CN107992979A - A kind of mobile platform crowdsourcing task price optimization method and system - Google Patents
A kind of mobile platform crowdsourcing task price optimization method and system Download PDFInfo
- Publication number
- CN107992979A CN107992979A CN201711403354.0A CN201711403354A CN107992979A CN 107992979 A CN107992979 A CN 107992979A CN 201711403354 A CN201711403354 A CN 201711403354A CN 107992979 A CN107992979 A CN 107992979A
- Authority
- CN
- China
- Prior art keywords
- task
- mission requirements
- crowdsourcing
- requirements person
- price
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06Q30/0283—Price estimation or determination
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0639—Item locations
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Human Resources & Organizations (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of mobile platform crowdsourcing task price optimization method and system, step (1):Data acquisition:The crowdsourcing mission bit stream of labor service crowdsourcing platform and the information of mission requirements person are gathered, crowdsourcing mission bit stream is visualized and pre-processed;Step (2):The data message that multi-angular analysis step (1) collects, the influence factor of optimization task price;Step (3):For the feature that step (2) is determined as the input value of GBDT algorithms, the output valve using task pricing model as GBDT algorithms, task pricing model is established using GBDT algorithms;Step (4):Using task pricing model, fix a price.The present invention carries out the data of crowdsourcing task performance the performance of data visualization, data analysis and characteristic optimization, the optimization of task pricing model, test and appraisal Optimized model, the task price optimization method of the self-service labor service crowdsourcing platform of final definite mobile Internet.
Description
Technical field
The present invention relates to data analysis technique field, more particularly to a kind of mobile platform crowdsourcing task price optimization method
And system.
Background technology
Today's society, mobile Internet already become the highly important terminal of acquisition of information, will be numerous and jumbled with crowdsourcing model
Investigation business perform chain and be transformed on mobile platform and implement, make numerous APP users be directly connected on a mobile platform task,
Perform and feed back, only enterprise does not provide various business inspections and information search service, and is effectively guaranteed survey data
Authenticity, shortens the cycle of investigation, reduces the cost of implementation of investigation business.And among these, task price formulates most important, conjunction
Suitable price can just attract user to get task and then improve task performance.
Problem existing in the prior art is:Substantial amounts of mission requirements person connects task into robbing in task issue peak period,
Pressure easily is brought to mobile platform server, if taking over business behavior to robbing without rational means and guiding, then
Large-scale rob takes over business behavior and will produce very detrimental effect to the operation of mobile platform server, if moreover, price
It is unreasonable to will not be able to A clear guidance mission requirements person and rob take over business, cause partial task to overstock nobody for a long time and rob and connect, waste and move
The memory space of moving platform server, but also can not guarantee upgrading in time for mobile platform server data;Taken over due to robbing
Business behavior is subordinated to the private wish of mission requirements person completely, and conventional scheduling means are not suitable for crowdsourcing task.
At present, some traditional task pricing methods are a kind of qualitatively pricing methods according to the theories of psychology, are not
For the feasible solution of crowdsourcing task, and they often pertain only to enterprise or the one-sided interests of demander, without more
Angle considers the characteristic of enterprise and demander.
The content of the invention
To solve the above-mentioned problems, the present invention provides a kind of mobile platform crowdsourcing task price optimization method and system, this
It is excellent that invention carries out data visualization, data analysis and characteristic optimization, task pricing model to the data of crowdsourcing task performance
Change, the performance of test and appraisal Optimized model, the task price optimization method of the self-service labor service crowdsourcing platform of final definite mobile Internet,
Solve the problems, such as price it is unreasonable caused by server stress.
To achieve these goals, the present invention adopts the following technical scheme that:
A kind of mobile platform crowdsourcing task price optimization method, comprises the following steps:
Step (1):Data acquisition:The crowdsourcing mission bit stream of labor service crowdsourcing platform and the information of mission requirements person are gathered, it is right
Crowdsourcing mission bit stream is visualized and pre-processed;
Step (2):The data message that multi-angular analysis step (1) collects, the influence factor of optimization task price;
Step (3):Using step (2) determine feature as GBDT algorithms input value, using task pricing model as
The output valve of GBDT algorithms, task pricing model is established using GBDT algorithms;
Step (4):Using task pricing model, fix a price.
The crowdsourcing mission bit stream includes:Crowdsourcing mission number, crowdsourcing task price, the longitude and latitude of crowdsourcing task and crowdsourcing
The implementation status of task.
The information of the mission requirements person, including the position longitude and latitude of mission requirements person, the task of mission requirements person are got
Amount, the credit value of mission requirements person.
In the step (1), crowdsourcing mission bit stream is visualized and pre-processed:
Step (1-1):According to the crowdsourcing mission bit stream of collection, extraction task price, histogram is drawn according to different prices,
Visualization processing is carried out, obtains task price distribution map;
Step (1-2):Coarse localization is carried out according to the longitude and latitude of crowdsourcing task, the city position of crowdsourcing task is visual
Change;
Step (1-3):Finely positioning is carried out according to the longitude and latitude of crowdsourcing task, the counties and districts position of crowdsourcing task is visual
Change;
Step (1-4):Pretreatment:Determine whether there is data and there is exception, and abnormal task pricing data is rejected.
The abnormal data, such as repeated data, such as longitude and latitude missing data, such as price is beyond the number of setting range
According to.
Step (2) step is:
Step (2-1):Setting task radiation radius, straight line between calculating task position and mission requirements person position away from
From;
Step (2-2):Air line distance between the task location obtained according to step (2-1) and mission requirements person position,
The mission requirements person in radiation radius is selected, statistics obtains mission requirements person's number in radiation radius;
Step (2-3):Air line distance between the task location obtained according to step (2-1) and mission requirements person position and
Mission requirements person's number in each task radiation radius that step (2-2) obtains, finds out task location and mission requirements person position
Most short actual range between putting, calculates the average distance that mission requirements person arrives task.
Step (2-4):According to Baidu map, the actual running distance of calculating task demander position and task location;
Step (2-5):Mission requirements person's prior limitation average in analysis task radiation radius:Using each in gathered data
The task of mission requirements person gets amount, the mission requirements person's number in each task radiation radius drawn according to step (2-2),
Calculate the task that is averaged of each mission requirements person in each task radiation radius and get amount;
Step (2-6):The prestige average of mission requirements person in analysis task radiation radius:Each appoint using in gathered data
The credit value for demander of being engaged in, mission requirements person's number in the task radiation radius drawn according to step (2-2), calculates each task spoke
Penetrate the average of mission requirements person's prestige in radius;
Step (2-7):Calculate each task location periphery others number of tasks:The analysis distance of task is set first, then
Calculate the distance of each task and other tasks of periphery;Count the number of tasks that each task location periphery is less than or equal to analysis distance
Mesh;
Step (2-8):The city position shown according to (1), obtains the GDP data in the city;
Step (2-9):Calculate average, variance, standard deviation, maximum, minimum value, mode, the mode number of each feature
With mode number accounting;The feature includes:It is straight between the credit worthiness of mission requirements person, mission requirements person position and task location
The task of air line distance, mission requirements person between linear distance, task location gets amount and the GDP data in city;
Step (2-10):Feature selecting:The variance of the feature in (if 2-9) is more than given threshold, the feature quilt
Selection is used to establish pricing model.
In the step (2-1), the step of air line distance d between calculating task position and mission requirements person position
For:
Dy=(BWD-GLAT) × ec × π/180.0; (2)
Dx=(BJD-GLON) × ed × π/180.0; (3)
Ec=Eb+ (Ea-Eb) × (90-GLAT)/90; (4)
Wherein, GLAT represents the latitude of task position, and GLON represents the longitude of task position,
BWD represents the latitude of mission requirements person position, and BJD represents the longitude of mission requirements person position,
Dy represents the vertical distance between task location and mission requirements person position, and dx represents task location and mission requirements
Lateral separation between person position;
Ed is the parallel of latitude radius where GLAT, and ec is used to correct because the continually changing earth radius length of latitude;
Ea represents equatorial radius, and Eb represents polar radius.
In the step (2-2), the step of obtaining mission requirements person's number in each task radiation radius is counted:
Step (2-2-1):Determine analyst coverage:The point centered on each task location, take set the circles of radiation radius as
Analyst coverage;
Step (2-2-2):The mission requirements person in each task distance range d is selected, statistics obtains each task radiation
Mission requirements person's number M in radius.
In the step (2-7), the step of calculating each other number of tasks of task location periphery:
Step (2-7-1):Calculate the air line distance of each two task location;
Step (2-7-3):The point centered on each task location, takes and sets radius as analyst coverage;
Step (2-7-4):For each task, other tasks less than setting radius are picked out, statistics obtains each
Other task numbers being engaged in radiation radius.
A kind of mobile platform crowdsourcing task price Optimization Platform, including:Memory, processor and storage are on a memory
And the computer instruction run on a processor, when the computer instruction is run by processor, complete as above either method institute
The step of stating.
A kind of computer-readable recording medium, thereon operation have computer program, and the computer program is transported by processor
During row, the as above step described in either method is completed.
GBDT:Gradient Boost Decision Tree.
Beneficial effects of the present invention:
The present invention carries out data visualization, data analysis and characteristic optimization, foundation to the data of crowdsourcing task performance
Task price Optimized model based on integrated study, the performance for Optimized model of testing and assessing, finally determine the self-service labor of mobile Internet
The task price optimization method of business crowdsourcing platform.By the present invention in that with GBDT (Gradient Boost Decision Tree)
Algorithm establishes the task pricing model of optimization, can realize the arm's length pricing of task, draws so as to fulfill connect correct is robbed to task
Lead, avoid task from robbing the pressure for connecing and being brought to server.
Brief description of the drawings
The accompanying drawings which form a part of this application are used for providing further understanding of the present application, and the application's shows
Meaning property embodiment and its explanation are used to explain the application, do not form the improper restriction to the application.
Fig. 1 is the flow chart of overall process of the present invention;
Fig. 2 is the task price distribution map of the present invention;
Fig. 3 is the task location distribution map of the present invention;
Fig. 4 is the task detail location distribution map of the present invention;
Fig. 5 is number of tasks in task radiation radius and mission requirements person's data/coherency figure.
Embodiment
It is noted that described further below is all illustrative, it is intended to provides further instruction to the application.It is unless another
Indicate, all technical and scientific terms used herein has usual with the application person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe embodiment, and be not intended to restricted root
According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singulative
It is also intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " bag
Include " when, it indicates existing characteristics, step, operation, device, component and/or combinations thereof.
Embodiment 1:
As shown in Figure 1, the present invention provides a kind of mobile Internet platform crowdsourcing task price optimization method.The present invention's
One embodiment is the task price being directed in " money-making of taking pictures " APP." money-making of taking pictures " is off the net a kind of self-service of mobile interchange
Formula service mode.User downloads APP, is registered as the mission requirements person of APP, and the needing to take pictures of the task is then got from APP
(for example upper supermarket goes to check the restocking situation of certain commodity), the reward that earning APP demarcates task.
(1) data visualization:The self-service labor service crowdsourcing data of mobile Internet are gathered, to data visualization and pretreatment;
This implementation collects the task items data and mission requirements person's information data of " money-making of taking pictures " APP, has terminated to appoint
Business project data contain each task position (using longitude and latitude represent), price and performance (" 1 " represent complete, " 0 "
Represent not completing) (such as table 1);Mission requirements person's information data has included the position (such as table 2) of mission requirements person, credit value, ginseng
Examine the task that its prestige provides to start the ticket reserving time and subscribe limit, mission requirements person's prestige is higher in principle, more preferential beginning
Task is selected, its quota is also bigger (being actually to be allotted according to reservation limit proportion when task is distributed).
Table 1 has ended task project data partial schematic diagram
Task number | Task gps latitudes | Task gps longitudes | Task is marked the price | Tasks carrying situation |
A0001 | 22.56614225 | 113.9808368 | 66 | 0 |
A0002 | 22.68620526 | 113.9405252 | 65.5 | 0 |
A0003 | 22.57651183 | 113.957198 | 65.5 | 1 |
A0004 | 22.56484081 | 114.2445711 | 75 | 0 |
A0005 | 22.55888775 | 113.9507227 | 65.5 | 0 |
A0006 | 22.55899906 | 114.2413174 | 75 | 0 |
A0007 | 22.54900371 | 113.9722597 | 65.5 | 1 |
A0008 | 22.56277351 | 113.9565735 | 65.5 | 0 |
A0009 | 22.50001192 | 113.8956606 | 66 | 0 |
A0010 | 22.5437861 | 113.9239778 | 66 | 1 |
A0011 | 22.52486369 | 113.9308596 | 65.5 | 0 |
A0012 | 22.519087 | 113.9358436 | 65.5 | 0 |
A0013 | 22.54797243 | 113.977909 | 65.5 | 1 |
A0014 | 22.50616871 | 113.9314284 | 66 | 1 |
A0015 | 22.49962566 | 113.9365145 | 66 | 1 |
A0016 | 22.54032142 | 113.9236456 | 66 | 1 |
A0017 | 22.52455419 | 113.9247319 | 65.5 | 1 |
A0018 | 22.4981901 | 113.8984817 | 66 | 0 |
A0019 | 22.54603946 | 113.9749684 | 65.5 | 1 |
2 mission requirements person's information data part schematic diagram of table
(1-1) extracts task price therein, Nogata is drawn according to different prices according to the crowdsourcing mission bit stream of collection
Figure, carries out visualization processing, obtains the distribution map of task price, such as Fig. 2.Figure it is seen that task price is concentrated mainly on
Between 65.0-75.0, occurring interruption between 75.0-85.0, thus analysis can obtain, and the task quantity of low price is relatively more, and
The task negligible amounts of high price.
(1-2) extraction has terminated the task latitude and longitude coordinates in project data, and sharp Baidu map API obtains each task
Actual geographic position distribution, such as Fig. 3.Fig. 3 can be seen that the distribution task of task is mainly distributed on five cities in Guangdong Province:It is deep
Zhen Shi, Dongguan City, Guangzhou, Foshan City, Qingyuan City.
(1-3) is accurately positioned crowdsourcing task data latitude and longitude coordinates, as shown in figure 4, data visualization is drawn
City carry out counties and districts divisions, show that Qingyuan City only has a task coordinate, it is abnormal thus to judge that this data exists, therefore handle
The task data of Qingyuan City is rejected.According to task price distribution figure in Fig. 2, find task number that price is 80.0 and 85.0 compared with
It is few, therefore temporarily the task data for being priced at the two prices is rejected, in addition analyzed.
(2) in characteristic optimization, data message that multi-angular analysis step (1) collects, the influence of optimization task price because
Element, concretely comprises the following steps:
(2-1) determines the actual longitude and latitude distance between task location and mission requirements person position:Obtained according to step (1)
Task location and mission requirements person position latitude and longitude coordinates data, determine between task location and mission requirements person position real
Border longitude and latitude distance.The actual longitude and latitude distance between 2 points is calculated, utilizes equation below:
Dy=(BWD-GLAT) × ec × π/180.0
Dx=(BJD-GLON) × ed × π/180.0
Wherein, GLAT represents the latitude of task position, and GLON represents the longitude of task position, and BWD represents to appoint
The latitude of demander of being engaged in position, BJD represent the longitude of mission requirements person position, and dy represents that task location and task need
Vertical distance between the person of asking position, dx represent the lateral separation between task location and mission requirements person position, and ed is GLAT
The parallel of latitude radius at place, ec are used to correct because the continually changing radius of a ball length of latitude, its calculation formula is:
Ec=Eb+ (Ea-Eb) × (90-GLAT)/90
Wherein, Ea represents equatorial radius, and Eb represents polar radius;
(2-2) determines mission requirements person's number in the range of task location:Firstly the need of the analyst coverage of definite task, then count
Calculation task and the actual range of mission requirements person, statistics obtain mission requirements person's number in each task radiation radius;Such as Fig. 5
It is shown.
In the step (2-2), the step of mission requirements person's number in calculating task position range:
(2-2-1) determines analyst coverage, the point centered on each task location, and the circle for taking 6 km of radius is analyst coverage,
The changing rule of analysis task price within this range, and it is to give tacit consent to the difficulty level phase of each task to provide task restriction condition
Together, without considering bad weather, take pictures and refused etc. to influence, each task radiation radius are identical;
Each mission requirements person position that (2-2-2) is obtained according to above-mentioned (2-1) and each task location laterally away from
From dx and fore-and-aft distance dy.Utilize range formulaCalculate the mission requirements in each task radiation radius
Distance d of the person to task locationti-j(i=1,2 ..., m;J=1,2 ... n), ti represents the ti task, and m, n are represented respectively
Number of tasks and mission requirements person's number.D is picked out for each taskti-j≤ 6 mission requirements person, statistics obtain each task spoke
Penetrate mission requirements person's number M in radiusti(i=1,2 ..., m).
The beeline and average distance of mission requirements person position and task location walk in (2-3) calculating task radiation radius
Suddenly:The d for utilizing (2-2) to obtainti-j(i=1,2 ..., m;J=1,2 ... n), find out beeline d thereinti-min, calculate
Mission requirements person arrives the average distance of mission requirements person position and task location in each task radiation radiusCalculation formula
For:
(2-4) calculating task demander position and the running distance of task location:Usual calculating task demander position with
Task location distance can use air line distance formulaSince demander is passed through during completion task
Distance should be between running distance, rather than two positions air line distance, it is therefore desirable to calculate mission requirements person position with
Running distance between task location.Method is:The pre-defined function for calling Baidu map to provide, calculates the reality between two coordinates
Running distance dsti-j(i=1,2 ..., m;J=1,2 ... n).
Mission requirements person's prior limitation average in (2-5) analysis task radiation radius:Utilize each task in gathered data
The preplanned mission limit r of demanderj(j=1,2 ..., n), the task in each task radiation radius drawn according to (2-2)
Demander number Mti(i=1,2 ..., m), calculate each mission requirements person in each task radiation radius and are averaged preplanned mission limit
Volume, utilizes equation below:
The prestige average of mission requirements person in (2-6) analysis task radiation radius:Needed using each task in gathered data
The credit value c for the person of askingi(j=1,2 ..., n), mission requirements person's number M in the task radiation radius drawn according to (2-2)ti(i=
1,2 ..., m), the average of mission requirements person's prestige in each task radiation radius is calculated, utilizes equation below:
(2-7) determines each task location periphery others number of tasks:Firstly the need of the analyst coverage of definite task, then count
Calculate the distance between each two task location, task number of the statistics less than or equal to certain distance range;
In the step (2-7), the step of calculating each other number of tasks of task location periphery:
(2-7-1) determines actual longitude and latitude the distance dx and dy between task location;
Longitude distance dx and latitude distance dy between each task location that (2-7-2) is obtained according to above-mentioned (2-7-1).
Utilize range formulaCalculate the air line distance d of each two task locationti-tj(i, j=1,2 ... m),
Ti represents the ti task, and m represents number of tasks;
(2-7-3) determines analyst coverage, the point centered on each task location, and the circle for taking 6 km of radius is analyst coverage;
(2-7-4) picks out d for each taskti-tj≤ 6 task, statistics are obtained in each task radiation radius
Other task number Nsti(i=1,2 ..., m).
(2-7-1) determines the step of actual longitude and latitude distance dx and dy between task location:Obtained according to step (1)
The latitude and longitude coordinates data of task location, determine actual longitude and latitude distance between task location.Calculate the actual warp between 2 points
Latitude distance, utilizes equation below:
Dy=(GLAT2-GLAT1) × ec × π/180.0
Dx=(GLON2-GLON1) × ed × π/180.0
Wherein, GLAT1 represents the latitude of 1 position of task, and GLAT2 represents the latitude of 2 position of task, GLON1
The longitude of expression task position, GLON2 represent the longitude of task 2 position, and dy represents vertical between task location
Distance, dx represent the lateral separation between task location, and ed is the parallel of latitude radius where GLAT, and ec is used to correct because latitude
Continually changing radius of a ball length, its calculation formula are:
Ec=Eb+ (Ea-Eb) × (90-GLAT)/90
Wherein, Ea represents equatorial radius, and Eb represents polar radius.
Embodiment 2:
(2-8) according to the city position of (1) displaying, by inspection information, we obtain Guangzhou, Guangdong, Shenzhen,
Dongguan City and Foshan City corresponding GDP data in 2016, as shown in table 3.According to the GDP values in four cities, four cities are drawn
It is divided into 4 grades, as shown in table 4:
Each city in 3 Guangdong Province of table GDP values in 2016
City | Guangzhou | Shenzhen | Foshan City | Dongguan City |
GDP (dollar) | 19610.94 | 19492.60 | 8630.00 | 6827.67 |
Each city in 4 Guangdong Province of table GDP grades in 2016
City | Guangzhou | Shenzhen | Foshan City | Dongguan City |
GDP grades | 4 | 3 | 2 | 1 |
In the step (2-9), feature extension includes:Mission requirements person's creditworthiness information (average, variance, the mark of extension
Accurate poor, maximum, minimum value, mode, mode number, mode number accounting), the mission requirements person position that extends and task location
Between air line distance information (average, variance, standard deviation, maximum, minimum value, mode, mode number, mode number accounting),
Running distance information (average, variance, standard deviation, maximum, minimum between the mission requirements person position of extension and task location
Value, mode, mode number, mode number accounting), air line distance information (average, variance, mark between the task location of extension
Accurate poor, maximum, minimum value, mode, mode number, mode number accounting), the mission requirements person of extension can subscribe task quota
Information (average, variance, standard deviation, maximum, minimum value, mode, mode number, mode number accounting) etc..Such as ask maximum
Value, can directly invoke max () function, min () function can be directly invoked by minimizing, and averaging can be straight in code
Connect and call mean () function, ask variance to directly invoke var () function etc., by some above-mentioned basic statistics parameters, also may be used
There are a relatively good overview and understanding with the feature to initial data;
Embodiment 3:
In the step (2-10), feature selecting is carried out using filtration method, to the feature in (2-9), is selected using variance
Method carries out feature selecting, and given threshold 10, select variance to be more than the feature of threshold value, the feature gone out as final choice, is used
In optimization pricing model.The feature one that this example is selected shares 39, the main number for including mission requirements person in analyst coverage
Mesh, mission requirements person's creditworthiness information (average, variance, standard deviation, maximum, minimum value, mode, mode number, mode number
Accounting), range information (average, variance, standard deviation, maximum, minimum value, crowd between mission requirements person position and task location
Number, mode number, mode number accounting), mission requirements person can subscribe task quota information (average, variance, standard deviation, maximum
Value, minimum value, mode, mode number, mode number accounting) etc., the Partial Feature calculated is as shown in table 5, table 6 and table 7:
5 feature selecting partial results of table
Table 6
Table 7
Embodiment 4:
In step (3), the optimization of task pricing model:The influence factor that the optimization task determined by step (2) is fixed a price, builds
Be based on the task price Optimized model of integrated study, and the present embodiment uses GBDT (Gradient Boost Decision
Tree) algorithm establishes the task pricing model of optimization.The basic thought of GBDT algorithms is:Assuming that strong that previous round iteration obtains
It is f to practise devicet-1(x), loss function is L (y, ft-1(x)), the target of epicycle iteration is find CART regression tree model weak
Learner ht(x), loss L (y, the f of epicycle are allowedt(x)=L (y, ft-1(x)+ht(x)) it is minimum.That is, epicycle iteration is found
Decision tree, will allow the loss of sample to become smaller as far as possible.GBDT arthmetic statements are:
Input:Training set sample { (x1,y1),(x2,y2),...(xm,ym), maximum iteration T, loss function L
Output:Strong learner f (x)
(3-1) initializes weak learner:
(3-2) to iteration wheel number t=1,2 ... T has:
(3-2-1) to sample i=1,2 ... m, calculates negative gradient:
(3-2-2) utilizes (xi, rti) one CART regression tree of (i=1,2 ..m) fitting, the t regression tree is obtained, its is right
The leaf node region answered is Rtj, j=1,2 ..., J.Wherein J is the number of the leaf node of regression tree t.
(3-2-3) to area foliage j=1,2 ..J, calculate best-fit values:
(3-2-4) renewal learning device:
(3-3) obtains the expression formula of strong learner f (x):
According to the algorithm, we obtain new task pricing model:
Wherein, qj(j=1,2,3,4) represents Dongguan City respectively, Foshan City, the GDP grades of Shenzhen and Guangzhou;M
Mission requirements person's number in expression task radiation radius, N represent number of tasks in task radiation radius, dsRepresent between two task coordinate points
Actual running distance.
The price Optimized model test and appraisal of step (4) task, the influence that analysis other factors fix a price task, chooses 635 tasks
Data are as training data, using the feature that step (3) obtains as independent variable, by the completion feelings corresponding to this this group task data
Condition is learnt as dependent variable;Then remaining 300 task datas in data are tested as test data, accuracy rate
Reach 0.903, have a distinct increment than the accuracy rate before optimization.
Analysis show that the peripheral tasks number in each task radiation radius is similar to be positively correlated with mission requirements person's number.
As shown in figure 5, this is consistent with the conclusion gone out of model.
The foregoing is merely the preferred embodiment of the application, the application is not limited to, for the skill of this area
For art personnel, the application can have various modifications and variations.It is all within spirit herein and principle, made any repair
Change, equivalent substitution, improvement etc., should be included within the protection domain of the application.
Claims (10)
- The optimization method 1. a kind of mobile platform crowdsourcing task is fixed a price, it is characterized in that, comprise the following steps:Step (1):Data acquisition:The crowdsourcing mission bit stream of labor service crowdsourcing platform and the information of mission requirements person are gathered, to crowdsourcing Mission bit stream is visualized and pre-processed;Step (2):The data message that multi-angular analysis step (1) collects, the influence factor of optimization task price;Step (3):Input value using the feature that step (2) determines as GBDT algorithms, is calculated task pricing model as GBDT The output valve of method, task pricing model is established using GBDT algorithms;Step (4):Using task pricing model, fix a price.
- 2. a kind of mobile platform crowdsourcing task price optimization method as claimed in claim 1, it is characterized in that, the crowdsourcing task Information includes:Crowdsourcing mission number, crowdsourcing task price, the implementation status of the longitude and latitude of crowdsourcing task and crowdsourcing task.
- 3. a kind of mobile platform crowdsourcing task price optimization method as claimed in claim 1, it is characterized in that, the mission requirements The information of person, including the position longitude and latitude of mission requirements person, the task of mission requirements person get amount, the prestige of mission requirements person Value.
- 4. a kind of mobile platform crowdsourcing task price optimization method as claimed in claim 1, it is characterized in that, the step (1) In, crowdsourcing mission bit stream is visualized and pre-processed:Step (1-1):According to the crowdsourcing mission bit stream of collection, extraction task price, draws histogram according to different prices, carries out Visualization processing, obtains task price distribution map;Step (1-2):Coarse localization is carried out according to the longitude and latitude of crowdsourcing task, the city position of crowdsourcing task is visualized;Step (1-3):Finely positioning is carried out according to the longitude and latitude of crowdsourcing task, by counties and districts' position visualization of crowdsourcing task;Step (1-4):Pretreatment:Determine whether there is data and there is exception, and abnormal task pricing data is rejected.
- 5. a kind of mobile platform crowdsourcing task price optimization method as claimed in claim 1, it is characterized in that, the step (2) Step is:Step (2-1):Setting task radiation radius, the air line distance between calculating task position and mission requirements person position;Step (2-2):Air line distance between the task location obtained according to step (2-1) and mission requirements person position, is selected Mission requirements person in radiation radius, statistics obtain mission requirements person's number in radiation radius;Step (2-3):Air line distance and step between the task location obtained according to step (2-1) and mission requirements person position Mission requirements person's number in each task radiation radius that (2-2) is obtained, find out task location and mission requirements person position it Between most short actual range, calculate mission requirements person arrive task average distance;Step (2-4):According to Baidu map, the actual running distance of calculating task demander position and task location;Step (2-5):Mission requirements person's prior limitation average in analysis task radiation radius:Utilize each task in gathered data The task of demander gets amount, the mission requirements person's number in each task radiation radius drawn according to step (2-2), calculates Each mission requirements person in each task radiation radius task that is averaged gets amount;Step (2-6):The prestige average of mission requirements person in analysis task radiation radius:Needed using each task in gathered data The credit value for the person of asking, mission requirements person's number in the task radiation radius drawn according to step (2-2), calculates each task radiation half The average of mission requirements person's prestige in footpath;Step (2-7):Calculate each task location periphery others number of tasks:The analysis distance of task is set first, then is calculated The distance of each task and other tasks of periphery;Count the task number that each task location periphery is less than or equal to analysis distance;Step (2-8):The city position shown according to (1), obtains the GDP data in the city;Step (2-9):Calculate average, variance, standard deviation, maximum, minimum value, mode, mode number and the crowd of each feature Several several accountings;The feature includes:Between the credit worthiness of mission requirements person, mission requirements person position and task location straight line away from Amount and the GDP data in city are got from the air line distance between, task location, the task of mission requirements person;Step (2-10):Feature selecting:If the variance of the feature in (2-9) is more than given threshold, the feature is chosen For establishing pricing model.
- 6. a kind of mobile platform crowdsourcing task price optimization method as claimed in claim 5, it is characterized in that, in the step In (2-1), the step of air line distance d between calculating task position and mission requirements person position, is:<mrow> <mi>d</mi> <mo>=</mo> <msqrt> <mrow> <msup> <mi>dx</mi> <mn>2</mn> </msup> <mo>+</mo> <msup> <mi>dy</mi> <mn>2</mn> </msup> </mrow> </msqrt> <mo>;</mo> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>Dy=(BWD-GLAT) × ec × π/180.0; (2)Dx=(BJD-GLON) × ed × π/180.0; (3)Ec=Eb+ (Ea-Eb) × (90-GLAT)/90; (4)Wherein, GLAT represents the latitude of task position, and GLON represents the longitude of task position,BWD represents the latitude of mission requirements person position, and BJD represents the longitude of mission requirements person position,Dy represents the vertical distance between task location and mission requirements person position, and dx represents task location and mission requirements person position Lateral separation between putting;Ed is the parallel of latitude radius where GLAT, and ec is used to correct because the continually changing earth radius length of latitude;Ea represents equatorial radius, and Eb represents polar radius.
- 7. a kind of mobile platform crowdsourcing task price optimization method as claimed in claim 5, it is characterized in that, the step (2- 2) in, the step of obtaining mission requirements person's number in each task radiation radius is counted:Step (2-2-1):Determine analyst coverage:The point centered on each task location, takes the circle for setting radiation radius as analysis Scope;Step (2-2-2):The mission requirements person in each task distance range d is selected, statistics obtains each task radiation radius Interior mission requirements person's number M.
- 8. a kind of mobile platform crowdsourcing task price optimization method as claimed in claim 5, it is characterized in that, the step (2- 7) in, the step of calculating each other number of tasks of task location periphery:Step (2-7-1):Calculate the air line distance of each two task location;Step (2-7-3):The point centered on each task location, takes and sets radius as analyst coverage;Step (2-7-4):For each task, other tasks less than setting radius are picked out, statistics obtains each task spoke Penetrate other task numbers in radius.
- The Optimization Platform 9. a kind of mobile platform crowdsourcing task is fixed a price, it is characterized in that, including:Memory, processor and it is stored in The computer instruction run on memory and on a processor, when the computer instruction is run by processor, completion is such as taken up an official post Step described in one method.
- 10. a kind of computer-readable recording medium, it is characterized in that, operation thereon has computer program, the computer program quilt When processor is run, the as above step described in either method is completed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711403354.0A CN107992979A (en) | 2017-12-22 | 2017-12-22 | A kind of mobile platform crowdsourcing task price optimization method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711403354.0A CN107992979A (en) | 2017-12-22 | 2017-12-22 | A kind of mobile platform crowdsourcing task price optimization method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107992979A true CN107992979A (en) | 2018-05-04 |
Family
ID=62041358
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711403354.0A Pending CN107992979A (en) | 2017-12-22 | 2017-12-22 | A kind of mobile platform crowdsourcing task price optimization method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107992979A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111260161A (en) * | 2018-11-30 | 2020-06-09 | 中移(杭州)信息技术有限公司 | Method and device for issuing crowdsourcing tasks |
CN111611204A (en) * | 2020-04-30 | 2020-09-01 | 中国舰船研究设计中心 | Distributed task progress data acquisition and analysis method |
CN111612135A (en) * | 2020-05-22 | 2020-09-01 | 京东数字科技控股有限公司 | Method and device for information interaction |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105976205A (en) * | 2016-05-04 | 2016-09-28 | 南京邮电大学 | Crowdsourcing sensing method and system for quality sensitive geographical regional information |
CN107301519A (en) * | 2017-06-16 | 2017-10-27 | 佛山科学技术学院 | A kind of task weight pricing method in mass-rent express system |
-
2017
- 2017-12-22 CN CN201711403354.0A patent/CN107992979A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105976205A (en) * | 2016-05-04 | 2016-09-28 | 南京邮电大学 | Crowdsourcing sensing method and system for quality sensitive geographical regional information |
CN107301519A (en) * | 2017-06-16 | 2017-10-27 | 佛山科学技术学院 | A kind of task weight pricing method in mass-rent express system |
Non-Patent Citations (1)
Title |
---|
袁锐莹: "基于用户信息和任务位置的动态打包任务定价模型", 《消费导刊》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111260161A (en) * | 2018-11-30 | 2020-06-09 | 中移(杭州)信息技术有限公司 | Method and device for issuing crowdsourcing tasks |
CN111260161B (en) * | 2018-11-30 | 2023-11-14 | 中移(杭州)信息技术有限公司 | Method and equipment for issuing crowdsourcing task |
CN111611204A (en) * | 2020-04-30 | 2020-09-01 | 中国舰船研究设计中心 | Distributed task progress data acquisition and analysis method |
CN111611204B (en) * | 2020-04-30 | 2024-03-01 | 中国舰船研究设计中心 | Distributed task progress data acquisition and analysis method |
CN111612135A (en) * | 2020-05-22 | 2020-09-01 | 京东数字科技控股有限公司 | Method and device for information interaction |
CN111612135B (en) * | 2020-05-22 | 2024-04-16 | 京东科技控股股份有限公司 | Method and device for information interaction |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106651392A (en) | Intelligent business location selection method, apparatus and system | |
CN110517080A (en) | Outdoor advertising put-on method, device, equipment and computer readable storage medium | |
CN107992979A (en) | A kind of mobile platform crowdsourcing task price optimization method and system | |
CN107766984A (en) | The optimal addressing system and method for businessman based on cloud computing | |
CN104965920A (en) | Method and apparatus for determining target address | |
CN105550199A (en) | Point position clustering method and point position clustering apparatus based on multi-source map | |
CN110991418B (en) | Synthetic aperture radar target image recognition method and system | |
Ribeiro et al. | Placing land cover pattern preferences on the map: Bridging methodological approaches of landscape preference surveys and spatial pattern analysis | |
US20210223433A1 (en) | Determination of location-specific weather information for agronomic decision support | |
CN107133289A (en) | A kind of method and apparatus of determination commercial circle | |
CN106767835A (en) | Localization method and device | |
CN104679942A (en) | Construction land bearing efficiency measuring method based on data mining | |
CN109766818A (en) | Pupil center's localization method and system, computer equipment and readable storage medium storing program for executing | |
Janus et al. | Differences in the effectiveness of land consolidation projects in various countries and their causes: Examples of Poland and Turkey | |
CN108205515A (en) | A kind of mobile platform crowdsourcing task pricing method and system | |
CN109325657A (en) | A kind of measures and procedures for the examination and approval of credit operation, storage medium and server | |
CN106886911A (en) | A kind of travelling products method and device for planning based on user's telecommunications behavioural characteristic | |
CN111179042A (en) | Method, system and device for determining consumption place of user | |
US20170236224A1 (en) | Identifying Points of Interest | |
CN109087508A (en) | Contiguous zone traffic analysis method and system based on high definition bayonet data | |
CN110233864B (en) | Crowdsourcing-based information acquisition method, device, equipment and storage medium | |
Fardani | Participatory Mapping Solution for Village Potential and Boundary | |
CN112258128B (en) | Target position estimation method, target position estimation device, electronic equipment and computer storage medium | |
CN113887993A (en) | Method, system, equipment and medium for evaluating coupling harmony of sports facilities and population | |
CN114399202A (en) | Big data visualization system for urban community |
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: 20180504 |