CN114943374A - Subway OD (optical density) quantity prediction method, equipment and storage medium - Google Patents

Subway OD (optical density) quantity prediction method, equipment and storage medium Download PDF

Info

Publication number
CN114943374A
CN114943374A CN202210542245.1A CN202210542245A CN114943374A CN 114943374 A CN114943374 A CN 114943374A CN 202210542245 A CN202210542245 A CN 202210542245A CN 114943374 A CN114943374 A CN 114943374A
Authority
CN
China
Prior art keywords
prediction
term
short
time period
long
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
Application number
CN202210542245.1A
Other languages
Chinese (zh)
Inventor
李子牧
张蕾
王伟
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.)
Traffic Control Technology TCT Co Ltd
Original Assignee
Traffic Control Technology TCT 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 Traffic Control Technology TCT Co Ltd filed Critical Traffic Control Technology TCT Co Ltd
Priority to CN202210542245.1A priority Critical patent/CN114943374A/en
Publication of CN114943374A publication Critical patent/CN114943374A/en
Pending legal-status Critical Current

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
    • 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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • 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
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • 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
    • G06N3/045—Combinations of networks
    • 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)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Business, Economics & Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • Strategic Management (AREA)
  • Computing Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Mathematical Physics (AREA)
  • General Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Evolutionary Biology (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Medical Informatics (AREA)
  • Primary Health Care (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The application provides a subway OD quantity prediction method, equipment and a storage medium, wherein the method comprises the following steps: determining an OD (optical density) long-term predicted value of a predicted time period; acquiring a current measurement variable; performing OD short-term prediction on the prediction time period according to the measurement variable to obtain an OD short-term prediction value; and synthesizing the OD quantity short-term predicted value and the OD quantity long-term predicted value to obtain the final OD quantity predicted value of the predicted time period. According to the method, the OD quantity short-term predicted value and the OD quantity long-term predicted value are integrated to obtain the OD quantity final predicted value in the prediction time period, so that the OD quantity final predicted value has the characteristics of few prediction parameters of OD quantity long-term prediction and short-term prediction accuracy, the prediction complexity is reduced while accurate prediction is provided, and further the reliable basis of operation plan design is provided for the subway operator.

Description

Subway OD (optical Density) quantity prediction method, equipment and storage medium
Technical Field
The application relates to the technical field of rail transit, in particular to a subway OD (optical density) quantity prediction method, equipment and a storage medium.
Background
The Origin-Destination (Origin-Destination) quantity of a subway is an important basis for describing the quantity of pedestrian travel traffic data between all starting points and all end points in a subway traffic network, reflecting the basic requirements of pedestrians on the subway traffic network, and carrying out subway traffic management and vehicle allocation and traffic change planning. The OD is typically expressed as a start-end passenger flow matrix divided by a time granularity (e.g., 5 minutes, 15 minutes, 1 hour, etc.).
The accurate OD quantity has important guiding significance for the design of a subway line running scheme and the adjustment of a running plan, so that how to acquire the accurate OD quantity in a long term or a trend and in a short term is always a key concern of subway operation.
For long-term prediction of the OD amount, namely obtaining the OD amount at a certain day/moment after a long period of time, the current condition basically does not play a guiding role. Short-term prediction of OD amounts refers to predicting OD amounts in a short time (several minutes to several hours), in which case the current state (OD amount, weather, etc.) will play a great guiding role, so short-term prediction results are often more accurate than long-term prediction, and if short-term prediction results can be obtained quickly and responded in time, the short-term prediction can play a greater role.
Therefore, how to accurately measure the OD amount becomes a problem to be solved at present.
Disclosure of Invention
In order to accurately measure the OD quantity, the application provides a subway OD quantity prediction method, subway OD quantity prediction equipment and a subway OD quantity storage medium.
In a first aspect of the present application, a method for predicting an OD of a subway is provided, the method including:
determining a long-term predicted value of a starting-to-end OD (origin-destination) quantity of a predicted time period;
acquiring a current measurement variable;
performing OD short-term prediction on the prediction time period according to the measurement variable to obtain an OD short-term prediction value;
and synthesizing the OD quantity short-term predicted value and the OD quantity long-term predicted value to obtain the OD quantity final predicted value of the prediction time period.
Optionally, before determining the long-term predicted value of the origin-destination OD amount of the prediction time period, the method further includes:
obtaining OD sample data;
acquiring a date characteristic label, wherein the date characteristic label is a week label and/or a holiday label;
constructing a training set according to the OD sample data and the date feature label;
training a neural network model based on the training set to obtain an OD long-term prediction model;
determining an OD quantity long-term prediction value of each time period according to the OD quantity long-term prediction model, wherein each time period comprises the prediction time period;
the determining of the long-term predicted value of the origin-destination OD of the predicted time period comprises the following steps:
and obtaining the OD quantity long-term predicted value of the predicted time period from the OD quantity long-term predicted value of each time period.
Optionally, the obtaining OD sample data includes:
obtaining card swiping records of all users from a platform gate machine;
determining OD trip data and corresponding trip time of each user according to the card swiping record;
and determining the total amount of OD trip data in each time period according to the trip time to form OD sample data.
Optionally, the card swiping record includes a user identifier, card swiping time, and a card swiping station identifier;
determining OD trip data and corresponding trip time of each user according to the card swiping record, wherein the determining comprises the following steps:
determining a card swiping record of each user according to the user identification;
sequentially selecting a card swiping record of a user, and determining an initial station, a termination station and a trip time period of each trip according to the card swiping time and the card swiping station identifier; determining target trips according to the starting station and the ending station, taking all the target trips as OD trip data, and determining trip time periods of the target trips as trip times corresponding to the OD trip data.
Optionally, the constructing a training set according to the OD sample data and the date feature tag includes:
and splicing and combining the total amount of the OD trip data in each time period with the date characteristic label to form a training set.
Optionally, the measured variable comprises one or more of: OD amount in a preset time period, traffic congestion degree, pedestrian flow of an OD starting station and vectorized weather information.
Optionally, the performing short-term OD measurement prediction on the prediction time period according to the measurement variable to obtain a short-term OD measurement prediction value includes:
determining a plurality of predictive models;
training each prediction model based on the measurement variables and the training set to obtain a plurality of OD short-term prediction models;
determining an OD short-term prediction to-be-processed value of a prediction time period according to each OD short-term prediction model;
determining the average value of all OD quantity short-term prediction to-be-processed values;
calculating the deviation between each OD short-term prediction to-be-processed value and the average value;
and determining the short-term predicted value of the OD quantity according to the deviation.
Optionally, the plurality of predictive models includes at least two of: the system comprises a neural network model, a long-term and short-term memory network model, a gradient lifting decision tree model, a support vector regression model and a convolution neural network model.
In a second aspect of the present application, there is provided an electronic device comprising:
a memory;
a processor; and
a computer program;
wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method according to the first aspect.
In a third aspect of the present application, there is provided a computer-readable storage medium having a computer program stored thereon; the computer program is executed by a processor to implement the method according to the first aspect as described above.
The application provides a subway OD quantity prediction method, equipment and a storage medium, wherein the method comprises the following steps: determining an OD (optical density) long-term predicted value of a predicted time period; acquiring a current measurement variable; performing OD short-term prediction on the prediction time period according to the measurement variable to obtain an OD short-term prediction value; and synthesizing the OD amount short-term predicted value and the OD amount long-term predicted value to obtain the final OD amount predicted value of the prediction time period.
According to the method, the OD quantity short-term predicted value and the OD quantity long-term predicted value are integrated to obtain the OD quantity final predicted value in the prediction time period, so that the OD quantity final predicted value has the characteristics of few prediction parameters of OD quantity long-term prediction and short-term prediction accuracy, the prediction complexity is reduced while accurate prediction is provided, and further the reliable basis of operation plan design is provided for the subway operator.
In addition, in one implementation, a training set is constructed according to OD sample data and date feature labels, an OD long-term prediction model is trained on the basis of the training set, and then an OD long-term prediction value of a prediction time period is determined according to the OD long-term prediction model, so that a specific implementation scheme of OD long-term prediction is provided. According to the scheme, long-term prediction of the OD quantity can be completed through the minimum amount of data, the long-term prediction requirement of the OD quantity can be met to the greatest extent when the data are incomplete, and meanwhile, only one long-term prediction model of the OD quantity needs to be established and trained independently when each new OD appears, so that the time cost for building the long-term prediction model of the OD quantity is greatly reduced.
In addition, in one implementation, OD sample data is formed according to the card swiping record, accurate and efficient acquisition of the OD sample data is achieved, the prediction time of the long-term predicted value of the OD quantity is shortened, and the prediction accuracy of the long-term predicted value of the OD quantity is improved.
In addition, in one implementation, the OD trip data and the corresponding trip time are determined according to the user identification, the card swiping time and the card swiping station identification, the determining scheme of the OD trip data and the corresponding trip time is further determined, and the OD trip data and the corresponding trip time are accurately and efficiently acquired.
In addition, in one implementation, the total amount of the OD trip data and the date characteristic labels in each time period are spliced and combined to form a training set, so that the training set not only has the attribute of the total amount of the OD trip data, but also has the date characteristic attribute, the OD amount can be accurately predicted aiming at a single date characteristic label, and the prediction precision is refined.
Additionally, in one implementation, the measured variable is specifically defined, specifying that the measured variable includes one or more of: OD amount, traffic congestion degree, pedestrian flow of an OD starting station and vectorized weather information in a preset time period can enrich the short-term prediction variable of the OD amount through the measurement variable, so that the short-term prediction value of the OD amount is more accurate.
In one implementation, short-term OD predictions are made by a plurality of short-term OD prediction models, and a final short-term OD prediction is determined from the plurality of short-term OD prediction results. Because the situation that prediction with large deviation is generated for some special input combinations may inevitably occur in any model, the usability of prediction is reduced, and the final OD quantity short-term prediction result is determined from a plurality of OD quantity short-term prediction results, so that the accidental large error situation caused by a single OD quantity short-term prediction model can be effectively avoided, and the prediction precision is improved.
In addition, in one implementation, the prediction model is specifically limited, and the accidental large error condition caused by a single OD short-term prediction model is reduced through abundant prediction models, so that the prediction precision is improved.
According to the electronic equipment, the computer program is executed by the processor to synthesize the OD quantity short-term predicted value and the OD quantity long-term predicted value to obtain the OD quantity final predicted value of the prediction time period, so that the OD quantity final predicted value has the characteristics of few prediction parameters of the OD quantity long-term prediction and the prediction accuracy of the short-term prediction, the prediction complexity is reduced while the accurate prediction is provided, and the reliable basis of the operation plan design is provided for the subway operator.
According to the computer-readable storage medium, the computer program is executed by the processor to synthesize the OD quantity short-term predicted value and the OD quantity long-term predicted value to obtain the OD quantity final predicted value of the prediction time period, so that the OD quantity final predicted value has the characteristics of few prediction parameters of the OD quantity long-term prediction and the prediction accuracy of the short-term prediction, the prediction complexity is reduced while the accurate prediction is provided, and the reliable basis of the operation plan design is provided for the subway operator.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
fig. 1 is a schematic flow chart of a method for predicting an OD of a subway according to an embodiment of the present disclosure;
fig. 2 is a schematic diagram of a neural network model for predicting long-term predicted values of OD quantities in various time periods according to an embodiment of the present disclosure;
FIG. 3 is a schematic diagram of a short-term prediction model of OD measurement of a long-term short-term memory network according to an embodiment of the present disclosure;
fig. 4 is a schematic diagram illustrating rolling update of a prediction result of a subway OD amount prediction method according to an embodiment of the present application.
Detailed Description
In order to make the technical solutions and advantages of the embodiments of the present application more apparent, the following further detailed description of the exemplary embodiments of the present application with reference to the accompanying drawings makes it clear that the described embodiments are only a part of the embodiments of the present application, and are not exhaustive of all embodiments. It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
In the process of implementing the present application, the inventor finds that the current situation basically does not play a guiding role in long-term prediction of the OD amount, i.e., obtaining the OD amount at a certain day/time after a long period of time. Short-term prediction of OD amounts refers to predicting OD amounts in a short time (several minutes to several hours), in which case the current state (OD amount, weather, etc.) will play a great guiding role, so short-term prediction results are often more accurate than long-term prediction, and if short-term prediction results can be obtained quickly and responded in time, the short-term prediction can play a greater role. Therefore, how to accurately measure the OD amount becomes a problem to be solved at present.
In view of the foregoing problems, embodiments of the present application provide a method, an apparatus, and a storage medium for predicting an OD of a subway, where the method includes: determining an OD (optical density) long-term predicted value of a predicted time period; acquiring a current measurement variable; performing OD short-term prediction on the prediction time period according to the measurement variable to obtain an OD short-term prediction value; and synthesizing the OD amount short-term predicted value and the OD amount long-term predicted value to obtain the final OD amount predicted value of the prediction time period. According to the method and the device, the OD quantity short-term predicted value and the OD quantity long-term predicted value are integrated to obtain the OD quantity final predicted value in the prediction time period, so that the OD quantity final predicted value has the characteristics of few prediction parameters of OD quantity long-term prediction and short-term prediction accuracy, the prediction complexity is reduced while accurate prediction is provided, and further a reliable basis of operation plan design is provided for a subway operator.
The OD amount is a traffic traveling amount between the starting and ending points. "O" is derived from ORIGIN, and refers to the departure location of the trip (i.e., the start site in the present embodiment), "D" is derived from DESTINATION, and refers to the DESTINATION of the trip (i.e., the end site in the present embodiment).
When the traffic travel amount between a certain starting station and a certain ending station needs to be predicted, the method for predicting the subway OD amount provided by the embodiment is executed.
For convenience of description, in this and subsequent embodiments, a starting station to a terminating station is referred to as an OD pair, where O of the OD pair is the starting station and D is the terminating station.
For example, if the traffic volume between a station a and a station B needs to be predicted, O in the OD pair is station a and D is station B.
Referring to fig. 1, the implementation process of the subway OD amount prediction method provided by this embodiment is as follows:
and 101, determining an OD (optical density) long-term predicted value of the prediction time period.
Before the step is executed, the long-term predicted value of the OD quantity of each time period is predicted, and each time period comprises a predicted time period. And obtaining the OD quantity long-term predicted value of the predicted time period from the OD quantity long-term predicted value of each time period.
Still taking the example of predicting the traffic volume of one hour in the future between the station a and the station B, the long-term predicted value of the OD volume of each hour today is predicted before the step 101 is executed, that is, there are 24 long-term predicted values of the OD volume before the step 101 is executed, each long-term predicted value of the OD volume corresponds to 1 hour today, and the hours corresponding to the long-term predicted values of the OD volume are different. Then a long-term predicted value of OD corresponding to one hour in the future is obtained from the 24 long-term predicted values of OD in step 101.
For the implementation schemes of predicting the long-term predicted value of the OD quantity of each time period, there are two existing schemes,
the first method is that modeling is performed from a single station, and the model predicts the passenger flow of the station entering and exiting the station according to the data of the station entering and exiting the station, and the model is very convenient and fast to make a passenger flow strategy in the station, but cannot support OD prediction, namely, passengers entering and exiting the station cannot be matched, so that the predicted number of passengers entering and exiting the station is different, and a corresponding OD pair cannot be generated.
The second method is to represent the OD quantity of the whole network as an O-D matrix and predict the OD quantity based on the OD quantity, which is just opposite to the first method, although the generation of the station quantity of the whole network can be ensured, the formalized matrix needs to integrate the data of each station of the whole network, which is very inconvenient in data preparation, and in addition, because the number of the stations is equal to the dimension of the matrix, the structure of the model can be changed when the concerned station is changed.
In this embodiment, when determining an implementation scheme for predicting a long-term predicted value of the OD amount in each time period, considering that when configuring a running scheme according to OD prediction, the OD amounts of the line and some related transfer lines often need to be concerned, while the OD prediction cannot be implemented by the existing first method, and the data in the whole network is considered to be redundant by the existing second method, therefore, the embodiment combines the existing two methods, and adopts a sub-OD model method for modeling not according to a station but according to an OD pair: each OD pair is modeled separately and the model is trained using only this OD data, so that a comprehensive large model of the line and road network is represented by a relevant combination of OD pair small models. The method can meet the requirements of modeling and predicting the OD quantity by using the minimum amount of data, can meet the preliminary modeling requirements to the maximum extent when the data are incomplete, and only needs to establish and train a small model to be added and combined when a new OD appears, so that the time consumption is less than that of a recombined data training wire-level model.
Specifically, the implementation process of predicting the long-term predicted value of the OD amount in each time period is as follows:
1. and obtaining OD sample data.
OD sample data can be obtained through card swiping records during implementation.
For example,
1) and obtaining the card swiping record of each user from the platform gate.
The card swiping record comprises a user identifier, card swiping time, card swiping station identifier, card swiping type (inbound/outbound), and the like.
2) And determining OD travel data and corresponding travel time of each user according to the card swiping record.
For example, the card swiping records are sorted according to time, and for the record of the same user identifier, the inbound card swiping record and the next earliest outbound card swiping record thereof constitute one trip of the user (since the user may go out many times a day, it is necessary that the user is a pair of records which are adjacent in time sequence and go in and out first), that is, the inbound and outbound time, of the user can be obtained.
3) And determining the total amount of the OD trip data in each time period according to the trip time to form OD sample data.
Specifically, the card-swiping record of each user is determined according to the user identification. And sequentially selecting a card swiping record of a user, and determining a starting station, a stopping station and a travel time period of each travel according to the card swiping time and the card swiping station identifier. And determining target trips according to the starting station and the stopping station, taking all the target trips as OD trip data, and determining the trip time period of the target trips as the trip time corresponding to the OD trip data.
Wherein the starting and ending sites of the target trip are matched to the OD pairs. Taking the prediction of the traffic volume from the station a to the station B for one hour in the future as an example, the trip with the starting station as the station a and the ending station as the station B is determined as the target trip.
After all target trips are taken as OD trip data, the OD trip data only comprise trip data of which the starting station is station A and the ending station is station B.
In addition, the trip data includes the card swiping time of the site a and the card swiping time of the site B. For a certain target trip (for example, the target trip i), if the card swiping time at the site a is time ai and the card swiping time at the site B is time bi, the trip time period is from time ai to time bi. The travel time when the target travel i is taken as the travel data is from time ai to time bi.
The OD trip data and the corresponding trip time are determined according to the user identification, the card swiping time and the card swiping station identification, the determining scheme of the OD trip data and the corresponding trip time is further determined, and the OD trip data and the corresponding trip time are accurately and efficiently obtained.
In addition, OD sample data are formed through card swiping record, accurate and efficient acquisition of the OD sample data is achieved, prediction duration of the OD quantity long-term prediction value is shortened, and prediction accuracy of the OD quantity long-term prediction value is improved.
2. A date signature tag is obtained.
Wherein, the date characteristic label is a week label and/or a holiday label.
The date characteristic label is a label describing date characteristics, such as the week, holiday conditions and the like, but attributes such as weather, road conditions and the like which cannot be or are difficult to accurately predict cannot be used as the date characteristic label.
3. And constructing a training set according to the OD sample data and the date feature label.
Specifically, the total amount of the OD trip data in each time period and the date feature labels are spliced and combined to form a training set.
And combining each OD sample data and the date feature label to form a sample which can be used for training, wherein the set of all samples is a training set.
By splicing and combining the total amount of the OD trip data and the date characteristic labels in each time period, a training set is formed, so that the training set not only has the attribute of the total amount of the OD trip data, but also has the date characteristic attribute, the OD quantity can be accurately predicted aiming at a single date characteristic label, and the prediction precision is refined.
4. And training the neural network model based on the training set to obtain the OD long-term prediction model.
The structure of the neural network model may be as shown in fig. 2.
After enough samples exist in the training set, the samples in the training set are divided into three types, wherein one type is used for training, the other type is used for verifying, the other type is used for testing, and the neural network model is trained through the training set. In view of the advantages of strong universality and simple training of the neural network, different OD pairs can adopt the same network model, the model matching the OD can be obtained only by using different OD sample sets for training, and for some ODs which can not meet the accuracy requirement all the time, special adjustment or model establishment is needed.
5. And determining the long-term predicted value of the OD quantity in each time period according to the long-term prediction model of the OD quantity.
After the OD quantity long-term prediction model is obtained, the information of the OD quantity of the passenger flow in the whole day time period can be predicted through the model only by determining the label information of a certain future date.
For example, a preset time period granularity, e.g., 15 minutes, 1 hour. Taking 1 hour as an example, dividing 1 day into 24 time periods, and determining the long-term predicted value of the OD quantity in each time period according to the long-term prediction model of the OD quantity to obtain 24 long-term predicted values of the OD quantity.
In the process of predicting the long-term predicted value of the OD quantity in each time period, a training set is constructed according to OD sample data and date feature labels, an OD quantity long-term prediction model is trained on the basis of the training set, the long-term predicted value of the OD quantity in the prediction time period is further determined according to the OD quantity long-term prediction model, and a specific implementation scheme of the OD quantity long-term prediction is provided. According to the scheme, long-term prediction of the OD quantity can be completed through the minimum amount of data, the long-term prediction requirement of the OD quantity can be met to the greatest extent when the data are incomplete, and meanwhile, only one long-term prediction model of the OD quantity needs to be established and trained independently when each new OD appears, so that the time cost for building the long-term prediction model of the OD quantity is greatly reduced.
Therefore, the embodiment adopts a neural network to predict the OD quantity long-term value.
102, the current measured variable is obtained.
Wherein the measured variable comprises one or more of: the OD amount in the preset time period (such as the OD amount in the previous time period), the traffic congestion degree, the pedestrian flow of the OD starting station and vectorized weather information.
By the measuring variable, the OD quantity short-term prediction variable can be enriched, so that the OD quantity short-term prediction value is more accurate.
And 103, performing OD quantity short-term prediction on the prediction time period according to the measurement variable to obtain an OD quantity short-term prediction value.
In particular, the method comprises the following steps of,
1. a plurality of predictive models is determined.
For example, the plurality of predictive models includes at least two of: the system comprises a neural network model, a long-term and short-term memory network model, a gradient lifting decision tree model, a support vector regression model and a convolution neural network model.
The situation that prediction with large deviation is generated for some special input combinations can be inevitably generated by any model, so that the usability of prediction is reduced, and therefore, the accidental large error situation caused by a single OD short-term prediction model is reduced through abundant prediction models, and the prediction precision is improved.
2. And training each prediction model based on the measured variables and a training set to obtain a plurality of OD short-term prediction models.
3. And determining the OD quantity short-term prediction to-be-processed value of the prediction time period according to each OD quantity short-term prediction model.
4. And determining the average value of all the short-term predicted to-be-processed values of the OD quantity.
5. And calculating the deviation between each OD quantity short-term prediction to-be-processed value and the average value.
6. And determining the short-term predicted value of the OD quantity according to the deviation.
The OD amount short-term prediction has more available information than the OD amount long-term prediction, and information such as passenger flow, weather, road conditions, etc. in the current/previous time period can provide important references, so that the short-term OD amount prediction model needs to be more complex in order to better use the abundant variables. The model also adopts a sub-OD modeling method, and a plurality of models are used for cooperative prediction to achieve better prediction accuracy.
The neural network: the OD short-term prediction model can also use a neural network, the process is basically the same as that of OD long-term prediction, but the input dimensionality needs to be increased, including the OD amount, the traffic congestion degree, the pedestrian volume of the OD starting station, the vectorized weather information and the like in the previous time period, and meanwhile, the complexity of the neural network is also improved.
② Long Short Term Memory network (LSTM, Long Short-Term Memory): the network adds a door controlled by the previous input on the basis of a general neural network, so that the network has certain long-term and short-term memory capability when processing time series information, therefore, the LSTM can be used for taking the passenger flow volume of a period of time as a time series, and the OD passenger flow volume of a future period of time is predicted in a rolling way by inputting real-time information such as road conditions, traffic conditions and the like, as shown in figure 3.
And thirdly, a Gradient Boosting Decision Tree (GBDT) is an iterative Decision Tree algorithm, the algorithm consists of a plurality of Decision trees, and the conclusions of all the trees are accumulated to be used as a final answer. The idea is to divide the input of each layer in a minimum error mode according to the number of trees in the layer, and represent the tree by the average value of sub-nodes in each tree as the node input value of the previous level. The GDBT has no memory, so the input variable is consistent with the neural network model, but has better accuracy and faster operation when the number of samples is less than that of the neural network model, thereby being capable of well complementing the method.
In addition, the model can be established and participated in prediction by using Support Vector Regression (SVR) model, Convolutional Neural Network (CNN) model and other modes.
Due to the characteristics of the machine learning model, the prediction condition with large deviation generated for some special input combinations may inevitably occur in any model, so that the prediction usability is reduced, therefore, under the condition that the calculation capacity is allowed, different models as much as possible can be used for simultaneous prediction, the deviation between the result of each model and the average value is calculated after the predicted result is averaged, and thus a model with the maximum deviation is eliminated or the model is continuously eliminated until the deviation is reduced to a certain range, so that the accidental large error condition of the model which can occur singly is avoided, the integral precision of the model is improved, and therefore, the implementation process of determining the short-term predicted value of the OD quantity according to the deviation in the step 6 is as follows: firstly, determining the mean value of all OD quantity short-term prediction to-be-processed values, then calculating the deviation between each OD quantity short-term prediction to-be-processed value and the mean value, determining whether the variance or the eliminated number meets the requirement, and if so, taking the mean value of the current remaining OD quantity short-term prediction to-be-processed values as the OD quantity short-term prediction value. And if the OD quantity short-term prediction model does not meet the requirement, eliminating the OD quantity short-term prediction value with the largest deviation, and further eliminating the OD quantity short-term prediction model with the OD quantity short-term prediction value with the largest deviation.
In addition, during specific elimination, a variance value can be added, namely, the OD quantity short-term prediction model with the maximum variance is eliminated.
The following tests were performed on Hangzhou subway card swiping data published by Ali cloud, and a neural network model, a support vector regression model, XGBOST and LightGBM (GBDT variants with two different operation modes) models were trained respectively, and the results predicted by each model were directly averaged and fused with the above algorithm respectively (one model result with the largest variance was discarded for each prediction), and the RMSE (root mean square error) is shown in Table 1.
TABLE 1 short-term prediction model for each OD
Model (model) RMSE (human)
NN 9.55
SVR 9.22
XGBoost 9.65
LightGBM 9.65
Average of results 9.19
Result fusion 9.12
It can be seen that the average result is improved to some extent compared with each original result, and the method using the proposal is improved compared with the method of simply averaging.
And 104, synthesizing the OD quantity short-term predicted value and the OD quantity long-term predicted value to obtain the final OD quantity predicted value of the prediction time period.
As shown in fig. 4, for subway operation companies, the method provided in this embodiment first uses an OD quantity long-term prediction model to predict OD quantities for a long period of time (one week, etc.) in the future, so as to design a preliminary operation plan. Then, with the acquisition of actual passenger flow, weather and road condition information, a more accurate short-time (several hours or the next day) result is predicted by using the OD short-time prediction model, and a scheme is adjusted or a plan is started in time.
The method for predicting the OD (origin-destination) quantity of the subway is a method for predicting the OD quantity of the subway based on machine learning, and long-term and short-term prediction of the OD quantity of the subway is comprehensively carried out by integrating various machine learning algorithms, so that reference is provided for setting and adjusting a subway operation scheme.
Specifically, 1) by using a neural network with few parameters and a simple model, the OD quantity of any date can be predicted according to the week and the period, so that long-term OD quantity prediction data can be obtained. 2) By using a neural network with more parameters and a complex model and combining other time series prediction models, the subsequent short-time OD quantity can be more accurately predicted according to the current weather, the OD quantity information of the previous period and the like. 3) By combining 1) and 2), a long-term prediction result is obtained according to historical data, and then a short-term prediction result is continuously updated along with the acquisition of the data, so that a long-term and short-term integrated OD prediction result is obtained.
The embodiment provides a subway OD quantity prediction method, which is used for determining an OD quantity long-term prediction value in a prediction time period; acquiring a current measurement variable; performing OD short-term prediction on the prediction time period according to the measurement variable to obtain an OD short-term prediction value; and synthesizing the OD amount short-term predicted value and the OD amount long-term predicted value to obtain the final OD amount predicted value of the prediction time period. The method provided by the embodiment synthesizes the OD quantity short-term predicted value and the OD quantity long-term predicted value to obtain the OD quantity final predicted value of the prediction time period, so that the OD quantity final predicted value has the characteristics of few prediction parameters of the OD quantity long-term prediction and short-term prediction accuracy, the prediction complexity is reduced while accurate prediction is provided, and further a reliable basis for operation plan design is provided for subway operators.
Based on the same inventive concept of the subway OD amount prediction method, the present embodiment provides an electronic device, including: memory, processor, and computer programs.
Wherein the computer program is stored in the memory and configured to be executed by the processor to implement the above-described subway OD amount prediction method.
In particular, the method comprises the following steps of,
and determining the OD quantity long-term predicted value of the prediction time period.
And acquiring the current measurement variable.
And performing OD quantity short-term prediction on the prediction time period according to the measurement variable to obtain an OD quantity short-term prediction value.
And synthesizing the OD quantity short-term predicted value and the OD quantity long-term predicted value to obtain the final OD quantity predicted value of the predicted time period.
Optionally, before determining the OD amount long-term prediction value of the prediction time period, the method further includes:
and obtaining OD sample data.
And acquiring a date characteristic label, wherein the date characteristic label is a week label and/or a holiday label.
And constructing a training set according to the OD sample data and the date feature label.
And training the neural network model based on the training set to obtain an OD (optical density) long-term prediction model.
And determining an OD quantity long-term prediction value of each time period according to the OD quantity long-term prediction model, wherein each time period comprises a prediction time period.
Determining an OD quantity long-term prediction value of a prediction time period, comprising:
and obtaining the OD quantity long-term predicted value of the predicted time period from the OD quantity long-term predicted value of each time period.
Optionally, obtaining OD sample data includes:
and obtaining the card swiping record of each user from the platform gate.
And determining OD travel data and corresponding travel time of each user according to the card swiping record.
And determining the total amount of the OD trip data in each time period according to the trip time to form OD sample data.
Optionally, the card swiping record includes a user identifier, a card swiping time, and a card swiping station identifier.
Determining OD trip data and corresponding trip time of each user according to the card swiping record, wherein the method comprises the following steps:
and determining the card swiping record of each user according to the user identification.
And sequentially selecting a card swiping record of a user, and determining a starting station, a stopping station and a travel time period of each travel according to the card swiping time and the card swiping station identifier. And determining target trips according to the starting station and the stopping station, taking all the target trips as OD trip data, and determining the trip time period of the target trips as the trip time corresponding to the OD trip data.
Optionally, constructing a training set according to the OD sample data and the date feature tag, including:
and splicing and combining the total amount of the OD trip data in each time period with the date characteristic label to form a training set.
Optionally, the measured variable comprises one or more of: OD amount in a preset time period, traffic congestion degree, pedestrian flow of an OD starting station and vectorized weather information.
Optionally, performing short-term OD measurement prediction on the prediction time period according to the measurement variable to obtain a short-term OD measurement prediction value, including:
a plurality of predictive models is determined.
And training each prediction model based on the measured variables and a training set to obtain a plurality of OD short-term prediction models.
And determining the OD quantity short-term prediction to-be-processed value of the prediction time period according to each OD quantity short-term prediction model.
And determining the average value of all the short-term predicted to-be-processed values of the OD quantity.
And calculating the deviation of each OD short-term prediction to-be-processed value and the average value.
And determining the short-term predicted value of the OD quantity according to the deviation.
Optionally, the plurality of predictive models includes at least two of: the system comprises a neural network model, a long-term and short-term memory network model, a gradient lifting decision tree model, a support vector regression model and a convolution neural network model.
In the electronic device provided by this embodiment, the computer program is executed by the processor to synthesize the short-term predicted value and the long-term predicted value of the OD amount, so as to obtain the final predicted value of the OD amount in the prediction time period, so that the final predicted value of the OD amount has the characteristics of few prediction parameters of the long-term prediction of the OD amount and the prediction accuracy of the short-term prediction, and the prediction complexity is reduced while providing accurate prediction, thereby providing a reliable basis for the operation plan design for the subway operator.
Based on the same inventive concept of the subway OD amount prediction method, the present embodiment provides a computer on which a computer program can be stored. The computer program is executed by a processor to implement the above-described subway OD amount prediction method.
In particular, the method comprises the following steps of,
and determining the OD quantity long-term predicted value of the prediction time period.
And acquiring the current measurement variable.
And performing OD quantity short-term prediction on the prediction time period according to the measurement variable to obtain an OD quantity short-term prediction value.
And synthesizing the OD amount short-term predicted value and the OD amount long-term predicted value to obtain the final OD amount predicted value of the prediction time period.
Optionally, before determining the OD amount long-term prediction value of the prediction time period, the method further includes:
and obtaining OD sample data.
And acquiring a date characteristic label, wherein the date characteristic label is a week label and/or a holiday label.
And constructing a training set according to the OD sample data and the date feature label.
And training the neural network model based on the training set to obtain an OD (optical density) long-term prediction model.
And determining the OD quantity long-term prediction value of each time period according to the OD quantity long-term prediction model, wherein each time period comprises a prediction time period.
Determining an OD quantity long-term prediction value of a prediction time period, comprising:
and obtaining the OD quantity long-term predicted value of the predicted time period from the OD quantity long-term predicted value of each time period.
Optionally, obtaining OD sample data includes:
and obtaining the card swiping record of each user from the platform gate.
And determining OD travel data and corresponding travel time of each user according to the card swiping record.
And determining the total amount of the OD trip data in each time period according to the trip time to form OD sample data.
Optionally, the card swiping record comprises a user identifier, a card swiping time, and a card swiping station identifier.
Determining OD trip data and corresponding trip time of each user according to the card swiping record, wherein the method comprises the following steps:
and determining the card swiping record of each user according to the user identification.
And sequentially selecting a card swiping record of a user, and determining the starting station, the stopping station and the trip time period of each trip according to the card swiping time and the card swiping station identification. And determining target trips according to the starting station and the stopping station, taking all the target trips as OD trip data, and determining the trip time period of the target trips as the trip time corresponding to the OD trip data.
Optionally, constructing a training set according to the OD sample data and the date feature tag, including:
and splicing and combining the total amount of the OD trip data in each time period with the date characteristic label to form a training set.
Optionally, the measured variable comprises one or more of: OD amount, traffic congestion degree, pedestrian flow of an OD starting station and vectorized weather information in a preset time period.
Optionally, performing short-term OD measurement prediction on the prediction time period according to the measurement variable to obtain a short-term OD measurement prediction value, including:
a plurality of predictive models is determined.
And training each prediction model based on the measured variables and a training set to obtain a plurality of OD quantity short-term prediction models.
And determining the OD quantity short-term prediction to-be-processed value of the prediction time period according to each OD quantity short-term prediction model.
And determining the average value of all the short-term predicted to-be-processed values of the OD quantity.
And calculating the deviation between each OD quantity short-term prediction to-be-processed value and the average value.
And determining the short-term predicted value of the OD quantity according to the deviation.
Optionally, the plurality of predictive models includes at least two of: the system comprises a neural network model, a long-term and short-term memory network model, a gradient lifting decision tree model, a support vector regression model and a convolution neural network model.
In the computer-readable storage medium provided in this embodiment, the computer program on the computer-readable storage medium is executed by the processor to synthesize the short-term predicted value and the long-term predicted value of the OD volume, so as to obtain the final predicted value of the OD volume in the prediction time period, so that the final predicted value of the OD volume has the characteristics of few prediction parameters of the long-term prediction of the OD volume and the accurate prediction characteristic of the short-term prediction, and the prediction complexity is reduced while providing the accurate prediction, thereby providing a reliable basis for the operation plan design for the subway operator.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and so forth) having computer-usable program code embodied therein. The scheme in the embodiment of the application can be implemented by adopting various computer languages, such as object-oriented programming language Java and transliterated scripting language JavaScript.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While the preferred embodiments of the present application have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all alterations and modifications as fall within the scope of the application.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.

Claims (10)

1. A subway OD quantity prediction method is characterized by comprising the following steps:
determining a long-term predicted value of a starting-to-end OD (origin-destination) quantity of a predicted time period;
acquiring a current measurement variable;
performing OD short-term prediction on the prediction time period according to the measurement variable to obtain an OD short-term prediction value;
and synthesizing the OD quantity short-term predicted value and the OD quantity long-term predicted value to obtain the OD quantity final predicted value of the prediction time period.
2. The method of claim 1, wherein prior to determining the long-term prediction value of the origin-destination OD amount for the prediction time period, further comprising:
obtaining OD sample data;
acquiring a date characteristic label, wherein the date characteristic label is a week label and/or a holiday label;
constructing a training set according to the OD sample data and the date feature label;
training a neural network model based on the training set to obtain an OD long-term prediction model;
determining an OD quantity long-term prediction value of each time period according to the OD quantity long-term prediction model, wherein each time period comprises the prediction time period;
the determining of the long-term predicted value of the origin-destination OD amount of the prediction time period comprises:
and obtaining the OD quantity long-term predicted value of the predicted time period from the OD quantity long-term predicted value of each time period.
3. The method of claim 2, wherein the obtaining OD sample data comprises:
obtaining card swiping records of all users from a platform gate machine;
determining OD trip data and corresponding trip time of each user according to the card swiping record;
and determining the total amount of OD trip data in each time period according to the trip time to form OD sample data.
4. The method of claim 3, wherein the swipe record includes a user identification, a swipe time, a swipe site identification;
determining OD trip data and corresponding trip time of each user according to the card swiping record, wherein the determining comprises the following steps of:
determining a card swiping record of each user according to the user identification;
sequentially selecting a card swiping record of a user, and determining an initial station, a termination station and a trip time period of each trip according to the card swiping time and the card swiping station identifier; determining target trips according to the starting station and the ending station, taking all the target trips as OD trip data, and determining trip time periods of the target trips as trip times corresponding to the OD trip data.
5. The method of claim 3, wherein said constructing a training set from said OD sample data and date signature comprises:
and splicing and combining the total amount of the OD trip data in each time period with the date characteristic label to form a training set.
6. The method of claim 1, wherein the measured variable comprises one or more of: OD amount in a preset time period, traffic congestion degree, pedestrian flow of an OD starting station and vectorized weather information.
7. The method of claim 1, wherein the performing OD short-term prediction on the prediction time period according to the measured variable to obtain an OD short-term prediction value comprises:
determining a plurality of predictive models;
training each prediction model based on the measurement variables and the training set to obtain a plurality of OD short-term prediction models;
determining an OD short-term prediction to-be-processed value of a prediction time period according to each OD short-term prediction model;
determining the average value of all OD quantity short-term prediction to-be-processed values;
calculating the deviation between each OD quantity short-term prediction to-be-processed value and the average value;
and determining the short-term predicted value of the OD quantity according to the deviation.
8. The method of claim 7, wherein the plurality of predictive models includes at least two of: a neural network model, a long-term and short-term memory network model, a gradient lifting decision tree model, a support vector regression model and a convolution neural network model.
9. An electronic device, comprising:
a memory;
a processor; and
a computer program;
wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method of any one of claims 1-8.
10. A computer-readable storage medium, having stored thereon a computer program; the computer program is executed by a processor to implement the method of any one of claims 1-8.
CN202210542245.1A 2022-05-18 2022-05-18 Subway OD (optical density) quantity prediction method, equipment and storage medium Pending CN114943374A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210542245.1A CN114943374A (en) 2022-05-18 2022-05-18 Subway OD (optical density) quantity prediction method, equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210542245.1A CN114943374A (en) 2022-05-18 2022-05-18 Subway OD (optical density) quantity prediction method, equipment and storage medium

Publications (1)

Publication Number Publication Date
CN114943374A true CN114943374A (en) 2022-08-26

Family

ID=82907019

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210542245.1A Pending CN114943374A (en) 2022-05-18 2022-05-18 Subway OD (optical density) quantity prediction method, equipment and storage medium

Country Status (1)

Country Link
CN (1) CN114943374A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115439206A (en) * 2022-11-08 2022-12-06 税友信息技术有限公司 Declaration data prediction method, device, equipment and medium

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115439206A (en) * 2022-11-08 2022-12-06 税友信息技术有限公司 Declaration data prediction method, device, equipment and medium
CN115439206B (en) * 2022-11-08 2023-03-07 税友信息技术有限公司 Declaration data prediction method, device, equipment and medium

Similar Documents

Publication Publication Date Title
US11143522B2 (en) Travel time prediction method, apparatus and server
CN109214584B (en) Method and device for predicting passenger flow
Chiappone et al. Traffic simulation models calibration using speed–density relationship: An automated procedure based on genetic algorithm
CN110119845A (en) A kind of application method of track traffic for passenger flow prediction
CN108444486B (en) Navigation route sorting method and device
CN110889558A (en) Road condition prediction method and device
CN109388480A (en) A kind of method and device handling cloud resource
Nuzzolo et al. A subjective optimal strategy for transit simulation models
CN114943374A (en) Subway OD (optical density) quantity prediction method, equipment and storage medium
Wepulanon et al. A real-time bus arrival time information system using crowdsourced smartphone data: a novel framework and simulation experiments
Pirc et al. Highway travel time estimation using multiple data sources
CN114491168B (en) Method and system for regulating and controlling cloud sample data sharing, computer equipment and storage medium
Büchel et al. What do we know when? modeling predictability of transit operations
CN116662815B (en) Training method of time prediction model and related equipment
CN111507541B (en) Goods quantity prediction model construction method, goods quantity measurement device and electronic equipment
CN111160594B (en) Method and device for estimating arrival time and storage medium
Kieu et al. Stochastic collective model of public transport passenger arrival process
Zhang et al. A Dynamic Short‐Turning Bus Control for Uncertain Demand
CN111079175A (en) Data processing method, data processing device, computer readable storage medium and computer equipment
Nuzzolo et al. Real-time modelling of normative travel strategies on unreliable dynamic transit networks: a framework analysis
CN112785044B (en) Real-time full-load rate prediction method, device, equipment and medium for public transport means
CN114463978A (en) Data monitoring method based on rail transit information processing terminal
CN115049464A (en) Operator package recommendation method and device and computer readable storage medium
CN114241385A (en) Subway station entrance and exit passenger flow prediction method based on deep learning
Seman et al. A framework to estimate dwell time of BRT systems using fuzzy regression

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