so as to provide refuse bag, the use
Family can gets refuse bag in taking mouth 3, and record is got if crossing, and into next step, if what is exchanged is commodity, enters
In next step；
Step 5：Server calls database data searches the existing integration of the user, judges whether integration reaches exchange product
Point, if existing integration reaches exchange integration, background server operating database deducts associated quad and preserves the product after deducting
Point, if existing integration is inadequate, into next step；
Step 6：Judge whether user uses e-payment, such as wechat or Alipay to pay, if without using commodity are converted
Refuse bag and commodity will not be provided by changing planes；If used, judging whether to deduct the remaining integration of user again, remained if agreeing to deduct
Coproduct point, then deduct the remaining integration of user and pay corresponding expense, if disagreeing the remaining integration of deduction, pay accordingly
Expense, user's can get article at taking mouth 3.
The present invention uses three forecast models based on LS-SVM, establishes the number, correct of user's garbage throwing respectively
Model between rate and acquisition rewards and punishments integration；The number of user's garbage throwing, accuracy, obtain between rewards and punishments integration and weights
Model；And market price corresponding to the weights of user, the weight of user's garbage throwing, rubbish type, different rubbish and user launch
Model between the integration that rubbish obtains.The changer of the present invention uses server and database, multiple changers can be joined
It is tied, can collects and handle substantial amounts of data, with the embedded limitation for carrying out single operation of bottom before changes
Property, the blank of Internet of Things is formd, is advantageous to manage concentratedly.The present invention can ensure secure user data, be advantageous to personnel after
Phase safeguards, is easy to add user data, is not influenceed by the damage of commodity exchange machine, reduces system to routed probability.Pass through system
Upload user information and receive the instruction that server issues, realize that commodity fall so as to control the motor of commodity exchange machine to rotate
Fall.The accumulated point exchanging technology of this patent links together user, commodity exchange machine, server, database, wechat public number,
The commodity exchange machine being wherein placed in cell confirms user profile by user affirming device, and number is uploaded by radio transmitting device
According to the information of server calls database deducts user's associated quad and implements remote auto behaviour to sending instructions under commodity exchange machine
Make, and can also realize that accumulated point exchanging and electronic article are paid using cell phone software.
Claims (10)
1. refuse classification integrates the changer for commodity exchange, the mechanism and taking mouth of storage commodity are provided with changer, is converted
Change planes and be provided with controller and coupled user affirming device and radio transmitting device, user affirming device confirms user's letter
Breath；It is characterized in that changer is provided with electronic fare payment system, user profile and payment information are reached clothes by radio transmitting device
Business device, server carry out confirmation user integral, by accumulated point exchanging commodity and completion electricity by radio transmitting device and controller
The operation that son is paid；User integral is launched by refuse classification and obtained, and establishes the first model, the second model respectively based on LS-SVN
With the 3rd model, the model that the first model is the number of user's garbage throwing and accuracy integrates with obtaining rewards and punishments, the second model
It is the model of the number, accuracy and rewards and punishments integration of user's garbage throwing with obtaining weights, the 3rd model is the weights of user, rubbish
Rubbish weight, rubbish type and market price corresponding to different rubbish and the model of the final integration obtained, the rewards and punishments integration that user obtains
It is used for commodity exchange with final integration deposit user integral account.
2. refuse classification according to claim 1 integrates the changer for commodity exchange, it is characterised in that on changer
Provided with operating display, operating display is used for user and operates and show user and merchandise news.
3. refuse classification according to claim 1 integrates the changer for commodity exchange, it is characterised in that user confirms
Device is scanner or Quick Response Code.
4. refuse classification according to claim 1 integrates the changer for commodity exchange, it is characterised in that user confirms
Device includes Quick Response Code and the mobile phone of user, and user is converted by carrying out commodity by cell phone software after the user profile that confirms phone
Change, user integral is inquired about and e-payment.
5. refuse classification according to claim 1 integrates the changer for commodity exchange, it is characterised in that using minimum
Square law SVMs LS-SVM establishes the first model, comprises the following steps successively：(1) for given data vector { xj,
yj}N, j=1,2 ..., N, input vector collection is combined into That is x ∈ R2；Output vector setThat is formula 1：y∈R1, yj=f (xj)+ej, f (xj) it is fitting
Value；ejIt is the deviation for being fitted codomain actual value；(2) Nonlinear Mapping is definedThe data of nonlinear correlation in step 1 from
Former space reflection is to feature spaceFormer nonlinear model is converted into the linear model of feature space,
That is the modular form 2 of LS-SVM：ω, b are the parameters for needing to be recognized in model；(3) destructing is utilized
Principle of minimization risk, the parameter for needing to recognize in formula 2 is handled to obtain formula 3：γ is regularization
Parameter；Remp is loss function；(4) based on LS-SVN rewards and punishments integration forecast model selected in object function loss function for
The quadratic power of errorThe expression formula 4 of function：Formula 5：(5) formula 4 is solved by Lagrangian, according to KKT conditions, to glug
Bright day function asks local derviation to obtain formula 6：(6) according to Mercer conditions, definition
Kernel function, Gauss radial direction machine kernel functions formula 7 is selected herein：K(x,xi)=exp-| | xj-xv||}2/σ2, in formula 7σ is core width；(7) formula 7 is substituted into formula 6, optimization problem is converted into linear equation solution
Problem, obtain formula 8：xjIt is training number of days input data vector；xvIt is pre- observation
Several input data vectors；f(xv) it is output vector set；Formula 8 is that the rewards and punishments established based on LS-SVM integrate forecast model；
(8) multiple data are normalized,||·||2Represent the 2- models of vector
Number；(9) it needs to be determined that parameter directed quantity machine regularization parameter C, and RBF nuclear parameters σ 2, parameter is determined using cross-validation method
Value, according to model checking collection on performance, it is determined that suitable parameter.
6. refuse classification according to claim 1 integrates the changer for commodity exchange, it is characterised in that according to user
The behavioural habits of garbage throwing, obtain the index that a weights obtain integration as user's garbage throwing.
7. refuse classification according to claim 1 integrates the changer for commodity exchange, it is characterised in that using minimum
Square law SVMs LS-SVM establishes the second model, comprises the following steps successively：(1) for given data vector { xj,
yj}N, j=1,2 ..., N, input vector collection is combined into That is x ∈ R2；Output vector collection is combined intoThat is formula 1：y∈R1, yj=f (xj)
+ej, f (xj) it is match value；ejIt is the deviation for being fitted codomain actual value；(2) Nonlinear Mapping is definedNon-thread in step 1
Property related data from former space reflection to feature spaceFormer nonlinear model conversion is characterized sky
Between linear model, i.e. the modular form 2 of LS-SVM：ω, b are the ginsengs for needing to be recognized in model
Number；(3) using principle of minimization risk is deconstructed, the parameter for needing to recognize in formula 2 is handled to obtain formula 3：γ is regularization parameter；Remp is loss function；(4) the rewards and punishments integration forecast model based on LS-SVN
Quadratic power of the loss function for error is selected in object functionThe expression formula 4 of function：Formula
5：(5) formula 4 is solved by Lagrangian, according to KKT conditions, to drawing
Ge Lang functions ask local derviation to obtain formula 6：(6) it is fixed according to Mercer conditions
Adopted kernel function, Gauss radial direction machine kernel functions formula 7 is selected herein：K(x,xi)=exp-| | xj-xv||}2/σ2, in formula 7σ is core width；(7) formula 7 is substituted into formula 6, optimization problem is converted into linear equation solution
Problem, obtain formula 8：xjIt is training number of days input data vector；xvIt is pre- observation
Several input data vectors；f(xv) it is output vector set；Formula 8 is that the rewards and punishments established based on LS-SVM integrate forecast model；
(8) multiple data are normalized,||·||2Represent the 2- models of vector
Number；(9) vector machine regularization parameter C in LS-SVM, and RBF nuclear parameters σ 2 parameter are determined.
8. refuse classification according to claim 1 integrates the changer for commodity exchange, it is characterised in that using minimum
Square law SVMs LS-SVM establishes the 3rd model, comprises the following steps successively：(1) for given data vector { xj,
yj}N, j=1,2 ..., N, input vector collection is combined into That is x ∈ R2；Output vector collection is combined intoThat is formula
1：y∈R1, yj=f (xj)+ej, f (xj) it is match value；ejIt is the deviation for being fitted codomain actual value；(2) Nonlinear Mapping is definedThe data of nonlinear correlation in step 1 from former space reflection to feature spaceIt is former non-thread
Property model conversation is characterized the linear model in space, the i.e. modular form 2 of LS-SVM：ω, b are models
The middle parameter for needing to be recognized；(3) using principle of minimization risk is deconstructed, the parameter for needing to recognize in formula 2 is handled
Obtain formula 3：γ is regularization parameter；Remp is loss function；(4) the rewards and punishments integration based on LS-SVN
Forecast model selects quadratic power of the loss function for error in object functionThe expression formula 4 of function：Formula 5： (5) formula is solved by Lagrangian
4, according to KKT conditions, ask Lagrangian local derviation to obtain formula 6：
(6) according to Mercer conditions, kernel function is defined, selects Gauss radial direction machine kernel functions formula 7 herein：K(x,xi)=exp-| | xj-
xv||}2/σ2, in formula 7σ is core width；(7) formula 7 is substituted into formula 6, optimization problem conversion
For linear equation Solve problems, formula 8 is obtained：Xj is training number of days input number
According to vector；xvIt is the input data vector for predicting number of days；f(xv) it is output vector set；Formula 8 is to be established based on LS-SVM
Rewards and punishments integrate forecast model；(8) multiple data are normalized,||·|
|2Represent the 2- norms of vector；(9) vector machine regularization parameter C in LS-SVM, and RBF nuclear parameters σ 2 parameter are determined.
9. refuse classification according to claim 1 integrates the changer for commodity exchange, it is characterised in that commodity include
Refuse bag and other commodity, it is as follows the step of row commodity exchange：(1) user authentication information, database authentication are next by entering
Step；(2) judge that user exchanges refuse bag or other commodity, if what user exchanged is refuse bag, calling database data after
It is continuous to judge the user got record whether the calendar month has, if not getting record, send instructions under server, the user
Refuse bag can be got, record is got if crossing, into next step, if what is exchanged is commodity, into next step；(3) take
Business device calls database data to search the existing integration of the user, judges whether integration reaches exchange integration, if existing integration
Reaching exchange integration, server operation database deducts associated quad and preserves the integration after deducting, if existing integration is inadequate,
Into in next step；(4) judge whether user uses e-payment, if without using refuse bag and other commodity can not be got；Such as
Fruit uses, then judges whether to deduct the remaining integration of user, if agreeing to deduct remaining integration, deducts the remaining integration of user simultaneously
Remaining expense is paid, if disagreeing the remaining integration of deduction, direct electron pays corresponding expense, and user can get rubbish
Bag or other commodity.
10. refuse classification according to claim 4 integrates the changer for commodity exchange, it is characterised in that cell phone software
It is wechat public number or refuse classification integration APP.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710681263.7A CN107492192A (en) | 2017-08-10 | 2017-08-10 | Refuse classification integrates the changer for commodity exchange |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710681263.7A CN107492192A (en) | 2017-08-10 | 2017-08-10 | Refuse classification integrates the changer for commodity exchange |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107492192A true CN107492192A (en) | 2017-12-19 |
Family
ID=60645207
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710681263.7A Pending CN107492192A (en) | 2017-08-10 | 2017-08-10 | Refuse classification integrates the changer for commodity exchange |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107492192A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109993880A (en) * | 2019-03-29 | 2019-07-09 | 克拉玛依市诚谊科技服务有限公司 | Encourage the integral management system of garbage classification |
| CN111047768A (en) * | 2019-12-11 | 2020-04-21 | 苏州纳故环保科技有限公司 | Integrated machine for issuing exchange function and exchange method thereof |
| CN111814519A (en) * | 2019-04-11 | 2020-10-23 | 深圳市家家分类科技有限公司 | Behavior score obtaining method and related product |
| CN111815338A (en) * | 2019-04-11 | 2020-10-23 | 深圳市家家分类科技有限公司 | Garbage source classification and reduction comprehensive service platform and related products |
| CN112085878A (en) * | 2020-09-07 | 2020-12-15 | 广州东软科技有限公司 | House supervisory systems based on waste classification |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104008607A (en) * | 2014-06-05 | 2014-08-27 | 陈芳 | Android system based garbage classification point exchange machine |
| CN104044842A (en) * | 2014-06-23 | 2014-09-17 | 陈斌 | Garbage sorting platform and method based on two-dimensional code recognition technology |
| CN204066242U (en) * | 2014-07-29 | 2014-12-31 | 陈斌 | Environmental protection accumulated point exchanging machine |
| CN105600218A (en) * | 2016-01-07 | 2016-05-25 | 南京志达环保科技有限公司 | Trash classification and intelligent recovery system and method |
| CN205721954U (en) * | 2016-04-15 | 2016-11-23 | 江苏中恒美宏建设设计有限公司 | Point reward exchanging system |
| CN106976668A (en) * | 2017-04-20 | 2017-07-25 | 安吉国千环境科技有限公司 | A kind of Intelligent refuse classification processing system |
-
2017
- 2017-08-10 CN CN201710681263.7A patent/CN107492192A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104008607A (en) * | 2014-06-05 | 2014-08-27 | 陈芳 | Android system based garbage classification point exchange machine |
| CN104044842A (en) * | 2014-06-23 | 2014-09-17 | 陈斌 | Garbage sorting platform and method based on two-dimensional code recognition technology |
| CN204066242U (en) * | 2014-07-29 | 2014-12-31 | 陈斌 | Environmental protection accumulated point exchanging machine |
| CN105600218A (en) * | 2016-01-07 | 2016-05-25 | 南京志达环保科技有限公司 | Trash classification and intelligent recovery system and method |
| CN205721954U (en) * | 2016-04-15 | 2016-11-23 | 江苏中恒美宏建设设计有限公司 | Point reward exchanging system |
| CN106976668A (en) * | 2017-04-20 | 2017-07-25 | 安吉国千环境科技有限公司 | A kind of Intelligent refuse classification processing system |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109993880A (en) * | 2019-03-29 | 2019-07-09 | 克拉玛依市诚谊科技服务有限公司 | Encourage the integral management system of garbage classification |
| CN111814519A (en) * | 2019-04-11 | 2020-10-23 | 深圳市家家分类科技有限公司 | Behavior score obtaining method and related product |
| CN111815338A (en) * | 2019-04-11 | 2020-10-23 | 深圳市家家分类科技有限公司 | Garbage source classification and reduction comprehensive service platform and related products |
| CN111814519B (en) * | 2019-04-11 | 2024-02-02 | 深圳市家家分类科技有限公司 | Behavior score acquisition method and related products |
| CN111047768A (en) * | 2019-12-11 | 2020-04-21 | 苏州纳故环保科技有限公司 | Integrated machine for issuing exchange function and exchange method thereof |
| CN112085878A (en) * | 2020-09-07 | 2020-12-15 | 广州东软科技有限公司 | House supervisory systems based on waste classification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107492192A (en) | Refuse classification integrates the changer for commodity exchange | |
| US11524842B2 (en) | Incentivized multi-stream recycling system with fill level, volume, weight, counters, shredder, compactor, consumer identification, display and liquid drainage system | |
| US20070260466A1 (en) | System and methods for a recycling program | |
| Elia et al. | Designing Pay-As-You-Throw schemes in municipal waste management services: A holistic approach | |
| CN105416918B (en) | A kind of refuse classification management system and its application method | |
| KR20130140285A (en) | Method for collecting recycling articles through mobile terminal and system therof | |
| CN105740978A (en) | Garbage reservation recovery system and recovery method based on Internet | |
| CN105046832A (en) | Smart automatic delivering package arranging box | |
| CN105953885A (en) | Weighing system | |
| CN205826100U (en) | A kind of weighing system | |
| CN204515921U (en) | A kind of intelligent self-care express mail case | |
| KR101003659B1 (en) | Weight treatment control system for food garbage of ubiquitous sensor network basis and method thereof | |
| KR20100043807A (en) | System of management food waste based upon rfid | |
| CN108615080A (en) | A kind of power marketing service movement operation client | |
| KR20120000704A (en) | Apparatus for collecting recyclable container and system for compensating recyclable container using the same | |
| KR20110123882A (en) | System and method for managing food waste collection | |
| CN205221644U (en) | Waste classification management system | |
| CN204515846U (en) | A kind of safety-type express system | |
| KR101227800B1 (en) | Internet based garbage managemnnt system, method for operating garbage vehicle using thesame and method for operating thereof | |
| CN103839181A (en) | Trading system of Internet of Things | |
| CN207095687U (en) | Refuse classification electronic scale | |
| CN205066919U (en) | A intelligent electronic scale for network transaction | |
| CN110386377A (en) | A kind of Intelligent refuse classification system | |
| CN111599083A (en) | Garbage classification system based on chip technology and application method | |
| CN107545462A (en) | The method of refuse classification accumulated point exchanging |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171219 |