CN111275482A - Machine learning-based real-time dynamic rail transit sorting method - Google Patents

Machine learning-based real-time dynamic rail transit sorting method Download PDF

Info

Publication number
CN111275482A
CN111275482A CN202010039532.1A CN202010039532A CN111275482A CN 111275482 A CN111275482 A CN 111275482A CN 202010039532 A CN202010039532 A CN 202010039532A CN 111275482 A CN111275482 A CN 111275482A
Authority
CN
China
Prior art keywords
path
time
passenger
station
paths
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.)
Granted
Application number
CN202010039532.1A
Other languages
Chinese (zh)
Other versions
CN111275482B (en
Inventor
胡鑫
丁康
曹辉
刘卫红
郭庆
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Supcon Information Technology Co ltd
Original Assignee
Zhejiang Supcon Information Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang Supcon Information Technology Co ltd filed Critical Zhejiang Supcon Information Technology Co ltd
Priority to CN202010039532.1A priority Critical patent/CN111275482B/en
Publication of CN111275482A publication Critical patent/CN111275482A/en
Application granted granted Critical
Publication of CN111275482B publication Critical patent/CN111275482B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Marketing (AREA)
  • General Health & Medical Sciences (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Game Theory and Decision Science (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Primary Health Care (AREA)
  • Tourism & Hospitality (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a machine learning-based real-time dynamic rail transit sorting method, which comprises the following steps of: collecting an input data set: in a statistical period, the time of each passenger from the station i to the station j to get in and out is collected from an AFC card swiping position of the subway, and the time difference of the time of each passenger from getting in and out is calculated; calculating the average required time of the paths: calculating the average required time of each path from the station i to the station j by using a train timetable released by a subway company; merging the similar paths into one path to obtain merged paths from the station i to the station j, wherein the number of the merged paths is A; the invention has the characteristics of solving the inconvenience and inaccuracy caused by counting in a manual investigation mode, saving human resources and improving the accuracy.

Description

Machine learning-based real-time dynamic rail transit sorting method
Technical Field
The invention relates to the technical field of rail transit monitoring, in particular to a high-accuracy rail transit real-time dynamic sorting method based on machine learning.
Background
Clearing: the data preparation stage of clearing is mainly to collect, sort and sort all the network transaction data of the current day according to the current generation, the other generation, the credit, the debit, the stroke number, the amount, the net amount of the netting, and the like among member lines.
OD path passenger flow: "O" is from English Origin and refers to the starting place of the trip, and "D" is from English Destination and refers to the Destination of the trip. In rail transit, there may be multiple reachable paths from O to D. The OD route passenger flow is the selection of each reachable route in all passenger flows from O to D.
LSTM: the Long Short-Term Memory network (LSTM) is a time-cycle neural network, which is specially designed to solve the Long-Term dependence problem of the general RNN (cyclic neural network), and all RNNs have a chain form of repeated neural network modules.
K-Means: the K-means clustering algorithm (K-means clustering algorithm) is an iterative solution clustering analysis algorithm, and the steps are that K objects are randomly selected as initial clustering centers, then the distance between each object and each seed clustering center is calculated, and each object is allocated to the clustering center closest to the object. The cluster centers and the objects assigned to them represent a cluster. The cluster center of a cluster is recalculated for each sample assigned based on the objects existing in the cluster. This process will be repeated until some termination condition is met. The termination condition may be that no (or minimum number) objects are reassigned to different clusters, no (or minimum number) cluster centers are changed again, and the sum of squared errors is locally minimal.
AFC system: the Automatic Fare Collection System is called as an Automatic Fare Collection System, refers to an Automatic Fare Collection System for urban rail transit, and can realize automation of rail transit ticket selling, Fare Collection, charging, statistics, score clearing and management.
In recent years, urban rail transit is rapidly developing in China and even in the world due to the characteristics of large traffic volume, punctuality, safety, rapidness and the like. The rail transit is gradually becoming the first choice traffic mode for daily trips of urban residents, and the development of the rail transit is a key link for promoting urban development, improving rail transit trip service and ensuring urban sustainable development. Meanwhile, due to the increase of lines, more choices are provided for travel paths of passengers taking subways, so that the passenger volume of each station and each path needs to be calculated and predicted accurately and accurately in real time, and related workers are helped to monitor, analyze data and make aid decisions in real time. Therefore, the rail transit OD path passenger flow volume prediction is also gradually becoming a hot spot of current research. The method can master the space-time distribution rule of the rail transit passenger flow, and has wide and profound social significance for guaranteeing the safe operation management of the rail transit.
At present, the implementation scheme of predicting the passenger flow volume of rail transit mainly focuses on constructing a prediction model of passenger flow at stations, in stations, time periods and the like, and the technology used in the model is also a popular machine learning algorithm at present. There is no relevant implementation for OD path traffic prediction. The traditional OD path passenger flow volume prediction is statistical analysis by a manual statistical sampling investigation mode, and the mode has certain credibility, but has many defects that firstly, certain human resources are consumed for frequent investigation; the objects of the second survey are not necessarily representative of stratification; thirdly, the information of the investigation is probably only the current idea of the opposite party and is not the final idea, and other factors are not considered and have uncertainty; the number of people in the final survey is too small relative to the proportion of the total rail transit occupancy, further reducing accuracy considerably.
Disclosure of Invention
The invention aims to overcome the defect of low accuracy rate of OD path passenger flow prediction in the prior art, and provides a machine learning-based rail transit real-time dynamic clearing method with high accuracy rate.
In order to achieve the purpose, the invention adopts the following technical scheme:
a rail transit real-time dynamic sorting method based on machine learning comprises the following steps: (1-1) acquiring an input data set: in a statistical period, the time of each passenger from the station i to the station j to get in and out is collected from an AFC card swiping position of the subway, and the time difference of the time of each passenger from getting in and out is calculated;
let OD be { O ═ OiDjI ≠ j, i ∈ N, j ∈ N }, where N denotes a set of all stations in the subway line network, OD denotes a set of any station i to any other station j in the subway line network,
(1-2) calculating a path average required time: calculating the average required time of each path from the station i to the station j by using a train timetable released by a subway company;
(1-3) merging the similar paths into one path to obtain a merged path from the station i to the station j, wherein the number of the merged paths is A;
(1-4) classifying by using k-means; taking A as the model category number of the k-means model, and taking the average required time of each merged path as the initialization center of the k-means model corresponding to the path;
inputting the time difference of the arrival time and the departure time of each passenger into a k-means model, and classifying the paths taken by each passenger by the k-means model according to the time difference of the arrival time and the departure time of each passenger to obtain the classified paths and the classified central point; wherein the classified center point represents the average travel time of the passenger selecting the path;
(1-5) if the path of a certain passenger after classification is not the combined path, storing the path information and the path passenger selection probability into a result set;
if the path is the path after merging, storing the information of each path before merging into a prediction set;
(1-5) traversing each element in the OD set to obtain a final result set and a prediction set, inputting the final result set into an LSTM neural network model for training, and training the LSTM neural network model;
(1-6) inputting the final prediction set into an LSTM neural network model for prediction, predicting the passenger path selection probability of each path, and storing the passenger path selection probability into a result set of the path after prediction;
and (1-7) realizing dynamic score clearing according to the passenger path selection probability obtained in the result set.
The invention mainly aims to predict the passenger flow of each OD path in the whole track network by reading data from an AFC card swiping system, and the prediction model has self-learning capability, can continuously adjust and optimize parameters along with the increase of the data volume along with the increase of time, and is more applicable. Therefore, various defects caused by manual statistics of the OD path passenger flow are overcome.
The method mainly carries out probability prediction on selection of a passenger about an OD (Origin-Destination) path in rail transit, firstly collects passenger riding information from an AFC (automatic fare control) system, and then combines similar paths (with similar time required by the paths) into one path according to the path information; establishing a k-means clustering model, and substituting passenger riding information into the model for classification; for the non-merged path, storing the classified result into a result set; and finally, establishing an LSTM neural network model, substituting the stored result set as a training set into the LSTM model for training, and predicting the selection probability of each similar path by using the LSTM model after training. And finally, the real-time dynamic clearing of the rail transit is realized.
The problem of inconvenience and inaccuracy caused by statistics in a manual investigation mode is solved, manpower resources are saved, accuracy is improved, and data support is provided for relevant management personnel to make schemes and analyze statistics.
Preferably, (1-3) comprises the steps of:
if the average time required by the two paths is respectively B1 and B2, B1 is greater than B2, and if ((B1-B2)/B2) < 10%, judging that the two paths are similar paths, and setting the average time required by the combined paths as the average value of the two path times.
Preferably, (1-5) comprises the steps of:
setting m lines in a subway network, wherein the m lines are 1, 2, 3, … and m respectively; w1，W2，W3，…，WmEach indicating the congestion degree of each line, the following:
the result set includes the following elements: route number, center point after classification, average time taken for route, W1，W2，W3，…，WmThe path transfer distance, the path transfer times, the number of stations passed by the path and the passenger path selection probability;
the prediction set includes the following elements: route number, center point after classification, average time taken for route, W1，W2，W3，…，WmPath transfer distance, number of times of path transfer, number of stations where the path passes.
Preferably, (1-7) comprises the steps of:
obtaining a certain O from AFC card swiping machineiDjThe number of all passengers;
calculating out OiDjThe income amount of (1): s ═ OiDjFare x OiDjNumber of people
OiDjThe set of paths in (1) is Y, for each path t, t belongs to Y, the passenger path selection probability is WtAnd calculating the proportion of the routes occupied by each route in each route: zt，kK represents a line number; and calculating the amount of money obtained by each line operation company according to the route fare:
OiDjsum of sigma obtained by the middle line kt∈YS×Wt×Zt，k
And calculating all OiDj elements in the OD set to obtain the total sum of each line.
Preferably, the average required route time is calculated from a train schedule published by a subway company, and represents an average of at least the time required to be taken and at most the time required to be taken by a passenger when the passenger selects the route under normal conditions.
Therefore, the invention has the following beneficial effects: the problem of inconvenience and inaccuracy caused by statistics in a manual investigation mode is solved, manpower resources are saved, accuracy is improved, and data support is provided for relevant management personnel to make schemes and analyze statistics.
Drawings
FIG. 1 is a schematic diagram of an OD path according to the present invention.
Detailed Description
The invention is further described with reference to the following figures and detailed description.
A rail transit real-time dynamic sorting method based on machine learning comprises the following steps:
(1-1) acquiring an input data set: in a statistical period, the time of each passenger from the station i to the station j to get in and out is collected from an AFC card swiping position of the subway, and the time difference of the time of each passenger from getting in and out is calculated;
let OD be { O ═ OiDjI ≠ j, i ∈ N, j ∈ N }, where N denotes a set of all stations in the subway line network, OD denotes a set of any station i to any other station j in the subway line network,
(1-2) calculating a path average required time: calculating the average required time of each path from the station i to the station j by using a train timetable released by a subway company;
the average required time of the route represents the average value of at least the time required to be spent and at most the time required to be spent when the passenger selects the route under normal conditions, and can be calculated by a train schedule published by a subway company. The average required time of the path is the train running time + the waiting time of the passengers + the time required for the passengers to transfer the train to walk.
(1-3) merging the similar paths into one path to obtain a merged path from the station i to the station j, wherein the number of the merged paths is A;
if the average time required by the two paths is respectively B1 and B2, B1 is greater than B2, and if ((B1-B2)/B2) < 10%, judging that the two paths are similar paths, and setting the average time required by the combined paths as the average value of the two path times.
(1-4) classifying by using k-means; taking A as the model category number of the k-means model, and taking the average required time of each merged path as the initialization center of the k-means model corresponding to the path;
inputting the time difference of the arrival time and the departure time of each passenger into a k-means model, and classifying the paths taken by each passenger by the k-means model according to the time difference of the arrival time and the departure time of each passenger to obtain the classified paths and the classified central point; wherein the classified center point represents the average travel time of the passenger selecting the path;
(1-5) if the path of a certain passenger after classification is not the combined path, storing the path information and the path passenger selection probability into a result set;
if the path is the path after merging, storing the information of each path before merging into a prediction set;
setting m lines in a subway network, wherein the m lines are 1, 2, 3, … and m respectively; w1，W2，W3，…，WmEach indicating the congestion degree of each line, the following:
the result set includes the following elements: route number, center point after classification, average time taken for route, W1，W2，W3，…，WmThe path transfer distance, the path transfer times, the number of stations passed by the path and the passenger path selection probability;
the prediction set includes the following elements: route number, center point after classification, average time taken for route, W1，W2，W3，…，WmPath transfer distance, number of times of path transfer, number of stations where the path passes.
(1-6) traversing each element in the OD set to obtain a final result set and a prediction set, inputting the final result set into an LSTM neural network model for training, and training the LSTM neural network model;
(1-7) inputting the final prediction set into an LSTM neural network model for prediction, predicting the passenger path selection probability of each path, and storing the passenger path selection probability into a result set of the path after prediction;
and (1-8) realizing dynamic score clearing according to the passenger path selection probability obtained in the result set.
Obtaining a certain O from AFC card swiping machineiDjThe number of all passengers;
calculating out OiDjThe income amount of (1): s ═ OiDjFare x OiDjNumber of people
OiDjThe set of paths in (1) is Y, for each path t, t belongs to Y, the passenger path selection probability is WtAnd calculating the proportion of the routes occupied by each route in each route: zt，kK represents a line number; and calculating the amount of money obtained by each line operation company according to the route fare:
OiDjsum of sigma obtained by the middle line kt∈YS×Wt×Zt，k
For all O in OD concentrationiDjAnd calculating the elements to obtain the total sum of each line.
Example (c):
as shown in fig. 1, the respective path passenger amounts from station a to station F are calculated. There are 4 paths from a to F, and the basic known information of the paths is: a route list, transfer times, passing stations, transfer walking distance, and route average required time, as shown in table 1.
Figure BDA0002366864190000081
Figure BDA0002366864190000091
Table 1 acquisition of data set samples
The method comprises the following specific steps:
1. an input data set is collected. And in a statistical period, the time from A to F for each passenger to get in and out of the station is collected from an AFC card swiping position of the subway, and the time difference of getting in and out of the station is calculated. As shown in table 1.
2. The path average required time is calculated. The average required time of each of the four paths a to F was calculated from a train schedule issued by the subway company, and the final results are shown in table 2.
Figure BDA0002366864190000092
TABLE 2A-F Path basic information
2. Merging the similar paths into one path. When the average required time of the paths of the two paths is close (the larger one is not more than 10% of the smaller one), the two paths are considered to be similar paths, and the two paths cannot be distinguished obviously by the average required time of the paths. And combining the paths into one path, wherein the average time required by the combined path is the average value of the time of the two paths, and the path identifier is the splicing of the original two path identifiers. Path 1 and path 2 are similar paths (45+ 0.1X 45 > 48) as in Table 2, and the merged path is shown in Table 3.
Figure BDA0002366864190000093
Table 3 merged path information
4. The classification is performed using k-means. Constructing a k-means model, taking the number of the merged path sets as the number of model categories (for example, 3 categories in table 3), taking the average time required by the merged path categories as the initialization centers of the corresponding categories in the model (for example, 3 initialization centers in table 3: 46.5, 55, 65), inputting the time difference of the input data set collected in the step 1 (for example, the column of "time difference" in table 1) into the model, and classifying the categories of the input data set. And obtaining the classified result set and the classified central point. Wherein the sorted center point represents the average travel time of the passenger selecting the path. The data in table 1 are sorted as shown in table 4.
5. For the result classified in the previous step, if the path is not the combined path, the classified passenger is accurately classified, and the proportion of the path information and the number of passengers is stored in a result set (such as path 3 and path 4 in table 4); if the route is the merged route, the classified passenger is not classified accurately, and the information of each route before merging is stored in the prediction set (for example, route 1-2 in table 4, and route 1 and route 2 information are stored in the prediction set). Assuming that 4 lines are totally arranged in the whole subway line network, namely a, b, c and d, wherein W represents the congestion degree of the current line, a is a No. 1 line, b is a No. 2 line, c is a No. 3 line, and d is a No. 4 line; the saved format is as follows:
and (4) result set: (Path number, center point after classification, path average required time, Wa，Wb，Wc，WdDistance of transfer, number of times of transfer, number of stations passed by path, probability of path selection)
Prediction set: (Path number, center point after classification, path average required time, Wa，Wb，Wc，WdPath transfer distance, number of times of path transfer, number of path passing stations) are stored in the sample format shown in table 5.
Figure BDA0002366864190000101
TABLE 5 result set prediction set preservation examples
6.OD{OiDjI ≠ j, i ∈ N, j ∈ N }, where N denotes a set of subway stations, and OD denotes a set from any station i to any other station j in a subway network. If the net has 4 stations, the OD set has 4 × 3 — 12 elements. And traversing each element of the OD set, and executing the steps 1-5 for each OD element to obtain a final result set and a prediction set.
7. Input to the LSTM neural network model. Input parameters of the model: (time difference between arrival and departure of passengers, average required time of route, Wa，Wb，Wc，WdPath transfer distance, path transfer times, number of stations passed by the path), and finally predicting an output result: passenger routing probability. Substituting the result set generated in the step 6 into a model for training, inputting the prediction set generated in the step 6 into the model for prediction after the model is trained, predicting the selection probability of each path, and storing the result into the result set after prediction;
8. according to the passenger selection probability of each path obtained in the result set, the score clearing calculation is dynamically realized, and the steps are as follows:
a) obtaining a certain O from AFC card swiping machineiDjThe number of all passengers.
b) Calculating the OiDjThe amount of the intermediate income: s ═ OiDjFare x OiDjNumber of people
c)OiDjThe set of the middle paths is Y, and for each path t belonging to Y, the passenger selection probability is WtAnd calculating the proportion of the routes occupied by each route in each route: zt，kAnd k represents a line.
d) The amount of money obtained by each line operator is calculated based on the OD fare.
OiDjSum of sigma obtained by the middle line kt∈YS×Wt×Zt，k
e) For all O in OD concentrationiDjAnd repeating the steps by the elements, and finally calculating to obtain the total sum of each line.
Sample calculations are shown in table 6.
Figure BDA0002366864190000111
Figure BDA0002366864190000121
Figure BDA0002366864190000131
TABLE 6 Scoring calculation example
It should be understood that this example is for illustrative purposes only and is not intended to limit the scope of the present invention. Further, it should be understood that various changes or modifications of the present invention may be made by those skilled in the art after reading the teaching of the present invention, and such equivalents may fall within the scope of the present invention as defined in the appended claims.

Claims (5)

1. A rail transit real-time dynamic sorting method based on machine learning is characterized by comprising the following steps:
(1-1) acquiring an input data set: in a statistical period, the time of each passenger from the station i to the station j to get in and out is collected from an AFC card swiping position of the subway, and the time difference of the time of each passenger from getting in and out is calculated;
let OD be { O ═ OiDjI ≠ j, i ∈ N, j ∈ N }, where N denotes a set of all stations in the subway line network, OD denotes a set of any station i to any other station j in the subway line network,
(1-2) calculating a path average required time: calculating the average required time of each path from the station i to the station j by using a train timetable released by a subway company;
(1-3) merging the similar paths into one path to obtain a merged path from the station i to the station j, wherein the number of the merged paths is A;
(1-4) classifying by using k-means; taking A as the model category number of the k-means model, and taking the average required time of each merged path as the initialization center of the k-means model corresponding to the path;
inputting the time difference of the arrival time and the departure time of each passenger into a k-means model, and classifying the paths taken by each passenger by the k-means model according to the time difference of the arrival time and the departure time of each passenger to obtain the classified paths and the classified central point;
(1-5) if the path of a certain passenger after classification is not the combined path, storing the path information and the path passenger selection probability into a result set;
if the path is the path after merging, storing the information of each path before merging into a prediction set;
(1-6) traversing each element in the 0D set to obtain a final result set and a prediction set, inputting the final result set into an LSTM neural network model for training, and training the LSTM neural network model;
(1-7) inputting the final prediction set into an LSTM neural network model for prediction, predicting the passenger path selection probability of each path, and storing the passenger path selection probability into a result set of the path after prediction;
and (1-8) realizing dynamic score clearing according to the passenger path selection probability obtained in the result set.
2. The rail transit real-time dynamic sorting method based on machine learning as claimed in claim 1, wherein (1-3) comprises the following steps:
if the average time required by the two paths is respectively B1 and B2, B1 is greater than B2, and if ((B1-B2)/B2) < 10%, judging that the two paths are similar paths, and setting the average time required by the combined paths as the average value of the two path times.
3. The rail transit real-time dynamic sorting method based on machine learning as claimed in claim 1, wherein (1-5) comprises the following steps:
setting m lines in a subway network, wherein the m lines are 1, 2, 3, … and m respectively; w1，W2，W3，…，WmEach indicating the congestion degree of each line, the following:
the result set includes the following elements: route number, center point after classification, average time taken for route, W1，W2，W3，…，WmThe path transfer distance, the path transfer times, the number of stations passed by the path and the passenger path selection probability;
the prediction set includes the following elements: route number, center point after classification, average time taken for route, W1，W2，W3，…，WmPath transfer distance, number of times of path transfer, number of stations where the path passes.
4. The rail transit real-time dynamic sorting method based on machine learning as claimed in claim 1, wherein (1-8) comprises the following steps:
obtaining a certain O from AFC card swiping machineiDjThe number of all passengers;
calculating out OiDjThe income amount of (1): s ═ OiDjFare x OiDjNumber of people
OiDjThe set of paths in (1) is Y, and for each path t, t belongs to Y, the passenger path selection summaryA rate of WtAnd calculating the proportion of the routes occupied by each route in each route: zt，kK represents a line number;
and calculating the amount of money obtained by each line operation company according to the route fare:
OiDjsum of sigma obtained by the middle line kt∈YS×Wt×Zt，k
For all O in OD concentrationiDjAnd calculating the elements to obtain the total sum of each line.
5. The real-time dynamic rail transit sorting method based on machine learning according to claim 1, 2, 3 or 4,
the average required time of the route is the average value of at least the time required to be spent and at most the time required to be spent when the passenger selects the route under normal conditions, and is calculated by a train schedule released by a subway company.
CN202010039532.1A 2020-01-14 2020-01-14 Machine learning-based real-time dynamic track traffic sorting method Active CN111275482B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010039532.1A CN111275482B (en) 2020-01-14 2020-01-14 Machine learning-based real-time dynamic track traffic sorting method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010039532.1A CN111275482B (en) 2020-01-14 2020-01-14 Machine learning-based real-time dynamic track traffic sorting method

Publications (2)

Publication Number Publication Date
CN111275482A true CN111275482A (en) 2020-06-12
CN111275482B CN111275482B (en) 2023-05-09

Family

ID=71002972

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010039532.1A Active CN111275482B (en) 2020-01-14 2020-01-14 Machine learning-based real-time dynamic track traffic sorting method

Country Status (1)

Country Link
CN (1) CN111275482B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113536493A (en) * 2021-06-23 2021-10-22 南京理工大学 Effective path generation method based on clustering reverse thrust and section passenger flow estimation method

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102279950A (en) * 2011-08-02 2011-12-14 中国铁道科学研究院电子计算技术研究所 Railway transit fare clearing method based on data mining and neural network
CN102306243A (en) * 2011-07-27 2012-01-04 高新现代智能系统股份有限公司 Method for sorting transfer passenger flow in rail traffic
CN103761589A (en) * 2014-02-18 2014-04-30 东南大学 Distribution method for urban rail transit
WO2016045195A1 (en) * 2014-09-22 2016-03-31 北京交通大学 Passenger flow estimation method for urban rail network
CN108108897A (en) * 2017-12-20 2018-06-01 中国科学院深圳先进技术研究院 A kind of track traffic for passenger flow classifying method, system and electronic equipment
US10176435B1 (en) * 2015-08-01 2019-01-08 Shyam Sundar Sarkar Method and apparatus for combining techniques of calculus, statistics and data normalization in machine learning for analyzing large volumes of data

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102306243A (en) * 2011-07-27 2012-01-04 高新现代智能系统股份有限公司 Method for sorting transfer passenger flow in rail traffic
CN102279950A (en) * 2011-08-02 2011-12-14 中国铁道科学研究院电子计算技术研究所 Railway transit fare clearing method based on data mining and neural network
CN103761589A (en) * 2014-02-18 2014-04-30 东南大学 Distribution method for urban rail transit
WO2016045195A1 (en) * 2014-09-22 2016-03-31 北京交通大学 Passenger flow estimation method for urban rail network
US10176435B1 (en) * 2015-08-01 2019-01-08 Shyam Sundar Sarkar Method and apparatus for combining techniques of calculus, statistics and data normalization in machine learning for analyzing large volumes of data
CN108108897A (en) * 2017-12-20 2018-06-01 中国科学院深圳先进技术研究院 A kind of track traffic for passenger flow classifying method, system and electronic equipment

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113536493A (en) * 2021-06-23 2021-10-22 南京理工大学 Effective path generation method based on clustering reverse thrust and section passenger flow estimation method

Also Published As

Publication number Publication date
CN111275482B (en) 2023-05-09

Similar Documents

Publication Publication Date Title
CN104318324B (en) Shuttle Bus website and route planning method based on taxi GPS records
CN111191816B (en) System for identifying travel time chain of urban rail transit passengers
CN103984994B (en) Method for predicting urban rail transit passenger flow peak duration
CN107729938B (en) Rail station classification method based on bus connection radiation zone characteristics
CN103632212A (en) System and method for predicating time-varying user dynamic equilibrium network-evolved passenger flow
Wilson et al. The potential impact of automated data collection systems on urban public transport planning.
CN112418562A (en) Urban rail transit network passenger trip scheme estimation method
CN107656987A (en) A kind of subway station function method for digging based on LDA models
CN109903553A (en) The bus that multi-source data excavates is got on or off the bus station recognition and the method for inspection
CN107610282A (en) A kind of bus passenger flow statistical system
CN110889092A (en) Short-time large-scale activity peripheral track station passenger flow volume prediction method based on track transaction data
CN106846214A (en) Method of the analysis transport hub accessibility to region public transportation mode competitive influence
CN115953186A (en) Network appointment demand pattern recognition and short-time demand prediction method
Li et al. Using smart card data trimmed by train schedule to analyze metro passenger route choice with synchronous clustering
CN113221472B (en) Passenger flow prediction method based on LSTM
CN111275482B (en) Machine learning-based real-time dynamic track traffic sorting method
CN113468243A (en) Subway passenger flow analysis and prediction method and system
CN109389240A (en) Trip mode discrimination method based on big data machine learning
Chen et al. Research on the classification of urban rail transit stations-taking Shanghai metro as an example
CN116451867A (en) Subway short-time passenger flow prediction method based on space-time travel path calibration
CN110532515A (en) Urban track traffic passenger&#39;s stroke Inversion System based on AFC and video data
CN110610446A (en) County town classification method based on two-step clustering thought
Yue et al. Classification and determinants of high-speed rail stations using multi-source data: A case study in Jiangsu Province, China
CN111931968B (en) Method for optimizing public bicycle station layout by using mobile phone GPS data
CN114119314A (en) Urban network comprehensive analysis method and system based on multi-source element flow

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
CB02 Change of applicant information
CB02 Change of applicant information

Address after: 310053 23-25, 2 building, 352 BINKANG Road, Binjiang District, Hangzhou, Zhejiang.

Applicant after: Zhejiang zhongkong Information Industry Co.,Ltd.

Address before: 310053 23-25, 2 building, 352 BINKANG Road, Binjiang District, Hangzhou, Zhejiang.

Applicant before: ZHEJIANG SUPCON INFORMATION TECHNOLOGY Co.,Ltd.

GR01 Patent grant
GR01 Patent grant