CN109345296A - Common people's Travel Demand Forecasting method, apparatus and terminal - Google Patents
Common people's Travel Demand Forecasting method, apparatus and terminal Download PDFInfo
- Publication number
- CN109345296A CN109345296A CN201811100013.0A CN201811100013A CN109345296A CN 109345296 A CN109345296 A CN 109345296A CN 201811100013 A CN201811100013 A CN 201811100013A CN 109345296 A CN109345296 A CN 109345296A
- Authority
- CN
- China
- Prior art keywords
- region
- common people
- demand forecasting
- travelling
- travel demand
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention discloses a kind of common people's Travel Demand Forecasting method, apparatus, terminal and computer readable storage mediums, the described method includes: obtaining the association attributes at least one region, and obtain common people's trip sample data of corresponding region, wherein, it is described trip sample data include the region association attributes and corresponding common people's travelling OD information；Construct Travel Demand Forecasting model neural network based；The associated property data in region to be measured is inputted into the Travel Demand Forecasting model in advance to survey common people's travelling OD information in region to be measured.The present invention is by obtaining the trip sample data of different zones to carry out modeling analysis and the association attributes in region are inputted parameter as it, it can be achieved to carry out Accurate Prediction to common people's travelling OD information in region to be measured, can also provide reliable data reference etc. in advance for traffic transport power etc..
Description
Technical field
The present invention relates to field of intelligent transportation technology more particularly to a kind of common people's Travel Demand Forecasting method, apparatus and meters
Calculation machine terminal.
Background technique
Resident trip is that it completes social property and the comprehensive sexual behaviour of personal attribute, and people need to complete work by trip
Make, communication and the social properties such as social activities, in different places shopping, amusement and self-indulged to realize its personal attribute.It is logical
It crosses the analysis to resident trip data and establishes corresponding travel behaviour model, realize the Accurate Prediction to resident trip demand,
Facilitate traffic control department and transit agency and adjust vasodilator effect and transport power arrangement in time, alleviate city area-traffic pressure and
Situation is detained in the aggregation of website passenger flow.
However, the traffic circulation in current cities and towns is in particular point in time such as peak period on and off duty, festivals or holidays and occasions
It is easy phenomena such as traffic congestion occurs in corresponding region or crowd massing is detained.These problems are largely derived to the common people
The expected inaccuracy of trip requirements, therefore the trip requirements under reply deadly condition are tended not to conventional transport power arrangement.
Summary of the invention
In view of the above problems, the present invention proposes a kind of common people's Travel Demand Forecasting method, by obtaining in different zones
Sample data of going on a journey carries out modeling analysis and inputs parameter for a variety of association attributes in region as it, and realization goes out the region
Row demand carries out Accurate Prediction, can solve the problems, such as that forecasting inaccuracy is true etc. in the prior art.
The embodiment of the present invention proposes a kind of common people's Travel Demand Forecasting method, comprising:
The association attributes at least one region are obtained, and obtain common people's trip sample data of corresponding region, wherein is described
Trip sample data include the region association attributes and corresponding common people's travelling OD information；
Construct Travel Demand Forecasting model neural network based；
Wherein, the building includes: using the association attributes in the region as input parameter with the determination neural network
Input layer；Node layer is exported using common people's travelling OD information in the region as output parameter to determine；It determines implicit
Node layer number；It is trained and verifies using the trip sample data after normalized；
The associated property data in region to be measured is inputted into the Travel Demand Forecasting model to predict the region to be measured
Common people's travelling OD information.
In above-mentioned common people's Travel Demand Forecasting method, optionally, the people in the trip sample data in each region
The acquisition of many travelling OD information, comprising:
Obtain the identifier of the mobile communication equipment in the region；
When the mobile communication equipment reaches the region or leaves the region, the corresponding shifting of the identifier is obtained
Dynamic position data of the communication equipment in preset period of time；
The travelling OD information of the common people in the region is obtained from the position data.
In above-mentioned common people's Travel Demand Forecasting method, optionally, the travelling OD packet includes row destination, sets out
Trip mode between ground and dwell point, it is described " the travelling OD information of the common people in the region is obtained from the position data ",
Include:
The speed of service of the mobile communication equipment is obtained according to the position data and comprising the travel path of dwell point；
Corresponding trip mode is judged according to the speed of service between the dwell point；
With judging the trip purpose and departure place according to the travel path.
In above-mentioned common people's Travel Demand Forecasting method, optionally, the dwell point includes short dwell point and long dwell point,
The judgement of the dwell point includes:
If the mobile communication equipment is greater than the first dwell time values in same position residence time and is no more than second
Dwell time values then determine that current location is a short dwell point；
If the mobile communication equipment is greater than the second dwell time values in the same position residence time, determine described current
Position is a long dwell point.
It is optionally, described " to utilize the trip sample after normalized in above-mentioned common people's Travel Demand Forecasting method
Notebook data is trained and verifies ", comprising:
The data of preset ratio are randomly selected from all trip sample datas after normalized as training data,
Remaining data are as test data；
During the training neural network, the transmission function of forward-propagating is all made of sigmoid function, backpropagation
Error function use sum of squared errors function.
In above-mentioned common people's Travel Demand Forecasting method, optionally, using gradient descent algorithm determination and common people's travelling OD
Information output valve each connection weight corresponding with the minimal error of actual common people's travelling OD information actual value and threshold value.
In above-mentioned common people's Travel Demand Forecasting method, optionally, the input layer of the Travel Demand Forecasting model
Number range is 10~14, and output layer interstitial content range is 3~5, and hidden layer node number range is 5~10.
According to above-mentioned common people's Travel Demand Forecasting method, the embodiment of the present invention also proposes a kind of common people's Travel Demand Forecasting
Device, comprising:
Sample data obtains module, for obtaining the association attributes at least one region, and obtains the common people of corresponding region
Trip sample data, wherein the trip sample data includes the association attributes and corresponding common people's travelling OD letter in the region
Breath；
Model construction module, for constructing Travel Demand Forecasting model neural network based；
Wherein, the building includes: using the association attributes in the region as input parameter with the determination neural network
Input layer；Node layer is exported using common people's travelling OD information in the region as output parameter to determine；It determines implicit
Node layer number；It is trained and verifies using the trip sample data after normalized；
Model prediction module, for the associated property data in region to be measured to be inputted the Travel Demand Forecasting model with pre-
Survey common people's travelling OD information in the region to be measured.
Another embodiment of the present invention proposes that a kind of terminal, the terminal include memory and processing
Device, the memory run the computer program so that the computer is whole for storing computer program, the processor
End executes above-mentioned common people's Travel Demand Forecasting method.
Another embodiment of the present invention also proposes a kind of computer readable storage medium, is stored with above-mentioned terminal
Used in computer program.
Technical solution of the present invention has the following beneficial effects:
The embodiment of the present invention, which passes through, obtains the corresponding common people's trip sample data of different zones, and these regions are corresponding
Input parameter when a variety of association attributes progress common people's Travel Demand Forecasting model constructions is, it can be achieved that go out the common people in region to be measured
Row OD information carries out Accurate Prediction, in addition, also reliable data reference can be provided in advance for traffic transport power etc., and then can prevent
Phenomena such as occurring crowd's congestion in the region place or being detained.
Detailed description of the invention
In order to illustrate more clearly of technical solution of the present invention, letter will be made to attached drawing needed in the embodiment below
It singly introduces, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as to the present invention
The restriction of protection scope.
Fig. 1 is the flow diagram of common people's Travel Demand Forecasting method of the embodiment of the present invention；
Fig. 2 is that one group of trip sample data constitutes schematic diagram；
Fig. 3 is the first-class of common people's trip sample data acquisition of common people's Travel Demand Forecasting method of the embodiment of the present invention
Journey schematic diagram；
Fig. 4 is the second that common people's trip sample data of common people's Travel Demand Forecasting method of the embodiment of the present invention obtains
Journey schematic diagram；
Fig. 5 is that the neural network of common people's Travel Demand Forecasting method of the embodiment of the present invention constructs schematic diagram；
Fig. 6 is the structural schematic diagram of common people's Travel Demand Forecasting device of the embodiment of the present invention.
Main element symbol description:
10- common people's Travel Demand Forecasting device；100- sample data obtains module；200- model construction module；300- mould
Type prediction module.
Specific embodiment
Below in conjunction with attached drawing in the embodiment of the present invention, technical solution in the embodiment of the present invention carries out clear, complete
Ground description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Usually exist
The component of the embodiment of the present invention described and illustrated in attached drawing can be arranged and be designed with a variety of different configurations herein.Cause
This, is not intended to limit claimed invention to the detailed description of the embodiment of the present invention provided in the accompanying drawings below
Range, but it is merely representative of selected embodiment of the invention.Based on the embodiment of the present invention, those skilled in the art are not doing
Every other embodiment obtained under the premise of creative work out, shall fall within the protection scope of the present invention.
In addition, term " first ", " second " are used for descriptive purposes only and cannot be understood as indicating or suggesting relative importance
Or implicitly indicate the quantity of indicated technical characteristic.Define " first " as a result, the feature of " second " can be expressed or
Implicitly include one or more of the features.In the description of the present invention, the meaning of " plurality " is two or more,
Unless otherwise specifically defined.
Unless otherwise defined, all technical and scientific terms used herein and belong to technical field of the invention
The normally understood meaning of technical staff is identical.Used term is intended merely to describe specifically to implement in the description herein
The purpose of example, it is not intended that the limitation present invention.Term " and or " used herein includes one or more relevant institute's lists
Any and all combinations of purpose.
When in view of some regional fields places such as festivals or holidays or holding occasion particular point in time, it is easy to appear traffic
Phenomena such as congestion or crowd massing are detained, this is mostly derived from the expected inaccuracy to common people's trip requirements in the region place, because
This, the present invention proposes a kind of common people's Travel Demand Forecasting method, by carrying out association attributes description, and root to different regional fields
Analysis modeling is carried out according to these attributive character and corresponding common people's trip data in these regions, to obtain a trip requirements
Prediction model, as long as therefore input a region to be measured association attributes, that is, can be predicted the region to be measured the common people's travelling OD letter
Breath.
Below with reference to specific embodiment, the present invention is described in detail.
Embodiment 1
Fig. 1 is please referred to, the present embodiment proposes a kind of common people's Travel Demand Forecasting method, can be applied to some region places
The common people go on a journey prediction, the common people's Travel Demand Forecasting model obtained through this embodiment can to different zones the common people's travelling OD letter
Breath carries out Accurate Prediction.Common people's Travel Demand Forecasting method is specifically described below.
Common people's Travel Demand Forecasting method of the present embodiment mainly comprises the steps that
Step S100: obtaining the association attributes at least one region, and obtains common people's trip sample data of corresponding region,
Wherein, it is described trip sample data include the region association attributes and corresponding common people's travelling OD information.
In view of different region places, association attributes are not often identical, such as the activity schedule situation in the region, gas
Image data and area attribute etc. cause its corresponding common people's trip situation generally can also have biggish difference.In the present embodiment,
By the common people that association attributes to multiple and different regions carry out feature description and obtain corresponding region go on a journey sample data so as to
Carry out subsequent analysis modeling.By using the association attributes in region as the later period in the input parameter of analysis modeling, can be more
Accurately, comprehensively the region is analyzed, and then the forecasting accuracy to common people's trip situation in the region can be improved.It can
To understand, the region of the present embodiment be can be, but not limited to as a market, public place of entertainment, industrial park, office building or hotel etc..
In the present embodiment, the association attributes in the region may include that the activity schedule situation in the region, the common people go on a journey period, pre-
Survey the meteorological data of period and the area attribute etc. in the region.
Feature description for above-mentioned association attributes, the activity schedule situation may include having activity or two kinds of feelings of non-activity
Condition.Further, if there is activity, which can also include activity scale, particularly may be divided into large-scale, medium-sized and small-sized.Institute
Time segments division can be carried out the trip period as unit of a time interval by stating the common people, in order to model analysis and be predicted in one day
Flow of the people in which more etc. period.Exemplarily, if carrying out Time segments division with 1 hour for unit, integral point is node, then one
It can divide 24 periods, such as 11:00-12:00 is one of those period.The meteorological data mainly includes the corresponding period
Weather condition, the weather pattern can be divided into it is fine, negative and three kinds of rain, optionally, it is contemplated that Changes in weather, in a period at most
It may include two kinds of weather patterns.Further, if fine, which can also include whether high temperature, for example, if temperature is super
30 degree are crossed, then regards as high temperature, otherwise regards as non high temperature weather.Further, if rain, which can also include
Rainfall degree can be divided into heavy rain, moderate rain and light rain, optionally, it is contemplated that Changes in weather at most may include two kinds in one period
Rainfall degree.
Further, it includes the industry attribute in the region that the area attribute in the region, which specifically may include but be not limited to, is advised
Mould size, skyscraper attribute, construction area, bus station quantity and subway station quantity etc..Wherein, the industry attribute
It is mainly used for indicating which industry the region belongs to, such as, it may include industrial park, office building, market, hotel, food and drink amusement
Place, transport hub, educational institution, medical institutions etc..The scale may include large-scale, medium-sized and three kinds small-sized.Institute
Stating skyscraper attribute then may include high level and non-high-rise two kinds, wherein the definition of skyscraper can be selected according to the actual situation
Take, for example, when a certain region place number of floor levels be more than 8 layers then its belong to skyscraper.The construction area can mainly include
Greatly, neutralization is three kinds small, and specific area division can be chosen according to the actual situation.
Wherein, feature descriptions are carried out to a variety of attributes such as the activity schedule situation in these regions, meteorological data, it is contemplated that arrive
Some such as markets, public place of entertainment zone of action place is in festivals or holidays or its flow of the people is often larger when holding activity, Huo Zhe
Its flow of the people often special circumstances such as less, are carried out more by the trip requirements to different zones in the case that weather is relatively severe
Add comprehensively effectively analysis, and then the accuracy to common people's Travel Demand Forecasting of corresponding region can be improved.
In the present embodiment, the corresponding common people in each region sample data of going on a journey mainly includes two large divisions, and a part is this
The information of the association attributes in region, another part are the corresponding common people's travelling OD information in the region, wherein O, that is, ORIGIN, it is indicated that
Capable departure place, D, that is, DESTINATION, it is indicated that capable destination.In the present embodiment, which mainly includes
Trip purpose, departure place and corresponding trip mode etc..Exemplarily, it is assumed that get the multiple groups trip sample number of region A
According to by taking wherein one group of sample data as an example, then this group of sample data can carry out field description as shown in Figure 2, in modeling
These fields can be indicated with corresponding data to carry out Function Fitting etc. according to the rule of agreement.For example, with destination sample
For notebook data, different destinations can be described with the data of presetting digit capacity, and the data of the presetting digit capacity specifically can be used about
Fixed digit indicates the geographical location information (such as longitude and latitude) and correlation attribute information (such as industry attribute) of the destination.
Wherein, as shown in figure 3, the acquisition of common people's travelling OD information in the trip sample data of corresponding region mainly includes following
Step:
Step S110: the identifier of the mobile communication equipment in region is obtained.
Step S120: when the mobile communication equipment reaches the region or leaves the region, the identification is obtained
Number position data of the corresponding mobile communication equipment in preset period of time.
Under the premise of licensed-in, the identifier of the mobile communication equipment of the sufficient amount of common people can be obtained, for example, can
Obtain the identifier for being linked into the equipment of common carrier, Internet application etc., wherein the identifier of an equipment will represent one
A object.Then, when the mobile communication equipment reaches the region or leaves the region, then it can obtain the mobile communication equipment and exist
Position data in preset period of time.
It should be appreciated that the preset period of time is to reach the region or time point from the region is with mobile communication equipment
Benchmark and a period of time chosen.Specifically, a longer period can be chosen, for example, when mobile communication equipment should in arrival
When region, then the position data in 2 hours before starting to be chosen at arrival time point, or when mobile communication equipment leaves the area
The position data etc. in 2 hours when domain, then after starting to be chosen at time departure point.It in this way can be more accurate convenient for obtaining
The trip purpose of the common people, the travelling ODs information such as departure place.It is main to consider generally common people's daily trip, it is preferable that described pre-
If the period can be chosen 2~3 hours.
Step S130: the travelling OD information of the correspondence common people in the region is obtained from the position data.
In the present embodiment, the travelling OD information may include trip purpose, the trip mode between departure place and dwell point etc..
As shown in figure 4, step S130 mainly includes following sub-step again:
Sub-step S131: according to the speed of service of position data acquisition mobile communication equipment and the stroke rail comprising dwell point
Mark.
Sub-step S132: corresponding trip mode is judged according to the speed of service between dwell point.
Sub-step S133: trip destination and departure place are judged according to travel path.
Exemplarily, its operation speed can be calculated according to the change in displacement of the mobile communication equipment within a preset time interval
Degree.It is appreciated that the prefixed time interval is smaller, then the speed of service of the mobile communication equipment got will be more acurrate,
Travel path is also more accurate.Preferably, which can be chosen for 30s~1min.
Optionally, when obtaining its travel path, which can also be compared with trip big data, example
Such as, can be by carrying out comparing with the stroke time-consuming in common trip software, and then judge that the common people's is closest
Trip mode.
Wherein, when obtaining its travel path, the dwell point information in preset period of time can be obtained, further to judge
The common people specifically can be according to the fortune between two dwell points with the presence or absence of different trip modes between two different dwell points
The corresponding trip mode of row velocity estimated, so as to obtain the travelling OD information of the more accurate common people.
Optionally, which may include short dwell point and long dwell point.Specifically, if the mobile communication equipment is same
Position residence time is greater than the first dwell time values and is no more than the second dwell time values, then determines that current location is one short to stop
Stationary point；If the mobile communication equipment is greater than the second dwell time values in the same position residence time, determine that the position is one
Long dwell point.It is appreciated that first dwell time values should be greater than the second dwell time values, for example, first residence time
Value can choose 5min~10min, and the second dwell time values can choose 30min.In addition, after being confirmed as dwell point, it can be into one
Step obtains the attribute of corresponding dwell point, such as industry attribute.
Still optionally further, it is contemplated that may have part way that can have the case where traffic traffic congestion, stop getting these
Behind stationary point, can also further exclude whether be traffic congestion when stop.Exemplarily, can according to the geographical location of the dwell point and
Traffic information of the dwell point etc. is excluded.If traffic congestion when stop, then can travel path describe during will be right
It is that dwell point marks that the dwell point answered, which is cancelled,.In addition, the case where for having no resting state on the way in the preset period of time, then should
The trip mode of two dwell points can trip mode between the region and destination or departure place.
Step S200: Travel Demand Forecasting model neural network based is constructed.Wherein, above-mentioned trip sample can be passed through
Data carry out neural network learning training, to obtain the Travel Demand Forecasting model.As shown in figure 5, the prediction model is main
Including input layer, hidden layer and output layer, the building of the model mainly includes following sub-step.
Sub-step S210: the association attributes in region are determined to the input layer of neural network as input parameter；It will
Common people's travelling OD information in region exports node layer as output parameter to determine；Determine hidden layer node number.
In the present embodiment, each attributive character in the association attributes in region is used as to the input layer section of the neural network
Point by the activity schedule situation of chosen area, activity scale, weather pattern, goes out according to the actual conditions that the common people in region go on a journey
The input variable of row period and industry attribute as the neural network.Further, neighbouring bus station's points can also be increased
Amount, subway station quantity, whether high temperature, rainfall degree, scale, whether the influence factors such as skyscraper and construction area make
For input variable.
It is as needed using the corresponding common people's travelling OD information characteristics in region as output variable for exporting node layer
Common people's trip requirements in the region of solution, the trip mode that can be selected between row destination, departure place and dwell point become as output
Amount.Further, inflow crowd amount and outflow crowd's amount in prediction period etc. can also be increased as output variable, so as to detailed
That carefully recognizes the region predicts market condition.
For hidden layer node number, generally, if the implicit number of plies is more, the result of training is more accurate, but the time
It consumes also more.In the present embodiment, the number of plies of hidden layer is 1, in order to protect between time efficiency and the balance of the accuracy of result
It is optimal to hold strategy, formula will be utilizedTo choose the interstitial content of the hidden layer, wherein h is hidden layer node
Number, m are input layer number, and n is output layer interstitial content, and a is the regulating constant between 1~10.
In the present embodiment, the input layer number range of the Travel Demand Forecasting model can be 10~14, output layer section
Point number range can be 3~5, and hidden layer node number range can be 5~10.In conjunction with actual training pattern, trip prediction
The input layer number of model is 12, and output layer interstitial content is 4, and hidden layer node number is preferably set to 7.
Sub-step S220: it is trained and verifies using the trip sample data after normalized.
In the present embodiment, before carrying out learning training using trip sample data, need to carry out the trip sample data
Normalized makes its range between [0,1].Normalizing formula isWherein, x is that the related of region belongs to
Property in a feature value, x ' be x normalized output, xminFor the minimum value of x, xmaxFor the maximum value of x.
Then, the trip sample data by these after being normalized will be divided into two parts by preset ratio, and one
It is divided into training data, for constructing the Travel Demand Forecasting model, that is, determines connection weight and threshold value in the neural network；Separately
A part is test data, for examining the Travel Demand Forecasting model.It, will be random from trip sample data in the present embodiment
The data of selection 2/3 are as training data, and the data of remainder 1/3 are as test data.
Specifically, during the training neural network, the initial weight of the network will carry out random assignment, learn in setting
After practising the relevant parameters such as error, learning rate and maximum cycle, the transmission function of forward-propagating, i.e. input layer are to hidden
The transmission function of transmission function and hidden layer to output layer containing layer is all made of the sigmoid function in nonlinear function.Into one
Step ground, the formula of the sigmoid function areWherein, x is variable.In the present embodiment, in conjunction with actual trained mould
Type sets learning error ε=5 × 10∧2, learning rate is set as η=0.2, maximum cycle 100000.
To keep the error amount for predicting output valve and actual value minimum, the net will be determined using root-mean-square deviation method is minimized
Each connection weight and threshold value of network.Specifically, the error function of backpropagation will use sum of squared errors function.Specifically,
The formula of sum of squared errors function isWherein, w is connection weight, and b is threshold value, and n is output layer
Interstitial content, j indicate j-th of output node, djIt is exported for the prediction of j-th of node in common people's Travel Demand Forecasting model
Value, yjFor j-th of corresponding sample actual value.It is successively anti-to input layer by calculating output error and it being passed through hidden layer
It passes, and error propagation is given to all nodes of each layer, from the error that each layer obtains as adjustment input node and hidden node
Connection weight and threshold value and hidden node and output node connection weight and threshold value.
In the present embodiment, it will be determined using gradient descent algorithm and common people's travelling OD information output valve and the actual common people
The corresponding each connection weight of the minimal error of travelling OD information actual value and threshold value.The gradient descent algorithm, refers to and passes through
Make above-mentioned output error along its negative gradient direction change, so that it is determined that corresponding connection weight or threshold value.
Exemplarily, as shown in figure 5, if using wijConnection weight of expression i-th of the node of input layer to j-th of node of hidden layer
Weight, δjIndicate the output of the backpropagation of j-th of node of output layer, upper right corner target 1 indicates first layer connection weight, then has the
The gradient of one layer of connection weight isThen, according to gradient descent algorithm rule, haveThe connection weight can be adjusted, thus the connection weight w after being adjustedij'.Together
It manages, the connection weight and threshold value between other nodes can be adjusted using the algorithm, and this will not be detailed here.
Then, when connection weight adjusted and threshold value can make its output error reach default error threshold or meet phase
When the accuracy rate answered, then deconditioning and determine the Travel Demand Forecasting model connection weight and threshold value.Specifically, available
Above-mentioned test data carries out the verifying of accuracy rate.
Step S300: common people's travelling OD information in the Travel Demand Forecasting model prediction region to be measured is utilized.
In the model after being trained, it can be used for predicting some region of common people's travelling OD information, for example, certain business
Area, industrial park etc..Specifically, can input region to be measured includes activity schedule situation, activity scale, weather pattern, trip
Period and industry attribute these association attributes parameters, then the corresponding common people's travelling OD information in the exportable region, can specifically wrap
With including the trip purpose of the common people in the region in the future anticipation period, the trip mode between departure place and two dwell points.Into one
Step ground, can also predict the discrepancy flow of the people in the region etc. in the future anticipation period.
Wherein, the region to be measured is either region in above-mentioned sample, is also possible to other areas not as sample
Domain.It is appreciated that the Travel Demand Forecasting model can be not limited to be served only for predicting common people's trip situation of a certain specific region, and
It is that corresponding accurate common people's trip situation can be exported according to the association attributes in different regions.
The present embodiment is collected by common people's trip data to different zones and is built using neural network algorithm
Mould training, to obtain corresponding common people's Travel Demand Forecasting model, by using the association attributes of estimation range as the defeated of modeling
Enter parameter, situation progress Accurate Prediction of going on a journey to multi-faceted effective analysis in the region and the corresponding common people may be implemented.Wherein,
When collecting these sample datas, its corresponding stroke rail can be described by the position data of the equipment to these common people groups
Mark etc., the trip mode also accounted for when describing its travel path to dwell point and stop further judges, so as to obtain more
Accurate sample data, and then improve the accuracy rate etc. of prediction.In addition, by providing reliable number in advance for traffic transport power etc.
According to reference, then traffic control department and transit agency etc. can adjust traffic power in time and arrange strategy etc., can further prevent or
Phenomena such as alleviating crowd's congestion of some regions proximal site website in time or being detained.
Embodiment 2
Fig. 6 is please referred to, 1 common people's Travel Demand Forecasting method, the present embodiment propose a kind of common people based on the above embodiment
Travel Demand Forecasting device 10, specifically, common people's Travel Demand Forecasting device 10 can include:
Sample data obtains module 100, for obtaining the association attributes at least one region, and obtains the people of corresponding region
Crowd's trip sample data, wherein the trip sample data includes the association attributes and corresponding common people's travelling OD in the region
Information.
Model construction module 200, for constructing Travel Demand Forecasting model neural network based.Wherein, the building
It include: using the association attributes in the region as input parameter with the input layer of the determination neural network；By the area
Common people's travelling OD information in domain exports node layer as output parameter to determine；Determine hidden layer node number；Using by returning
One change treated trip sample data be trained and verify.
Model prediction module 300, for common people's travelling OD using the Travel Demand Forecasting model prediction region to be measured
Information.
Above-mentioned common people's Travel Demand Forecasting device 10 corresponds to common people's Travel Demand Forecasting method of embodiment 1.Implement
Any option in example 1 is also applied for the present embodiment, and I will not elaborate.
The present invention also provides a kind of terminal, the terminal includes memory and processor, described to deposit
Reservoir runs the computer program so that the terminal perform claim for storing computer program, the processor
It is required that the function of above-mentioned common people's Travel Demand Forecasting method or the modules in above-mentioned common people's Travel Demand Forecasting device 10
Energy.
Memory may include storing program area and storage data area, wherein storing program area can storage program area, at least
Application program needed for one function；Storage data area, which can be stored, uses created data (such as sound according to mobile terminal
Frequency evidence, phone directory etc.) etc..In addition, memory may include high-speed random access memory, it can also include non-volatile deposit
Reservoir, for example, at least a disk memory, flush memory device or other volatile solid-state parts.
Also a kind of computer readable storage medium of the present invention, is stored with computer journey used in above-mentioned terminal
Sequence.
In several embodiments provided herein, it should be understood that disclosed device and method can also pass through
Other modes are realized.The apparatus embodiments described above are merely exemplary, for example, flow chart and structure in attached drawing
Figure shows the system frame in the cards of the device of multiple embodiments according to the present invention, method and computer program product
Structure, function and operation.In this regard, each box in flowchart or block diagram can represent a module, section or code
A part, a part of the module, section or code includes one or more for implementing the specified logical function
Executable instruction.
It should also be noted that function marked in the box can also be attached to be different from the implementation as replacement
The sequence marked in figure occurs.For example, two continuous boxes can actually be basically executed in parallel, they sometimes may be used
To execute in the opposite order, this depends on the function involved.It is also noted that every in structure chart and/or flow chart
The combination of a box and the box in structure chart and/or flow chart, can function or movement as defined in executing it is dedicated
Hardware based system is realized, or can be realized using a combination of dedicated hardware and computer instructions.
In addition, each functional module or unit in each embodiment of the present invention can integrate one independence of formation together
Part, be also possible to modules individualism, an independent part can also be integrated to form with two or more modules.
It, can be with if the function is realized and when sold or used as an independent product in the form of software function module
It is stored in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be intelligence
Can mobile phone, personal computer, server or network equipment etc.) execute each embodiment the method for the present invention whole or
Part steps.And storage medium above-mentioned include: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory),
Random access memory (RAM, Random Access Memory), magnetic or disk etc. be various to can store program code
Medium.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.
Claims (10)
1. a kind of common people's Travel Demand Forecasting method characterized by comprising
The association attributes at least one region are obtained, and obtain common people's trip sample data of corresponding region, wherein the trip
Sample data include the region association attributes and corresponding common people's travelling OD information；
Construct Travel Demand Forecasting model neural network based；
Wherein, the building includes: using the association attributes in the region as input parameter with the defeated of the determination neural network
Enter node layer；Node layer is exported using common people's travelling OD information in the region as output parameter to determine；Determine hidden layer section
Point number；It is trained and verifies using the trip sample data after normalized；
The associated property data in region to be measured is inputted into the Travel Demand Forecasting model to predict the common people in the region to be measured
Travelling OD information.
2. common people's Travel Demand Forecasting method according to claim 1, which is characterized in that the trip sample in each region
The acquisition of common people's travelling OD information in notebook data, comprising:
Obtain the identifier of the mobile communication equipment in the region；
When the mobile communication equipment reaches the region or leaves the region, it is logical to obtain the corresponding movement of the identifier
Believe position data of the equipment in preset period of time；
The travelling OD information of the common people in the region is obtained from the position data.
3. common people's Travel Demand Forecasting method according to claim 2, which is characterized in that the travelling OD packet includes
Trip mode between row destination, departure place and dwell point, it is described " to obtain the common people's in the region from the position data
Travelling OD information ", comprising:
The speed of service of the mobile communication equipment is obtained according to the position data and comprising the travel path of dwell point；
Corresponding trip mode is judged according to the speed of service between the dwell point；
With judging the trip purpose and departure place according to the travel path.
4. common people's Travel Demand Forecasting method according to claim 3, which is characterized in that the dwell point includes short stop
Point and long dwell point, the judgement of the dwell point include:
It is stopped if the mobile communication equipment is greater than the first dwell time values and is no more than second in same position residence time
Time value then determines that current location is a short dwell point；
If the mobile communication equipment is greater than the second dwell time values in the same position residence time, the current location is determined
For a long dwell point.
5. common people's Travel Demand Forecasting method according to claim 1, which is characterized in that described " using by normalization
Treated, and trip sample data is trained and verifies ", comprising:
The data of preset ratio are randomly selected from all trip sample datas after normalized as training data, remainder
Data as test data；
During the training neural network, the transmission function of forward-propagating is all made of sigmoid function, the mistake of backpropagation
Difference function uses sum of squared errors function.
6. common people's Travel Demand Forecasting method according to claim 5, which is characterized in that determined using gradient descent algorithm
And common people's travelling OD information output valve each connection weight corresponding with the minimal error of actual common people's travelling OD information actual value
Weight and threshold value.
7. common people's Travel Demand Forecasting method according to claim 1, which is characterized in that the Travel Demand Forecasting model
Input layer number range be 10~14, output layer interstitial content range be 3~5, hidden layer node number range be 5~
10。
8. a kind of common people's Travel Demand Forecasting device characterized by comprising
Sample data obtains module, for obtaining the association attributes at least one region, and obtains common people's trip of corresponding region
Sample data, wherein it is described trip sample data include the region association attributes and corresponding common people's travelling OD information；
Model construction module, for constructing Travel Demand Forecasting model neural network based；
Wherein, the building includes: using the association attributes in the region as input parameter with the defeated of the determination neural network
Enter node layer；Node layer is exported using common people's travelling OD information in the region as output parameter to determine；Determine hidden layer section
Point number；It is trained and verifies using the trip sample data after normalized；
Model prediction module, for the associated property data in region to be measured to be inputted the Travel Demand Forecasting model to predict
State common people's travelling OD information in region to be measured.
9. a kind of terminal, which is characterized in that the terminal includes memory and processor, and the memory is used
In storage computer program, the processor runs the computer program so that the terminal perform claim requires 1
To 7 described in any item common people's Travel Demand Forecasting methods.
10. a kind of computer readable storage medium, which is characterized in that it is stored with institute in terminal described in claim 9
Computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811100013.0A CN109345296A (en) | 2018-09-20 | 2018-09-20 | Common people's Travel Demand Forecasting method, apparatus and terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811100013.0A CN109345296A (en) | 2018-09-20 | 2018-09-20 | Common people's Travel Demand Forecasting method, apparatus and terminal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109345296A true CN109345296A (en) | 2019-02-15 |
Family
ID=65306263
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811100013.0A Pending CN109345296A (en) | 2018-09-20 | 2018-09-20 | Common people's Travel Demand Forecasting method, apparatus and terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109345296A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110443408A (en) * | 2019-07-04 | 2019-11-12 | 特斯联（北京）科技有限公司 | Travel forecasting approaches and device |
| CN111008729A (en) * | 2019-11-06 | 2020-04-14 | 阿里巴巴集团控股有限公司 | Migration prediction method and device |
| CN112488419A (en) * | 2020-12-15 | 2021-03-12 | 平安国际智慧城市科技股份有限公司 | Passenger flow distribution prediction method, device, equipment and storage medium based on OD analysis |
| CN112949784A (en) * | 2021-05-13 | 2021-06-11 | 深圳市城市交通规划设计研究中心股份有限公司 | Resident trip chain model construction method and resident trip chain acquisition method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102595323A (en) * | 2012-03-20 | 2012-07-18 | 北京交通发展研究中心 | Method for obtaining resident travel characteristic parameter based on mobile phone positioning data |
| CN102708680A (en) * | 2012-06-06 | 2012-10-03 | 北京交通大学 | Commute travel mode identification method based on AGPS technology |
| US20150344036A1 (en) * | 2014-05-30 | 2015-12-03 | The Regents Of The University Of Michigan | Vehicle speed profile prediction using neural networks |
| CN105809962A (en) * | 2016-06-13 | 2016-07-27 | 中南大学 | Traffic trip mode splitting method based on mobile phone data |
| CN106503829A (en) * | 2016-09-26 | 2017-03-15 | 同济大学 | A kind of crowding Forecasting Methodology of the Urban Public Open Space based on multi-source data |
| CN107967532A (en) * | 2017-10-30 | 2018-04-27 | 厦门大学 | The Forecast of Urban Traffic Flow Forecasting Methodology of integration region vigor |
-
2018
- 2018-09-20 CN CN201811100013.0A patent/CN109345296A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102595323A (en) * | 2012-03-20 | 2012-07-18 | 北京交通发展研究中心 | Method for obtaining resident travel characteristic parameter based on mobile phone positioning data |
| CN102708680A (en) * | 2012-06-06 | 2012-10-03 | 北京交通大学 | Commute travel mode identification method based on AGPS technology |
| US20150344036A1 (en) * | 2014-05-30 | 2015-12-03 | The Regents Of The University Of Michigan | Vehicle speed profile prediction using neural networks |
| CN105809962A (en) * | 2016-06-13 | 2016-07-27 | 中南大学 | Traffic trip mode splitting method based on mobile phone data |
| CN106503829A (en) * | 2016-09-26 | 2017-03-15 | 同济大学 | A kind of crowding Forecasting Methodology of the Urban Public Open Space based on multi-source data |
| CN107967532A (en) * | 2017-10-30 | 2018-04-27 | 厦门大学 | The Forecast of Urban Traffic Flow Forecasting Methodology of integration region vigor |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110443408A (en) * | 2019-07-04 | 2019-11-12 | 特斯联（北京）科技有限公司 | Travel forecasting approaches and device |
| CN111008729A (en) * | 2019-11-06 | 2020-04-14 | 阿里巴巴集团控股有限公司 | Migration prediction method and device |
| CN111008729B (en) * | 2019-11-06 | 2023-04-28 | 阿里巴巴集团控股有限公司 | Migration prediction method and device |
| CN112488419A (en) * | 2020-12-15 | 2021-03-12 | 平安国际智慧城市科技股份有限公司 | Passenger flow distribution prediction method, device, equipment and storage medium based on OD analysis |
| CN112949784A (en) * | 2021-05-13 | 2021-06-11 | 深圳市城市交通规划设计研究中心股份有限公司 | Resident trip chain model construction method and resident trip chain acquisition method |
| CN112949784B (en) * | 2021-05-13 | 2021-10-29 | 深圳市城市交通规划设计研究中心股份有限公司 | Resident trip chain model construction method and resident trip chain acquisition method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Johnson et al. | Optimization by simulated annealing: an experimental evaluation; part II, graph coloring and number partitioning | |
| Sánchez-Medina et al. | Using machine learning and big data for efficient forecasting of hotel booking cancellations | |
| Rajamoorthy et al. | A novel intelligent transport system charging scheduling for electric vehicles using Grey Wolf Optimizer and Sail Fish Optimization algorithms | |
| CN109345296A (en) | Common people's Travel Demand Forecasting method, apparatus and terminal | |
| Musolino et al. | Freight vehicle routing with reliable link travel times: A method based on network fundamental diagram | |
| Shang et al. | A unified framework for multicriteria evaluation of transportation projects | |
| Ke et al. | A CA-based land system change model: LANDSCAPE | |
| CN106952181A (en) | Electric Load Prediction System based on long Memory Neural Networks in short-term | |
| Liu et al. | Control method of urban intelligent parking guidance system based on Internet of Things | |
| CN106781489A (en) | A kind of road network trend prediction method based on recurrent neural network | |
| CN106910199A (en) | Towards the car networking mass-rent method of city space information gathering | |
| CN109785618A (en) | Short-term traffic flow prediction method based on combinational logic | |
| Yu et al. | Bus travel-time prediction based on bus speed | |
| Mei et al. | Real-time multistep prediction of public parking spaces based on Fourier transform–least squares support vector regression | |
| Asadi et al. | A Markov decision process approach for managing medical drone deliveries | |
| Cartlidge et al. | Spatio-temporal prediction of shopping behaviours using taxi trajectory data | |
| Vitins et al. | Optimization of large transport networks using the ant colony heuristic | |
| Dai et al. | A simulation-based approach to investigate the driver route choice behavior under the connected vehicle environment | |
| Ng et al. | On joint railway and housing development: Housing-led versus railway-led schemes | |
| Yu et al. | Passenger flow prediction for new line using region dividing and fuzzy boundary processing | |
| Zhang et al. | Allocation strategies in a dockless bike sharing system: a community structure-based approach | |
| Viverit et al. | Application of machine learning to cluster hotel booking curves for hotel demand forecasting | |
| Karthik et al. | OELF: short term load forecasting for an optimal electrical load forecasting using hybrid whale optimization based convolutional neural network | |
| Gu et al. | Enhancing market service and enterprise operations through a large-scale GIS-based distribution system | |
| Wu et al. | Visiting probability model: a new method for tourist volume forecasting |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190215 |
|
| RJ01 | Rejection of invention patent application after publication |