CN108876056A - A kind of shared bicycle Demand Forecast method, apparatus, equipment and storage medium - Google Patents
A kind of shared bicycle Demand Forecast method, apparatus, equipment and storage medium Download PDFInfo
- Publication number
- CN108876056A CN108876056A CN201810804908.6A CN201810804908A CN108876056A CN 108876056 A CN108876056 A CN 108876056A CN 201810804908 A CN201810804908 A CN 201810804908A CN 108876056 A CN108876056 A CN 108876056A
- Authority
- CN
- China
- Prior art keywords
- regression model
- data
- time
- bicycle
- weight
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000003860 storage Methods 0.000 title claims abstract description 13
- 238000012549 training Methods 0.000 claims abstract description 66
- 238000012795 verification Methods 0.000 claims abstract description 20
- 230000006870 function Effects 0.000 claims description 37
- 238000004590 computer program Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 3
- 241000208340 Araliaceae Species 0.000 claims 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims 1
- 235000003140 Panax quinquefolius Nutrition 0.000 claims 1
- 235000008434 ginseng Nutrition 0.000 claims 1
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000009826 distribution Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000001351 cycling effect Effects 0.000 description 1
- 230000002354 daily effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000155 melt Substances 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/40—Business processes related to the transportation industry
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
This application discloses a kind of shared bicycle Demand Forecast method, apparatus, equipment and storage medium, this method to include:It obtains and shares bicycle in set period of time and borrow car data at multiple and go back car data;Set period of time is divided into training time section and verification time section;The data that training time section is obtained carry out the building of regression model by training set as training set, and the data that verification time section obtains are collected as verifying, and the determination for carrying out Parameters in Regression Model is collected by verifying;It finds out the factor weight for influencing shared bicycle demand and incorporates regression model, obtain time series Weight Regression Model;It determines and shares bicycle in the period to be measured in the prediction result of each demand.The time series Weight Regression Model that the application establishes does not need a large amount of data, can constantly add influence factor in need of consideration on the basis of original model sufficiently simultaneously in view of a variety of factors for influencing shared bicycle demand, have preferable generalization.
Description
Technical field
The present invention relates to technical field of data processing, more particularly to a kind of shared bicycle Demand Forecast method, dress
It sets, equipment and storage medium.
Background technique
Under the fast-developing promotion of Internet technology, there are more and more " shared economy " in our countries, in recent years
Come, with ofo, Mo Bai, Yongan etc. is the enterprise of representative, relies on the Internet industry rapidly developed now, is proposed and shares certainly
Driving, and it is short in two years, city is shared bicycle system and is gradually penetrated into each city, to the " last of Public Traveling
One kilometer " bring great convenience.
But as user's usage amount increases the increase with frequency, it is following for sharing bicycle optimizing management efficiency
Important topic.With the competitive influence of enterprise, user encounters " a vehicle hardly possible when using shared bicycle sometimes
Ask ", " vehicle is completely trouble " can be also encountered sometimes.And the investment of too many shared bicycle eruption type, to natively crowded city
City street brings great pressure.Therefore, our business and government relevant departments have to adopt an effective measure, it is established that
The stronger prediction model of robustness, overall planning dispatch shared bicycle, allow shared economy that can reach the maximization of interests, allow
There is vehicle to ride when people need, while city will not be caused to bear.
Meanwhile from the angle of city management, it is special that the behaviour in service of shared bicycle also projects urban population flowing
Sign, has important references value to urban planning, urban traffic control.Therefore, public certainly in Chinese development public bicycles, research
Driving will all have great importance for this low-carbon trip mode future in global development.
Nowadays, the implementation most like for such problem is problem as conventional time series problem, with system
It counts method-autoregression integral moving average model (ARIMA model) learned and carries out modeling processing, time series numerical value is carried out
Difference acquires as steady difference sequence, and determines autoregression item p and rolling average item q.We can be daily shared bicycle
Borrow car data and go back car data and import in ARIMA model, find out after the parameter of needs then to it followed by prediction.But
It is stable (stationary) using ARIMA model needs time series data, or passes through differencing (differencing)
It is stable afterwards.But for demand this problem for sharing bicycle, it may be subjected to weather and festivals or holidays
It influences, causes ordered series of numbers unstable.In addition, ARIMA model can only substantially capture linear relationship, and nonlinear dependence cannot be captured
System;Due to the change of urban planning, a migration of partially self knee point is had, the influence of reconstruction will lead to bicycle and borrow and returns the car
Situations such as not planning a successor in timing, will lead to data becomes nonlinear data.If ARIMA model will capture the week in timing
If phase property or Seasonal, a large amount of data are needed, but since shared bicycle is to start to rise in recent years, because
This data volume may not be able to meet the needs of model.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of shared bicycle Demand Forecast method, apparatus, equipment and
Storage medium does not need a large amount of data, and can add influence factor in need of consideration on the basis of original model, have compared with
Good generalization.Its concrete scheme is as follows:
A kind of shared bicycle Demand Forecast method, including:
It obtains and shares bicycle in set period of time and borrow car data at multiple and go back car data;The set period of time
It is divided into training time section and verification time section;
The data for obtaining the training time section under online carry out regression model by the training set as training set
Building, and the data that verification time section is obtained pass through the verifying and collect the progress recurrence mould as verifying collection
The determination of shape parameter;
It finds out the factor weight for influencing the shared bicycle demand and incorporates the regression model, obtain time series
Weight Regression Model;
It is determined by the time series weighted model and shares bicycle in the period to be measured in each demand
Prediction result.
Preferably, in above-mentioned shared bicycle Demand Forecast method provided in an embodiment of the present invention, pass through the instruction
Practice the building that collection carries out regression model, specifically includes:
The data obtained by the training time section, define loss function;
The minimum number of the loss function is found as building regression model.
Preferably, in above-mentioned shared bicycle Demand Forecast method provided in an embodiment of the present invention, finding out influences institute
It states the factor weight of shared bicycle demand and incorporates the regression model, specifically include:
Reciprocal, two kinds of forms of negative exponent time gap weighting functions are separately designed to be predicted;
According to the score on line, optimal time gap weight is selected;
Export the minimum value of the loss function with the product for the time gap weight picked out.
Preferably, in above-mentioned shared bicycle Demand Forecast method provided in an embodiment of the present invention, finding out influences institute
It states the factor weight of shared bicycle demand and incorporates the regression model, specifically further include:
When multiple have periodically by means of car data with car data is gone back, and when the periodically difference of piles with different, design sunlight
Day weighting function;
It is with sunlight day or is all work when the sunlight day of a sunlight day and period to be measured in the training time section
When day or weekend, the return value of the sunlight day weighting function is larger, and otherwise return value is smaller;
According to the return value of the sunlight day weighting function, sunlight day weight is determined;
Export the minimum value of the product of the loss function, the time gap weight and the sunlight day weight.
Preferably, in above-mentioned shared bicycle Demand Forecast method provided in an embodiment of the present invention, when determining to be measured
Between bicycle is shared in section in the prediction result of each demand, specifically include:
Bicycle will be shared in set period of time car data and goes back car data as new training set in multiple borrow;
According to new training set, new loss function and each sunlight day corresponding sunlight day coefficient in the period to be measured are found out;
By the minimum value of the product of the new loss function, the time gap weight coefficient and the sunlight day weight coefficient
Multiplied by the sunlight day coefficient found out, the prediction knot that demand of the bicycle at each is shared in the period to be measured is obtained
Fruit.
The embodiment of the invention also provides a kind of shared bicycle Demand Forecast devices, including:
Data acquisition module shares bicycle at multiple by means of car data and number of returning the car for obtaining in set period of time
According to;The set period of time is divided into training time section and verification time section;
Model building module, the data for obtaining the training time section under online are as training set, by described
Training set carries out the building of regression model, and the data that the verification time section obtains are collected as verifying, is tested by described
Card collection carries out the determination of the Parameters in Regression Model;
Weight calculation module influences the factor weight of the shared bicycle demand and incorporates the recurrence for finding out
Model obtains time series Weight Regression Model;
Demand Forecast module shares bicycle for determining by the time series weighted model in the period to be measured
In the prediction result of each demand.
The embodiment of the invention also provides a kind of shared bicycle Demand Forecast equipment, including processor and memory,
Wherein, it realizes when the processor executes the computer program saved in the memory as provided in an embodiment of the present invention above-mentioned
Shared bicycle Demand Forecast method.
The embodiment of the invention also provides a kind of computer readable storage mediums, for storing computer program, wherein institute
It states and realizes such as above-mentioned shared bicycle Demand Forecast side provided in an embodiment of the present invention when computer program is executed by processor
Method.
A kind of shared bicycle Demand Forecast method, apparatus, equipment and storage medium provided by the present invention, this method
Including:It obtains and shares bicycle in set period of time and borrow car data at multiple and go back car data;The set period of time point
For training time section and verification time section;The data for obtaining the training time section under online are as training set, by described
Training set carries out the building of regression model, and the data that the verification time section obtains are collected as verifying, is tested by described
Card collection carries out the determination of the Parameters in Regression Model;It finds out the factor weight for influencing the shared bicycle demand and incorporates institute
Regression model is stated, obtains time series Weight Regression Model;It is determined in the period to be measured by the time series weighted model
Prediction result of the shared bicycle in each demand.The time series Weight Regression Model that the present invention establishes does not need greatly
The data of amount can solve the problem of time series problem data amount is few, dimension is low and unstable, Nonlinear Time Series, only
One baseline and basic data distribution need to be provided, and sufficiently can influence shared bicycle in view of a variety of simultaneously
The factor of demand can constantly add influence factor in need of consideration on the basis of original model, have preferable extensive
Property.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is the flow chart of shared bicycle Demand Forecast method provided in an embodiment of the present invention;
Fig. 2 is the structural schematic diagram of shared bicycle Demand Forecast device provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The present invention provides a kind of shared bicycle Demand Forecast method, as shown in Figure 1, including the following steps:
S101, it obtains and shares bicycle in set period of time and borrow car data at multiple and go back car data;The setting
Period is divided into training time section and verification time section;
S102, it is online under the data that obtain the training time section as training set, returned by the training set
Return the building of model, and the data that the verification time section obtains are collected as verifying, is collected described in carrying out by the verifying
The determination of Parameters in Regression Model;
S103, it finds out the factor weight for influencing the shared bicycle demand and incorporates the regression model, when obtaining
Between sequence Weight Regression Model;
S104, it is determined by the time series weighted model and shares demand of the bicycle at each in the period to be measured
The prediction result of amount.
In above-mentioned shared bicycle Demand Forecast method provided in an embodiment of the present invention, set period of time is obtained first
Interior shared bicycle borrows car data and goes back car data at multiple;Wherein the set period of time is divided into training time section and tests
Demonstrate,prove the period;Then the data for obtaining the training time section under online are returned as training set by the training set
Return the building of model, and the data that the verification time section obtains are collected as verifying, is collected described in carrying out by the verifying
The determination of Parameters in Regression Model;The factor weight for influencing the shared bicycle demand is found out later and incorporates the recurrence mould
Type obtains time series Weight Regression Model;It determines in the period to be measured and shares finally by the time series weighted model
Prediction result of the bicycle in each demand.The time series Weight Regression Model being built such that does not need a large amount of number
According to can solve the problem of time series problem data amount is few, dimension is low and unstable, Nonlinear Time Series, only need to provide
One baseline and basic data distribution, and can shared bicycle demand sufficiently be influenced in view of a variety of simultaneously
Factor, can constantly on the basis of original model add influence factor in need of consideration, have preferable generalization.
It should be noted that assuming that the period to be measured is in September, 2015, set period of time be can choose apart from the time to be measured
Closer in May, 2015 to August, can preferably reflect the trend of period to be measured, the accuracy rate of prediction result can be higher in this way.
Specifically, training time section can be chosen for in May, 2015 to July (removal 3 days May Day, 3 days Dragon Boat Festivals), verification time Duan Xuan
It is taken as in August, 2015.
Further, in the specific implementation, in above-mentioned shared bicycle Demand Forecast side provided in an embodiment of the present invention
In method, step S102 carries out the building of regression model by the training set, can specifically include:When passing through described trained first
Between section obtain data, define loss function;Then the minimum number of the loss function is found as building regression model.
In practical applications, for predicting to borrow vehicle amount, a most basic thinking is, for each stake, when will be to be measured
Between section 61 days (2 months) borrow vehicle amount be predicted as its training time section the mean value by means of vehicle amount, i.e.,:
Wherein, predAIndicate that stake A borrows the predicted value of vehicle amount;D indicates the date set of training time section;| D | indicate training
The number of days of period;LeaseA,dIndicate that stake A borrows vehicle amount in day d.Mean value is considered as at this time and borrows vehicle amount in training time section
It returns.However, it can be apparent that a regressand value more more excellent than mean value can be found by machine learning, when this value should make to train
Between penalty values in section it is minimum.Predictions using this value as this A in 61 the skys of period to be measured.I.e.:
In formula, L is loss function.Obvious predAMeet:
predA∈[mind∈D(LeaseA,d),maxd∈D(LeaseA,d)] (3)
Influence very little due to the predicted value difference less than 1 to achievement traverses (3) formula institute in the implementation for the sake of simplicity
It determines all integers in range, has selected wherein to enable the smallest pred of penalty valuesA。
Further, in the specific implementation, in above-mentioned shared bicycle Demand Forecast side provided in an embodiment of the present invention
In method, since by the observation to data and to the analysis of problem, find stake is dynamic change by means of vehicle and situation of returning the car, certain
The trend risen or fallen can be presented in a little stakes, it is contemplated that the numerical value that distance needs predicted time closer will give bigger weight, step
Rapid S103 finds out the factor weight for influencing the shared bicycle demand and incorporates the regression model, can specifically include:
Reciprocal, two kinds of forms of negative exponent time gap weighting functions are separately designed to be predicted;According to the score on line, therefrom select
Optimal time gap weight out;Export the minimum value of the loss function with the product for the time gap weight picked out.
In practical applications, it added a time gap weight every day to the training set of selection, so that closer to
The time gap weight of the sample of period to be measured is bigger.Reciprocal, two kinds of forms of negative exponent weighted time is devised apart from letter
Number:
Wherein, d0Indicate first day (such as 2015 on September 1) of period to be measured;Two dates, which subtract each other, represents the two
Number of days is poor;m1And m2It is by verifying the determining constant of collection under line.After being predicted respectively using two kinds of weights, according to point on line
The weighting that number carries out, selects optimal time gap weights omegaA,d。
At this point, (2) formula can be improved to:
Further, in the specific implementation, in above-mentioned shared bicycle Demand Forecast provided in an embodiment of the present invention
In method, due to consideration that day has larger impact for the demand for sharing bicycle, by observation to data and to asking
The analysis of topic, find many stakes borrows vehicle and measurer of returning the car to have apparent periodicity, and piles with different is periodically different, such as
Borrowed when the stake near middle school and government bodies, weekend and festivals or holidays/amount of returning the car can be far below working day;And for street of lying fallow
Or shopping place, situation is then complete on the contrary, step S103, which is found out, influences the factor weight of the shared bicycle demand simultaneously
The regression model is incorporated, specifically can also include:It borrows car data when multiple and goes back car data and have periodically, and is different
When the periodically difference of stake, sunlight day weighting function is designed;When the sunlight day and period to be measured in training time section
When one sunlight day is with sunlight day or the working day of being all or weekend, the return value of the sunlight day weighting function is larger, otherwise return value
It is smaller;According to the return value of the sunlight day weighting function, sunlight day weight is determined;Export the loss function, the time gap
The minimum value of the product of weight and the sunlight day weight.The possible period can be given on the basis of priori in this way, allow mould
Type can also go to capture cyclical trend in the case where low volume data collection.
In practical applications, devising sunlight day weighting function is υ (ψ1,ψ2);Wherein, ψ1For one in training time section
Sunlight day;ψ2For the sunlight day of period to be measured;υ is sunlight day weighting function.Satisfaction works as ψ1, ψ2It for same sunlight day, or is all working day
Or when weekend, the return value of the function is larger, and otherwise return value is smaller.The concrete form of υ has been determined by verifying collection under line.
At this point, (2) be improved further for:
Wherein, μ indicates one day of period to be measured;ψ (d) indicates d days sunlight days (i.e. what day), ψ ∈ [0,6].
In addition, also have a great impact since rainy, festivals or holidays etc. choose whether selection cycling trip for people, model
From actual scene, it may be considered that weather, the factor of the more various dimensions such as festivals or holidays find out weather coefficient, holiday factor etc.,
Constantly multiplied by weather coefficient, holiday factor on the basis of original model, this will not be repeated here.
Further, in the specific implementation, in above-mentioned shared bicycle Demand Forecast side provided in an embodiment of the present invention
In method, step S104, which is determined, shares bicycle in the prediction result of each demand in the period to be measured, specifically can wrap
It includes:Bicycle will be shared in set period of time car data and goes back car data as new training set in multiple borrow;According to new
Training set, find out new loss function and each sunlight day corresponding sunlight day coefficient in the period to be measured;By the new loss
Function, the time gap weight coefficient and the minimum value of the product of the sunlight day weight coefficient are Japanese multiplied by the sunlight found out
Number obtains the prediction result that demand of the bicycle at each is shared in the period to be measured.
In practical applications, the problem of formula (7) is not high to the producing level of training set, when predicting some day, Bu Nengjin
With each training set.And the characteristics of time series problem is that time samples are less, thus training set producing level is extremely important.
For this purpose, having done following improvement again:The new training set after excluding the abnormal factors such as festivals or holidays, rainfall is chosen as new instruction
Practice period (such as set period of time in May, 2015 to August).It over this time period, is each sunlight day to calculate a sunlight day
Coefficient, i.e.,:
Wherein, hA,ψIndicate that stake A averagely borrows vehicle amount, γ on sunlight day ψA,ψH after indicating normalizedA,ψ, for making
For the sunlight day coefficient on stake A sunlight day ψ;D' indicates the date set of the second training time section, D'ψIt indicates to belong in second time period
The subset of sunlight day ψ;ψ'∈[0,6].
It is possible thereby to exclude cyclic component in training set, then multiply again when predicting the period to be measured
It returns, i.e.,:
Based on the same inventive concept, the embodiment of the invention also provides a kind of shared bicycle Demand Forecast device, by
The principle and a kind of aforementioned shared bicycle Demand Forecast method that bicycle Demand Forecast device solves the problems, such as are shared in this
It is similar, therefore the implementation of the shared bicycle Demand Forecast device may refer to the reality of shared bicycle Demand Forecast method
It applies, overlaps will not be repeated.
In the specific implementation, shared bicycle Demand Forecast device provided in an embodiment of the present invention, as shown in Fig. 2, tool
Body includes:
Data acquisition module 11 shares bicycle and car data and returns the car in multiple borrow for obtaining in set period of time
Data;The set period of time is divided into training time section and verification time section;
Model building module 12, the data for obtaining the training time section under online pass through institute as training set
The building that training set carries out regression model is stated, and the data that the verification time section obtains are collected as verifying, by described
Verifying collection carries out the determination of the Parameters in Regression Model;
Weight calculation module 13 influences the factor weight of the shared bicycle demand and incorporates described return for finding out
Return model, obtains time series Weight Regression Model;
Demand Forecast module 14 is shared voluntarily for being determined in the period to be measured by the time series weighted model
Prediction result of the vehicle in each demand.
In above-mentioned shared bicycle Demand Forecast device provided in an embodiment of the present invention, aforementioned four mould can be passed through
The interaction of block does not need a large amount of data, can solve the problem that time series problem data amount is few, dimension is low, only needs
One baseline and basic data distribution are provided, and can sufficiently be needed simultaneously in view of the shared bicycle of a variety of influences
The factor for the amount of asking can constantly add influence factor in need of consideration on the basis of original model, have preferable generalization.
Corresponding contents disclosed in previous embodiment can be referred to about the more specifical course of work of above-mentioned modules,
This is no longer repeated.
Correspondingly, the embodiment of the invention also discloses a kind of shared bicycle Demand Forecast equipment, including processor and
Memory;Wherein, it realizes when processor executes the computer program saved in memory and is shared voluntarily disclosed in previous embodiment
Vehicle Demand Forecast method.
It can be with reference to corresponding contents disclosed in previous embodiment, herein no longer about the more specifical process of the above method
It is repeated.
Further, the invention also discloses a kind of computer readable storage mediums, for storing computer program;It calculates
Machine program realizes aforementioned disclosed shared bicycle Demand Forecast method when being executed by processor.
It can be with reference to corresponding contents disclosed in previous embodiment, herein no longer about the more specifical process of the above method
It is repeated.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with it is other
The difference of embodiment, same or similar part may refer to each other between each embodiment.For being filled disclosed in embodiment
It sets, for equipment, storage medium, since it is corresponded to the methods disclosed in the examples, so be described relatively simple, correlation
Place is referring to method part illustration.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond scope of the present application.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
A kind of shared bicycle Demand Forecast method, apparatus, equipment and storage medium provided in an embodiment of the present invention, should
Method includes:It obtains and shares bicycle in set period of time and borrow car data at multiple and go back car data;The setting time
Section is divided into training time section and verification time section;The data for obtaining the training time section under online pass through as training set
The training set carries out the building of regression model, and the data that the verification time section obtains are collected as verifying, passes through institute
State the determination that verifying collection carries out the Parameters in Regression Model;It finds out the factor weight for influencing the shared bicycle demand and melts
Enter the regression model, obtains time series Weight Regression Model;The time to be measured is determined by the time series weighted model
Bicycle is shared in section in the prediction result of each demand.The time series Weight Regression Model that the present invention establishes is not required to
A large amount of data are wanted, can solve that time series problem data amount is few, dimension is low and unstable, Nonlinear Time Series asks
Topic need to only provide a baseline and basic data distribution, and can be sufficiently shared in view of a variety of influences simultaneously
The factor of bicycle demand can constantly add influence factor in need of consideration on the basis of original model, have preferable
Generalization.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning
Covering non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes that
A little elements, but also including other elements that are not explicitly listed, or further include for this process, method, article or
The intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence "including a ...", is not arranged
Except there is also other identical elements in the process, method, article or apparatus that includes the element.
Shared bicycle Demand Forecast method, apparatus provided by the present invention, equipment and storage medium are carried out above
It is discussed in detail, used herein a specific example illustrates the principle and implementation of the invention, above embodiments
Explanation be merely used to help understand method and its core concept of the invention;At the same time, for those skilled in the art,
According to the thought of the present invention, there will be changes in the specific implementation manner and application range, in conclusion in this specification
Appearance should not be construed as limiting the invention.
Claims (8)
1. a kind of shared bicycle Demand Forecast method, which is characterized in that including:
It obtains and shares bicycle in set period of time and borrow car data at multiple and go back car data;The set period of time is divided into
Training time section and verification time section;
The data for obtaining the training time section under online carry out the structure of regression model by the training set as training set
It builds, and the data that the verification time section obtains is collected as verifying, the regression model ginseng is carried out by verifying collection
Several determinations;
It finds out the factor weight for influencing the shared bicycle demand and incorporates the regression model, show that time series weights
Regression model;
The prediction that demand of the bicycle at each is shared in the period to be measured is determined by the time series weighted model
As a result.
2. shared bicycle Demand Forecast method according to claim 1, which is characterized in that by the training set into
The building of row regression model, specifically includes:
The data obtained by the training time section, define loss function;
The minimum number of the loss function is found as building regression model.
3. shared bicycle Demand Forecast method according to claim 2, which is characterized in that finding out influences described share
The factor weight of bicycle demand simultaneously incorporates the regression model, specifically includes:
Reciprocal, two kinds of forms of negative exponent time gap weighting functions are separately designed to be predicted;
According to the score on line, optimal time gap weight is selected;
Export the minimum value of the loss function with the product for the time gap weight picked out.
4. shared bicycle Demand Forecast method according to claim 3, which is characterized in that finding out influences described share
The factor weight of bicycle demand simultaneously incorporates the regression model, specifically further includes:
When multiple have periodically by means of car data with car data is gone back, and when the periodically difference of piles with different, sunlight day power is designed
Weight function;
When the sunlight day of a sunlight day and period to be measured in training time section be with sunlight day or the working day of being all or
When weekend, the return value of the sunlight day weighting function is larger, and otherwise return value is smaller;
According to the return value of the sunlight day weighting function, sunlight day weight is determined;
Export the minimum value of the product of the loss function, the time gap weight and the sunlight day weight.
5. shared bicycle Demand Forecast method according to claim 4, which is characterized in that determine in the period to be measured
Bicycle is shared in the prediction result of each demand, is specifically included:
Bicycle will be shared in set period of time car data and goes back car data as new training set in multiple borrow;
According to new training set, new loss function and each sunlight day corresponding sunlight day coefficient in the period to be measured are found out;
By the minimum value of the product of the new loss function, the time gap weight coefficient and the sunlight day weight coefficient multiplied by
The sunlight day coefficient found out obtains the prediction result that demand of the bicycle at each is shared in the period to be measured.
6. a kind of shared bicycle Demand Forecast device, which is characterized in that including:
Data acquisition module shares bicycle and borrows car data at multiple and go back car data for obtaining in set period of time;
The set period of time is divided into training time section and verification time section;
Model building module, the data for obtaining the training time section under online pass through the training as training set
Collection carries out the building of regression model, and the data that the verification time section obtains are collected as verifying, is collected by the verifying
Carry out the determination of the Parameters in Regression Model;
Weight calculation module influences the factor weight of the shared bicycle demand and incorporates the recurrence mould for finding out
Type obtains time series Weight Regression Model;
Demand Forecast module shares bicycle every for determining by the time series weighted model in the period to be measured
The prediction result of a demand.
7. a kind of shared bicycle Demand Forecast equipment, which is characterized in that including processor and memory, wherein the place
Reason device is realized when executing the computer program saved in the memory to be shared voluntarily as described in any one of claim 1 to 5
Vehicle Demand Forecast method.
8. a kind of computer readable storage medium, which is characterized in that for storing computer program, wherein the computer journey
Such as shared bicycle Demand Forecast method described in any one of claim 1 to 5 is realized when sequence is executed by processor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810804908.6A CN108876056A (en) | 2018-07-20 | 2018-07-20 | A kind of shared bicycle Demand Forecast method, apparatus, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810804908.6A CN108876056A (en) | 2018-07-20 | 2018-07-20 | A kind of shared bicycle Demand Forecast method, apparatus, equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108876056A true CN108876056A (en) | 2018-11-23 |
Family
ID=64303742
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810804908.6A Pending CN108876056A (en) | 2018-07-20 | 2018-07-20 | A kind of shared bicycle Demand Forecast method, apparatus, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108876056A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109840272A (en) * | 2018-12-19 | 2019-06-04 | 东南大学 | A kind of shared electric car site users needing forecasting method |
CN110543699A (en) * | 2019-08-15 | 2019-12-06 | 阿里巴巴集团控股有限公司 | shared vehicle travel data simulation and shared vehicle scheduling method, device and equipment |
CN110807921A (en) * | 2019-10-24 | 2020-02-18 | 上海钧正网络科技有限公司 | Vehicle scheduling method, device, equipment and storage medium |
CN111292106A (en) * | 2018-12-06 | 2020-06-16 | 北京嘀嘀无限科技发展有限公司 | Method and device for determining business demand influence factors |
CN111832600A (en) * | 2019-12-31 | 2020-10-27 | 北京骑胜科技有限公司 | Data processing method and device, electronic equipment and computer readable storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130321178A1 (en) * | 2012-05-29 | 2013-12-05 | Akhtar Jameel | Shared vehicle rental system including transmission of reservation information and targeted advertising |
CN104636828A (en) * | 2015-02-02 | 2015-05-20 | 西南交通大学 | Public bicycle station supply and demand prediction method based on Markov chain |
CN104778508A (en) * | 2015-04-02 | 2015-07-15 | 杭州电子科技大学 | Public bicycle renting forecasting method based on multi-source data fusion |
CN107067283A (en) * | 2017-04-21 | 2017-08-18 | 重庆邮电大学 | Based on history merchant record and the electric business of user behavior consumption passenger flow forecast method |
CN107767659A (en) * | 2017-10-13 | 2018-03-06 | 东南大学 | Shared bicycle traffic attraction and prediction of emergence size method based on ARIMA models |
-
2018
- 2018-07-20 CN CN201810804908.6A patent/CN108876056A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130321178A1 (en) * | 2012-05-29 | 2013-12-05 | Akhtar Jameel | Shared vehicle rental system including transmission of reservation information and targeted advertising |
CN104636828A (en) * | 2015-02-02 | 2015-05-20 | 西南交通大学 | Public bicycle station supply and demand prediction method based on Markov chain |
CN104778508A (en) * | 2015-04-02 | 2015-07-15 | 杭州电子科技大学 | Public bicycle renting forecasting method based on multi-source data fusion |
CN107067283A (en) * | 2017-04-21 | 2017-08-18 | 重庆邮电大学 | Based on history merchant record and the electric business of user behavior consumption passenger flow forecast method |
CN107767659A (en) * | 2017-10-13 | 2018-03-06 | 东南大学 | Shared bicycle traffic attraction and prediction of emergence size method based on ARIMA models |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111292106A (en) * | 2018-12-06 | 2020-06-16 | 北京嘀嘀无限科技发展有限公司 | Method and device for determining business demand influence factors |
CN109840272A (en) * | 2018-12-19 | 2019-06-04 | 东南大学 | A kind of shared electric car site users needing forecasting method |
CN109840272B (en) * | 2018-12-19 | 2023-10-27 | 东南大学 | Method for predicting user demand of shared electric automobile station |
CN110543699A (en) * | 2019-08-15 | 2019-12-06 | 阿里巴巴集团控股有限公司 | shared vehicle travel data simulation and shared vehicle scheduling method, device and equipment |
CN110543699B (en) * | 2019-08-15 | 2023-06-13 | 创新先进技术有限公司 | Shared vehicle travel data simulation and shared vehicle scheduling method, device and equipment |
CN110807921A (en) * | 2019-10-24 | 2020-02-18 | 上海钧正网络科技有限公司 | Vehicle scheduling method, device, equipment and storage medium |
CN111832600A (en) * | 2019-12-31 | 2020-10-27 | 北京骑胜科技有限公司 | Data processing method and device, electronic equipment and computer readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108876056A (en) | A kind of shared bicycle Demand Forecast method, apparatus, equipment and storage medium | |
Selby et al. | Spatial prediction of traffic levels in unmeasured locations: applications of universal kriging and geographically weighted regression | |
CN110517479B (en) | Urban road traffic prediction method and device and electronic equipment | |
CN109376935B (en) | Bus passenger flow time interval combined prediction method based on neural network | |
CN107832866A (en) | A kind of method for predicting and device, storage medium, terminal | |
CN110517492A (en) | Based on the traffic route recommended method of parallel integrated study, system, device | |
Yao et al. | Hybrid methods for short-term traffic flow prediction based on ARIMA-GARCH model and wavelet neural network | |
CN105405293A (en) | Short-term prediction method of road travel time and system | |
Li et al. | Deep learning based parking prediction on cloud platform | |
CN110956807A (en) | Highway flow prediction method based on combination of multi-source data and sliding window | |
Kim et al. | Determining optimal sensor locations in freeway using genetic algorithm-based optimization | |
CN112419711B (en) | Closed parking lot parking demand prediction method based on improved GMDH algorithm | |
CN112069635B (en) | Method and device for deploying battery changing cabinet, medium and electronic equipment | |
CN110459050A (en) | A kind of short-term bus passenger flow prediction technique based on hybrid decision tree | |
CN105303838B (en) | The method and apparatus for determining vehicle flow | |
Dai et al. | Assessing spatiotemporal bikeability using multi-source geospatial big data: A case study of Xiamen, China | |
Klunder et al. | Improvement of Network Performance by In‐Vehicle Routing Using Floating Car Data | |
Arnesen et al. | A forward Markov model for predicting bicycle speed | |
CN111489039B (en) | Method and system for predicting total quantity of shared bicycle | |
Pal et al. | Analyzing mobility patterns and imbalance of free floating bike sharing systems | |
Cirianni et al. | Road traffic noise prediction models in the metropolitan area of the Strait of Messina | |
CN105355038A (en) | Method for predicting short-term traffic flow through employing PMA modeling | |
Carteni et al. | Travel time cost functions for urban roads: A case study in Italy | |
Karunanithi et al. | A predictive analytics approach for forecasting bike rental demand | |
CN113611106A (en) | Road traffic condition prediction method, device and equipment based on smart city |
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: 20181123 |