CN111931998B - Individual travel mode prediction method and system based on mobile positioning data - Google Patents
Individual travel mode prediction method and system based on mobile positioning data Download PDFInfo
- Publication number
- CN111931998B CN111931998B CN202010733720.4A CN202010733720A CN111931998B CN 111931998 B CN111931998 B CN 111931998B CN 202010733720 A CN202010733720 A CN 202010733720A CN 111931998 B CN111931998 B CN 111931998B
- Authority
- CN
- China
- Prior art keywords
- data
- travel
- traffic cell
- traffic
- points
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 230000008569 process Effects 0.000 claims abstract description 8
- 239000013598 vector Substances 0.000 claims description 55
- 239000011159 matrix material Substances 0.000 claims description 50
- 230000006870 function Effects 0.000 claims description 24
- 238000012545 processing Methods 0.000 claims description 14
- 230000002159 abnormal effect Effects 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 8
- 238000000354 decomposition reaction Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000003058 natural language processing Methods 0.000 claims description 5
- 230000006399 behavior Effects 0.000 abstract description 11
- 210000004027 cell Anatomy 0.000 description 76
- 238000012163 sequencing technique Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000009825 accumulation Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000011835 investigation Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
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
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Traffic Control Systems (AREA)
- Navigation (AREA)
Abstract
The invention discloses an individual travel mode prediction method and system based on mobile positioning data. The method integrates the characteristics of resident travel behaviors and the characteristics of place interest points, can comprehensively obtain the characteristics of individual travel modes, has less limitation on the sources of data, can process and calculate mobile positioning data of different sources or interest point data of different sources, and can be suitable for individual travel mode prediction under different time and space environments, so that the method is more efficient and reliable in individual travel mode prediction process and is more suitable for popularization and application.
Description
Technical Field
The invention relates to the technical field of public transportation information processing, in particular to an individual travel mode prediction method and system based on mobile positioning data.
Background
At present, with the continuous development of economy and the continuous promotion of urbanization, the vehicle maintenance amount in the city is further improved, so that the traffic flow on the original city road exceeds the pre-designed road traffic amount, the traffic jam phenomenon is more and more serious, the traffic transportation efficiency in the city is reduced, the service level is reduced, the urban traffic jam problem caused by the increase of urban population and the increase of vehicle maintenance amount is avoided, and the preparation and implementation of urban reasonable layout, road planning and construction and traffic policy are the main means for solving the traffic jam problem. And traffic demands and urban resident travel behaviors and travel modes are indistinguishable, so that important influences on the traffic system state can be generated. Therefore, studying the individual travel mode can obtain the resident travel demands hidden behind, and the travel demands can provide theoretical support for improving the transportation efficiency and the service level, and the prediction of the individual travel mode is an indispensable key link in the establishment and implementation of urban reasonable layout, road planning and construction and traffic policy.
However, existing resident trip pattern prediction methods, the data sources are mainly based on manual traffic surveys. For the manual traffic investigation method, the defects of higher investigation cost, lower efficiency, smaller data volume, stronger subjectivity of partial data and insufficient reliability exist, so that the finally obtained individual travel mode prediction result has low reference value.
Therefore, how to provide an efficient and reliable individual travel mode prediction method is a problem to be solved by those skilled in the art.
Disclosure of Invention
In view of the above, the invention provides an individual travel mode prediction method based on mobile positioning data and interest point data, which is characterized in that the mobile positioning data are processed to obtain travel behavior data of residents, traffic cells are vectorized based on the travel behavior data and the interest point data respectively, matrix decomposition is utilized to synthesize, and finally, prediction of travel behaviors of the residents is realized by calculating vector distances, so that the problems of high cost, low efficiency, small data volume, strong subjectivity of part of data and insufficient reliability of the existing travel mode prediction method are solved.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
in one aspect, the invention provides a method for predicting an individual travel mode based on mobile positioning data, which comprises the following steps:
acquiring travel data: acquiring mobile positioning original data of an individual trip, and processing the mobile positioning original data to obtain trip origin-destination data of the individual trip;
traffic cell vectorization: vectorizing the traffic cell according to the trip origin-destination data and the city interest point data obtained in advance to obtain a traffic cell vector;
predicting a travel mode: solving the distance between the traffic cell vectors, and sequencing the obtained distance between the traffic cell vectors to realize the prediction of the travel mode.
Further, in the step of obtaining travel data, a process of processing the mobile positioning raw data specifically includes:
step 1: clustering the mobile positioning original data according to the user and time, and removing abnormal points in the clustered data;
the step is essentially a process of preprocessing and cleaning the mobile positioning original data, and is mainly divided into the following two parts: (1) clustering data by user and time. (2) The method comprises the steps of removing abnormal points in data, and splitting the whole data which is too huge into sub-data clustered according to time and users. The abnormal points mentioned in the second section mainly comprise repetition points, drift points, points beyond the research range and the like.
Step 2: presetting a time threshold and a space threshold, dividing track points in clustered data into two attributes of travel and stay to obtain travel points and stay points, and taking the stay points at two sides of the travel points as the origin-destination points of one-time travel to obtain origin-destination data;
the method comprises the steps of extracting the origin and destination points by setting time and space thresholds, converting the origin and destination point extraction into the equivalent problem of dividing the track points into two attributes of travel and stay, dividing the track points into travel and stay, wherein after the travel points between the stay points are the travel corresponding tracks, and the stay points on two sides of the travel points are the origin and destination points of the travel.
Step 3: and associating the obtained origin-destination data with the pre-divided traffic cells.
Further, the traffic cell vectorizing step specifically includes:
step 1: based on the travel origin-destination data, constructing a feature vector of a traffic cell by referring to a natural language processing method through characteristics of resident travel data, and finding a corresponding vocabulary vector by constructing a maximum objective function to maximize an objective function value
Solving a maximum objective function to obtain a location matrix based on a travel track chain;
the maximum objective function is:
wherein ：
in the formula, the set v= { V 1 ，v 2 ，...，v n All words are represented, set w= { W 1 ，w 2 ，...，w T And is a training set, c represents the size of the vocabulary window, i.e., the front/back c vocabularies of the center vocabulary are considered to be associated contexts, t= |w| represents the number of all vocabularies in the training set,
the vector m representing the context vocabulary and the center vocabulary, respectively, represents the variables in the accumulation, m during the accumulation is +.>
Variations within this range. For example c=2m= -2, -1, 2.w (w) n+m Is the center w n A lexical window of size 2.
Step 2: acquiring urban interest point (Point of Interest, POI) data in advance, dividing traffic cell characteristics by using a hierarchical clustering algorithm according to the interest point characteristics of places, and obtaining a place matrix based on urban interest point information;
step 3: and processing the location matrix based on the travel track chain and the location matrix based on the urban interest point information, constructing and solving a minimum objective function, and obtaining a location essential feature matrix of the traffic cell.
The minimum objective function is:
wherein ,
where R represents a location matrix based on urban interest point information, M represents a location matrix based on travel track chains,
the conversion relationship between R, M and L is shown.
Further, the distance calculation formula between the traffic cell vectors (i.e. the location essential feature matrix of the traffic cell) is as follows:
D C (A，B)＝1-S C (A，B)
wherein ,
in the formula ,DC (A, B) represents the distance between traffic cell vectors A, B, S C (A, B) represents the similarity between the traffic cell vectors A, B, A i 、B i Each component of the vectors a, B is represented, respectively, and n represents the dimension of the vectors a, B.
On the other hand, the invention also provides an individual travel mode prediction system based on the mobile positioning data, which uses the individual travel mode prediction method and specifically comprises the following parts:
the trip data acquisition module is used for acquiring the mobile positioning original data of the trip individuals and processing the mobile positioning original data to obtain trip origin-destination data of the trip individuals;
the traffic cell vectorization module is used for vectorizing the traffic cells according to the travel origin-destination data and the city interest point data obtained in advance to obtain traffic cell vectors; and
and the travel mode prediction module is used for solving the distance between the traffic cell vectors, sequencing the obtained distance between the traffic cell vectors and completing the prediction of the travel mode.
Further, the travel data acquisition module specifically includes:
the acquisition unit is used for acquiring mobile positioning original data of the trip individuals;
the preprocessing unit is used for clustering the mobile positioning original data according to the user and time and removing abnormal points in the clustered data;
the origin-destination extraction unit is used for presetting a time threshold and a space threshold, dividing the track points in the clustered data into travel and stay two attributes to obtain travel points and stay points, and taking the stay points at two sides of the travel points as origin-destination points of one travel to obtain origin-destination data; and
and the association unit is used for associating the obtained origin-destination data with the traffic cells which are divided in advance.
Further, the traffic cell vectorization module specifically includes:
the track determining unit is used for constructing a maximum objective function of the traffic cell through a natural language processing method based on the travel origin-destination data, solving the maximum objective function and obtaining a location matrix based on a travel track chain;
the interest point determining unit is used for obtaining urban interest point data in advance, dividing traffic cell characteristics according to interest point characteristics of places, and obtaining a place matrix based on urban interest point information; and
and the decomposition unit is used for processing the location matrix based on the travel track chain and the location matrix based on the urban interest point information, constructing and solving a minimum objective function, and obtaining a location essential feature matrix of the traffic cell.
Compared with the prior art, the individual travel mode prediction method and system based on the mobile positioning data provided by the invention have the advantages that the characteristics of resident travel behaviors and the characteristics of place interest points are combined, the characteristics of individual travel modes can be comprehensively obtained, the method has less limitation on the sources of data, the mobile positioning data of different sources or the interest point data of different sources can be processed and operated, and the method can be suitable for individual travel mode prediction under different time and space environments, so that the method is more efficient and reliable in the individual travel mode prediction process and is more suitable for popularization and application.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of an individual travel mode prediction method based on mobile positioning data;
fig. 2 is a schematic flow chart of an implementation of an individual travel mode prediction method based on mobile positioning data in an embodiment of the present invention;
fig. 3 is a schematic diagram of a start-stop point extraction step in an individual travel mode prediction method based on mobile positioning data according to an embodiment of the present invention;
fig. 4 is a schematic diagram showing a prediction result display state in an individual travel mode prediction method based on mobile positioning data according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a structure of an individual travel mode prediction system based on mobile positioning data according to the present invention;
fig. 6 is a schematic structural diagram of a trip data acquisition module according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In one aspect, referring to fig. 1, an embodiment of the present invention discloses a method for predicting an individual travel mode based on mobile positioning data, where the method includes:
s1: acquiring travel data: acquiring mobile positioning original data of an individual trip, and processing the mobile positioning original data to acquire trip origin-destination data of the individual trip;
s2: traffic cell vectorization: vectorizing the traffic cell according to the travel origin-destination data and the city interest point data obtained in advance to obtain a traffic cell vector;
s3: predicting a travel mode: solving the distance between the traffic cell vectors, and sequencing the distance between the obtained traffic cell vectors to realize the prediction of the travel mode.
The above method will be specifically described below by taking Beijing as a prediction horizon by way of specific example and referring to FIG. 2.
The present embodiment extracts mobile location data of about 300 ten thousand anonymous users in Beijing city as an example on day 3, 9 in 2018, and the point of interest data is provided by a certain open source data website.
The mobile positioning data origin-destination extraction mainly comprises the following steps:
1. data preprocessing
The data preprocessing is mainly divided into the following two parts: (1) clustering data by user and time. (2) removing outliers in the data. The first part is mainly used for splitting the whole data which is too huge originally into sub-data clustered by time and users. Traversing the whole data set, and writing the rows with the same contents in the user ID column and the time column into a file with the path of "\time\user name". The second portion of data cleansing is mainly aimed at removing the following three types of abnormal data:
(1) repeat points: since objectively a user can only appear at one location at one time, each time should only correspond to one latitude and longitude data. For a certain user at a plurality of pieces of data at the same moment, the embodiment selects the longitude and latitude with the largest occurrence number as the true longitude and latitude at the moment, discards other data, and takes the longitude and latitude with the front row number in the data as the longitude and latitude at the moment if the occurrence number of the plurality of longitudes and latitudes is the same;
(2) drift point: points where the locating point deviates significantly from the due position or time due to anomalies in the time or space of the locating data are called drift points. A simple method of determining the drift point is to calculate the instantaneous speed of the track point, i.e. the quotient of the displacement of the adjacent track point and the time difference of the adjacent track point. In this embodiment, it is assumed that the instantaneous speed of any urban traffic should be within 50m/s (the travel modes mainly considered here are walking, non-motor vehicles, motor vehicles and subways, and high-speed railways and airplanes are not within the research range of urban traffic), so that the two track points before and after the instantaneous speed exceeds this value are discarded.
(3) Points outside the scope of the study: the study range of the present embodiment is the travel pattern of residents in beijing city, and therefore, for points where the latitude is not between 39 ° 26 ° to 41 ° 03' in north latitude and the longitude is not between 115 ° 25' to 117 ° 30' in east longitude, the study is not performed.
2. Origin-destination extraction
In this embodiment, the time threshold is set to 30 minutes and the space threshold is set to 300 meters. As shown in FIG. 3, the P1, P2, P3 … … P12 abstractions represent trace points in the APP data, and the arrows represent the order in which they are generated over time. First time space layer clustering is performed, namely, according to a given time threshold and a given space threshold, the track points which are close enough in space and stay long enough in time are clustered as a starting point (or an end point), and the track points between the rest of the origin points are taken as moving track points. Such as P3 to P6; p8 and P9; P10-P12 are close enough in time and space, are extracted and divided into stay points, and the rest points do not meet the conditions and are divided into travel behavior track points between two stay points. And then clustering the space level, clustering the points which are close enough in space and meet the threshold value as places, and extracting the travel track from A to B and returning to A. In the present embodiment, for the track point { p } within the location A 1 ，p 2 ，p 3 ，...，p n }＝{(x 1 ，y 1 )，(x 2 ，y 2 )，(x 3 ，y 3 )，...，(x n ，y n ) E A, calculating the average longitude and latitude of the track point
As the location point of location a. The purpose of this is to convert a from a range to a point, facilitating subsequent operations.
3. Traffic cell matching
The example uses the PostgreSQL system to combine with the PostGIS plug-in, and utilizes the built-in function to realize the efficient and convenient matching and output of the track data and the traffic cell
The traffic cell vectorization mainly comprises the following steps:
(1) Based on resident trip data
In this embodiment, the objective function of this step is solved by using a neural network method, and the neural network structure is shown in table 1:
TABLE 1 neural network architecture
| Layer(s) | Output dimension | Number of parameters |
| Full tie layer 1 | (none, 300) | 516300 |
| Activation function 1 | (none, 300) | 0 |
| Discard layer 1 | (none, 300) | 0 |
| Full tie layer 2 | (none, 1720) | 517720 |
| Activation function 2 | (none, 1720) | 0 |
Wherein the total parameter number is 1,034,020
Number of trainable parameters 1,034,020
Number of untrainable parameters 0
The parameter matrix of the full connection layer 1 after training is the corresponding traffic cell vector matrix.
(2) Based on point of interest data
For traffic cell K e K {1,2,3. }, p is used kj Represents the number of J e J {1,2,3. } class POI for that cell. Using r kj Represents p kj The percentages among class j are ranked. R is then kj The larger the j-feature of the traffic cell is, the more pronounced in the city. For each traffic cell k, a corresponding vector r can be obtained k ＝(r k1 ，r k2 ，…，r kJ )。
Meanwhile, in order to reduce the influence of mean and variance of POI information, namely to ensure that the influence of various POI indexes is the same, R is treated k The normalization process is performed as shown in the following formula:
in the formula ,
meaning reducing the influence between different j,/j>
Represents r k Average value of r kj Represents p kj The percentages among class j are ranked.
Through the processing, each traffic cell obtains the position feature vector based on the POI. Then hierarchical clustering is performed, 7 is selected as a clustering cluster number based on the dunn index (DVI), the result is matched with the geographic information of the map, each category is explained, and the meaning summary corresponding to each traffic cell can be expressed as follows:
1. commercial ring or high people stream density scenic spot
2. Park scenic spot
3. Local hub for government authorities, bus stops, train stops and the like
4. Suburban, low people flow density areas
5. Industrial park
6. Multifunctional mixing zone
7. Residential area
(3) Matrix decomposition
In the embodiment, based on a matrix decomposition model, a place essential feature matrix L, a place matrix R based on POI information and a place matrix M extracted based on a travel track chain are decomposed respectively, and the decomposed hidden vectors are ensured to be consistent. And solving the minimum objective function in CPLEX software to obtain a place essential characteristic matrix.
The matrix factorization model (i.e., the minimum objective function) described above is:
wherein ,
where R represents a location matrix based on urban interest point information, M represents a location matrix based on travel track chains,
the conversion relationship between R, M and L is shown.
In this embodiment, matrix R is a 1911×27 matrix, where 1911 comes from the number of traffic cells, and 27 comes from POI data category J e J {1,2, 3..27 }; the matrix M is 1911×300, wherein 1911 is the number of traffic cells, 300 is the total number of neurons in the full-connection layer 1, the value is manually set, and the set basis is obtained by multiple experiments; the matrix L is a result obtained by solving, and is a 1911×300 matrix, the size of the matrix is set manually, the basis of the setting of 1911 is to keep the number of traffic cells unchanged, and the basis of the setting of 300 is to experiment and reduce the difference of the matrix shape of the result.
The travel mode prediction step specifically comprises the following steps:
in this embodiment, the common cosine distance D is used c As the vector distance, the following formula is shown:
D C (A，B)＝1-S C (A，B)
wherein ,
in the formula ,DC (A, B) represents the distance between traffic cell vectors A, B, S C (A, B) represents the similarity between the traffic cell vectors A, B, A i 、B i Each component of the vectors a, B is represented, respectively, and n represents the dimension of the traffic cell vector a, B. The traffic cell vector mentioned in this embodiment is essentially the location essence feature matrix obtained in the above step.
For a given traffic cell k m Respectively calculating traffic cell k m K e {1,2,3. } K with traffic cell m Distance D of each element in (3) c And sequencing the calculation results from small to large, and taking the first q values, wherein the number of the corresponding traffic cell, namely the first q traffic cells from which the cell is most likely to come or go to, is taken.
Taking a traffic cell with a traffic cell number 666 as an example, respectively calculating cosine distances between corresponding vectors of the traffic cell 666 and corresponding vectors of the traffic cells 1,2, … 665,667, … and 1910,1911, and arranging the results from small to large, wherein the numbers of the first nine traffic cells which are most relevant to the traffic cell 666 are as follows in sequence as shown in fig. 4: 491,493,667,665,398,662,494,68,510. Then for an individual in 666 th traffic cell, their next egress destination can be predicted from the result. And then predicting the next destination again by the same method, and finally obtaining the prediction result of the individual trip mode.
In summary, in the individual trip mode prediction method and system based on mobile positioning data and interest point data disclosed in the embodiments of the present invention, the method functional core is to implement vectorization of traffic cells. For this, first, the extraction of the travel origin-destination of the resident in the mobile positioning data is realized, and the embodiment divides the travel origin-destination data of the resident by extracting the movement (travel) and stop (activity) behaviors from the trajectory data. According to the characteristics of travel origin-destination distribution frequency and natural language vocabulary distribution frequency, the method is combined with a context prediction analysis method of natural language processing, so that location vectorization based on travel behaviors is realized. Through the interest point data, each traffic cell is classified into a plurality of categories based on hierarchical clustering, and vectorization of the traffic cells is realized. The unified place comprehensive feature vector exists behind the place vectors with different obtaining modes, a matrix decomposition method is provided, and the vector results based on travel and interest points are unified. And finally, obtaining an individual travel mode prediction result according to the vector distance.
On the other hand, referring to fig. 5, the embodiment of the invention also discloses an individual travel mode prediction system based on mobile positioning data, which uses the individual travel mode prediction method, and specifically comprises the following parts:
the trip data acquisition module 1 is used for acquiring mobile positioning original data of a trip individual and processing the mobile positioning original data to acquire trip origin-destination data of the trip individual;
the traffic cell vectorization module 2 is used for vectorizing the traffic cells according to the trip origin-destination data and the city interest point data obtained in advance to obtain traffic cell vectors; and
and the travel mode prediction module 3 is used for solving the distance between the traffic cell vectors, sequencing the obtained distance between the traffic cell vectors and completing the prediction of the travel mode.
Referring to fig. 6, the trip data acquisition module 1 specifically includes:
an acquiring unit 11, configured to acquire mobile positioning raw data of an individual traveling;
a preprocessing unit 12, configured to cluster the mobile positioning raw data according to a user and time, and remove abnormal points in the clustered data;
the origin-destination extraction unit 13 is configured to preset a time threshold and a space threshold, divide the track points in the clustered data into two attributes of travel and stay, obtain travel points and stay points, and use the stay points on two sides of the travel points as origin-destination points of one-time travel to obtain origin-destination data; and
and the association unit 14 is configured to associate the obtained origin-destination data with the traffic cells that are divided in advance.
Referring to fig. 5, the traffic cell vectoring module 2 mentioned in the above embodiment specifically includes:
the track determining unit 21 constructs a maximum objective function of the traffic cell based on the travel origin-destination data through a natural language processing method, and solves the maximum objective function to obtain a location matrix based on a travel track chain;
the interest point determining unit 22 obtains city interest point data in advance, and divides the characteristics of the traffic cell according to the interest point characteristics of the places to obtain a place matrix based on the city interest point information; and
and the decomposition unit 23 is used for processing the location matrix based on the travel track chain and the location matrix based on the urban interest point information, constructing and solving a minimum objective function, and obtaining the location essential feature matrix of the traffic cell.
In this embodiment, the mobile positioning includes two parts, where the first part is to obtain position information, and includes a smart phone, a user terminal, and the like; the second part is a result visualization, and corresponding positions are marked on the electronic map based on the position information obtained by the first part. The mobile positioning data has the advantages of large scale, low cost and high reliability, and a new idea is opened for resident travel mode research. The embodiment designs and proposes a resident track node attribute judging method, a resident travel track dividing and extracting method and an individual travel mode prediction method based on resident travel behaviors based on the smart phones which are already highly popular in cities at present and by using mobile positioning data generated by the smart phones as a basis, and provides important references for studying resident travel modes.
It is easy to find that the individual travel mode prediction method and system based on the mobile positioning data and the interest point data provided by the embodiment of the invention have the following advantages compared with the prior art:
1. the method integrates the characteristics of resident travel behaviors and the characteristics of place interest points, and can comprehensively obtain the characteristics of individual travel modes;
2. the method has less limitation on the source of the data, can process and calculate the mobile positioning data from different sources or the interest point data from different sources, can set specific parameters according to different use scenes and can carry out slicing screening on the input data so as to meet specific requirements, and the method can be suitable for individual travel mode prediction under different time space environments and has popularization.
3. The vectorization of the traffic cell realized by the method can be combined with other travel modes or urban traffic planning researches (OD matrix calculation among urban traffic cells, traffic cell category division, individual travel destination prediction and the like), and has expansibility.
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (3)
1. The individual travel mode prediction method based on the mobile positioning data is characterized by comprising the following steps of:
acquiring travel data: acquiring mobile positioning original data of an individual trip, and processing the mobile positioning original data to obtain trip origin-destination data of the individual trip;
traffic cell vectorization: vectorizing the traffic cell according to the trip origin-destination data and the city interest point data obtained in advance to obtain a traffic cell vector;
predicting a travel mode: numbering the traffic cells, selecting a target traffic cell, solving the distance between the target traffic cell and corresponding vectors of other traffic cells, sorting the obtained distance between the traffic cell vectors from small to large, and taking the first q values, wherein the number of the corresponding traffic cell is the q traffic cells from or to which the target traffic cell is most likely to come, so as to realize prediction of a travel mode;
in the step of obtaining travel data, a process of processing the mobile positioning original data specifically includes:
clustering the mobile positioning original data according to the user and time, and removing abnormal points in the clustered data;
presetting a time threshold and a space threshold, dividing track points in clustered data into two attributes of travel and stay to obtain travel points and stay points, and taking the stay points at two sides of the travel points as the origin-destination points of one-time travel to obtain origin-destination data;
associating the obtained origin-destination data with the traffic cells divided in advance;
the traffic cell vectorization step specifically comprises the following steps:
constructing a maximum objective function of a traffic cell based on the travel origin-destination data, and solving the maximum objective function to obtain a location matrix based on a travel track chain;
acquiring urban interest point data in advance, dividing traffic cell characteristics according to interest point characteristics of places, and obtaining a place matrix based on urban interest point information;
processing the location matrix based on the travel track chain and the location matrix based on the urban interest point information, constructing and solving a minimum objective function to obtain a location essential feature matrix of the traffic cell, namely obtaining a traffic cell vector;
the distance calculation formula between the traffic cell vectors is as follows:
D c (A,B)＝1-S c (A,B)
wherein ,
in the formula ,DC (A, B) represents the distance between traffic cell vectors A, B, S C (A, B) represents the similarity between the traffic cell vectors A, B, A i 、B i Each component of the vectors a, B is represented, respectively, and e represents the dimension of the traffic cell vector a, B.
2. The method for predicting individual travel patterns based on mobile location data as recited in claim 1, wherein the outlier points include a repetition point, a drift point, and a point out of a prediction range.
3. An individual travel pattern prediction system based on mobile positioning data, characterized in that an individual travel pattern prediction is implemented using an individual travel pattern prediction method based on mobile positioning data according to any one of claims 1-2, the system comprising:
the trip data acquisition module is used for acquiring the mobile positioning original data of the trip individuals and processing the mobile positioning original data to obtain trip origin-destination data of the trip individuals;
the traffic cell vectorization module is used for vectorizing the traffic cells according to the travel origin-destination data and the city interest point data obtained in advance to obtain traffic cell vectors; and
the travel mode prediction module is used for numbering the traffic cells, selecting a target traffic cell, solving the distance between the target traffic cell and corresponding vectors of other traffic cells, sorting the obtained distance between the traffic cell vectors from small to large, taking the first q values, and predicting the travel mode by the number of the corresponding traffic cell, namely the q traffic cells from which the target traffic cell is most likely to come or go;
the travel data acquisition module specifically comprises:
the acquisition unit is used for acquiring mobile positioning original data of the trip individuals;
the preprocessing unit is used for clustering the mobile positioning original data according to the user and time and removing abnormal points in the clustered data;
the origin-destination extraction unit is used for presetting a time threshold and a space threshold, dividing the track points in the clustered data into travel and stay two attributes to obtain travel points and stay points, and taking the stay points at two sides of the travel points as origin-destination points of one travel to obtain origin-destination data; the association unit is used for associating the obtained origin-destination data with the traffic cells which are divided in advance;
the traffic cell vectorization module specifically comprises:
the track determining unit is used for constructing a maximum objective function of the traffic cell through a natural language processing method based on the travel origin-destination data, solving the maximum objective function and obtaining a location matrix based on a travel track chain;
the interest point determining unit is used for obtaining urban interest point data in advance, dividing traffic cell characteristics according to interest point characteristics of places, and obtaining a place matrix based on urban interest point information; and
the decomposition unit is used for processing the location matrix based on the travel track chain and the location matrix based on the urban interest point information, constructing and solving a minimum objective function to obtain a location essential feature matrix of the traffic cell, and obtaining a traffic cell vector;
the distance calculation formula between the traffic cell vectors is as follows:
D c (A,B)＝1-S c (A,B)
wherein ,
in the formula ,DC (A, B) represents the distance between traffic cell vectors A, B, S C (A, B) represents the similarity between the traffic cell vectors A, B, A i 、B i Each component of the vectors a, B is represented, respectively, and e represents the dimension of the traffic cell vector a, B.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010733720.4A CN111931998B (en) | 2020-07-27 | 2020-07-27 | Individual travel mode prediction method and system based on mobile positioning data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010733720.4A CN111931998B (en) | 2020-07-27 | 2020-07-27 | Individual travel mode prediction method and system based on mobile positioning data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111931998A CN111931998A (en) | 2020-11-13 |
| CN111931998B true CN111931998B (en) | 2023-05-02 |
Family
ID=73314230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010733720.4A Expired - Fee Related CN111931998B (en) | 2020-07-27 | 2020-07-27 | Individual travel mode prediction method and system based on mobile positioning data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111931998B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112288048B (en) * | 2020-12-28 | 2021-03-26 | 湖南师范大学 | Urban crowd trip identification method based on multi-source data driving |
| CN112784901B (en) * | 2021-01-22 | 2023-04-07 | 北京航空航天大学 | Pedestrian travel classification detection method and system based on position data |
| CN112883133B (en) * | 2021-02-01 | 2024-05-24 | 南京航空航天大学 | Flow prediction method based on time sequence data and functional evolution data |
| CN113192321A (en) * | 2021-03-17 | 2021-07-30 | 东南大学 | Traffic demand distribution extraction method for comprehensive land utilization |
| CN113011512A (en) * | 2021-03-29 | 2021-06-22 | 长沙理工大学 | Traffic generation prediction method and system based on RBF neural network model |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106600960A (en) * | 2016-12-22 | 2017-04-26 | 西南交通大学 | Traffic travel origin and destination identification method based on space-time clustering analysis algorithm |
| CN107977734A (en) * | 2017-11-10 | 2018-05-01 | 河南城建学院 | A kind of Forecasting Methodology based on mobile Markov model under space-time big data |
| WO2018120424A1 (en) * | 2016-12-29 | 2018-07-05 | 平安科技（深圳）有限公司 | Location service-based method, device, equipment for crowd classification, and storage medium |
| CN108256914A (en) * | 2018-01-17 | 2018-07-06 | 华东师范大学 | A kind of point of interest category forecasting method based on tensor resolution model |
| CN109583640A (en) * | 2018-11-23 | 2019-04-05 | 东南大学 | A kind of Urban Traffic passenger flow attribute recognition approach based on multi-source location data |
-
2020
- 2020-07-27 CN CN202010733720.4A patent/CN111931998B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106600960A (en) * | 2016-12-22 | 2017-04-26 | 西南交通大学 | Traffic travel origin and destination identification method based on space-time clustering analysis algorithm |
| WO2018120424A1 (en) * | 2016-12-29 | 2018-07-05 | 平安科技（深圳）有限公司 | Location service-based method, device, equipment for crowd classification, and storage medium |
| CN107977734A (en) * | 2017-11-10 | 2018-05-01 | 河南城建学院 | A kind of Forecasting Methodology based on mobile Markov model under space-time big data |
| CN108256914A (en) * | 2018-01-17 | 2018-07-06 | 华东师范大学 | A kind of point of interest category forecasting method based on tensor resolution model |
| CN109583640A (en) * | 2018-11-23 | 2019-04-05 | 东南大学 | A kind of Urban Traffic passenger flow attribute recognition approach based on multi-source location data |
Non-Patent Citations (2)
| Title |
|---|
| 《impacts of free-floating bikesharing system on public transit ridership》;Xiaolei Ma 等;《Transportation Research Part D Transport and Environment 》;20191130;全文 * |
| 基于手机信令大数据的城市居民出行OD预测;孙卓等;《数学的实践与认识》;20190608(第11期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111931998A (en) | 2020-11-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111931998B (en) | Individual travel mode prediction method and system based on mobile positioning data | |
| CN111222744B (en) | Method for determining distribution relation between built environment and rail passenger flow based on signaling data | |
| CN107270922B (en) | Traffic accident spatial positioning method based on POI (Point of interest) index | |
| CN112749825B (en) | Method and device for predicting destination of vehicle | |
| CN111612670B (en) | Method, device and computer equipment for constructing emission list of motor vehicle | |
| CN110836675B (en) | Decision tree-based automatic driving search decision method | |
| CN110472999B (en) | Passenger flow mode analysis method and device based on subway and shared bicycle data | |
| CN112863182B (en) | Cross-modal data prediction method based on transfer learning | |
| CN112418696B (en) | Construction method and device of urban traffic dynamic knowledge graph | |
| CN110969861B (en) | Vehicle identification method, device, equipment and computer storage medium | |
| CN111046937A (en) | Two-segment passenger crowd trip purpose analysis method fusing public transportation data and POI data | |
| CN113298314A (en) | Rail transit passenger flow prediction method considering dynamic space-time correlation | |
| CN107798440B (en) | Subway connection bus line planning method based on line candidate set | |
| CN110070255A (en) | The method for introducing commuter's Passenger Traveling Choice modeling and analysis after sharing bicycle | |
| CN112258029A (en) | Demand prediction method for shared bicycle around subway station | |
| CN110889086A (en) | Method for evaluating influence of shared electric rental car on urban automobile exhaust emission | |
| CN113159371B (en) | Unknown target feature modeling and demand prediction method based on cross-modal data fusion | |
| CN109740957A (en) | A kind of urban traffic network node-classification method | |
| CN117648621A (en) | Subway station asymmetric passenger flow space-time mode mining method based on graph convolution | |
| Polyzos et al. | Determining the changes in commuting after the Ionian Motorway’s construction | |
| CN108960684B (en) | Public transport passenger closed travel chain mining method based on incomplete track segment | |
| CN107194505A (en) | A kind of method and system that bus travel amount is predicted based on city big data | |
| CN114862001B (en) | Urban crowd flow prediction method and system based on regional function enhancement features | |
| CN115345354A (en) | Urban multi-mode traffic hyper-network situation prediction method based on hyper-map deep network | |
| Zheng et al. | Discovering urban functional regions with call detail records and points of interest: A case study of Guangzhou 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20230502 |