CN107194505B - Method and system for predicting bus traffic based on urban big data - Google Patents
Method and system for predicting bus traffic based on urban big data Download PDFInfo
- Publication number
- CN107194505B CN107194505B CN201710324534.3A CN201710324534A CN107194505B CN 107194505 B CN107194505 B CN 107194505B CN 201710324534 A CN201710324534 A CN 201710324534A CN 107194505 B CN107194505 B CN 107194505B
- Authority
- CN
- China
- Prior art keywords
- bus
- data
- prediction model
- line
- layer
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 19
- 239000013598 vector Substances 0.000 claims abstract description 66
- 238000012549 training Methods 0.000 claims abstract description 26
- 238000010219 correlation analysis Methods 0.000 claims abstract description 8
- 230000006870 function Effects 0.000 claims description 24
- 238000007477 logistic regression Methods 0.000 claims description 18
- 230000003068 static effect Effects 0.000 claims description 11
- 238000004140 cleaning Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 2
- 230000009286 beneficial effect Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 2
- 238000011835 investigation Methods 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 238000012098 association analyses Methods 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/40—Business processes related to the transportation industry
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a system for predicting the traffic volume of a bus based on urban big data, wherein the method is realized by the following steps: extracting sample data from urban big data, calculating the bus traffic volume on each bus line according to the bus IC card data on each bus line, performing correlation analysis on the sample data, extracting a feature set related to the bus traffic volume on each bus line, and taking the bus traffic volume and the feature set as input vectors; carrying out unsupervised pre-training and supervised fine tuning on the prediction model by using the input vector to obtain a trained prediction model; and acquiring data of the line to be predicted, extracting a feature set of the line to be predicted, and inputting the feature set of the line to be predicted into a trained prediction model to obtain the bus traffic of the line to be predicted. The method utilizes the input vector to train the prediction model, and can be used for accurately predicting the future bus traffic of the line to be predicted.
Description
Technical Field
The invention belongs to the field of smart cities, and particularly relates to a method and a system for predicting the traffic volume of a bus based on urban big data.
Background
With the rapid promotion of the urbanization process and the rapid increase of economy in China, the population and the motor vehicle reserves of cities are increased rapidly, residents go out more and more frequently, and cities face severe traffic problems. On one hand, a large amount of urban traffic exceeds the bearing capacity of roads, so that traffic jam, a large amount of energy loss and environmental pollution are caused. On the other hand, the low carrying capacity and the relatively large road occupation area of private vehicles also cause a great deal of resource waste.
As urban public transport has the advantages of large carrying capacity, high transport efficiency, low energy consumption, small relative pollution and the like, more and more areas begin to actively promote urban public transport construction, and public transport priority development strategies are implemented forcefully so as to optimize urban transport structures, relieve traffic jams, save resources and reduce traffic carbon emission. The efficient operation of the urban public transportation system depends on investment of large-scale infrastructure and also depends on reasonable operation decision and scientific management means. A key factor for efficiently utilizing the urban public transportation system is to sense the running quantity of each bus line in a certain period of time in the future in real time. According to the passenger traffic volume of a certain period of time in the future on each bus line, the system assists the traffic decision management department to make a reasonable operation scheme so as to optimize urban traffic, improve the transportation scheduling level and meet the travel demands of people. However, due to the unique geographical and cultural environmental characteristics of each region, the amount of bus traffic varies from time to time and from place to place for people in each region of a city. Conventionally, traffic management departments count the traveling volume of each area through special investigation of each area, the investigation needs a lot of time and expensive labor cost, the period for acquiring data is long, and the traveling situation of passengers in a reaction area is limited.
Therefore, the traditional technology is used for predicting the passenger traffic in a certain period of time in the future on each bus line, a large amount of time and expensive labor cost are consumed, and the period for acquiring data is long.
Disclosure of Invention
The invention provides a method and a system for predicting the traffic volume of a bus based on urban big data, aiming at extracting input vectors based on the urban big data, carrying out unsupervised pre-training and supervised fine tuning on a prediction model by using the input vectors to obtain a trained prediction model, and predicting the traffic volume of the bus by using the trained prediction model, thereby solving the technical problems that the traditional technology predicts the traffic volume of passengers in a certain period of time in the future on each bus line, consumes a large amount of time and expensive labor cost, and has a long period for acquiring data.
To achieve the above object, according to one aspect of the present invention, there is provided a method for predicting a bus traffic based on urban big data, comprising:
(1) extracting sample data from the urban big data, wherein the sample data comprises: bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data;
(2) calculating the bus traffic volume on each bus line according to the bus IC card data on each bus line, performing correlation analysis on sample data, extracting a feature set related to the bus traffic volume on each bus line, and taking the bus traffic volume and the feature set as input vectors;
(3) carrying out unsupervised pre-training and supervised fine tuning on the prediction model by using the input vector to obtain a trained prediction model;
(4) the method comprises the steps of collecting bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data of a line to be predicted, extracting a feature set of the line to be predicted, inputting the feature set of the line to be predicted into a trained prediction model, and obtaining the bus traffic of the line to be predicted.
Further, the step (1) further comprises the step of cleaning redundant and incomplete data in the sample data.
Further, the specific implementation manner of performing association analysis on the sample data in step (2) is as follows: taking an influence area by taking each bus stop on a bus line as a center, and analyzing the association relation between the mobile characteristic, the geographic characteristic and the meteorological characteristic in each influence area and the resident trip by using sample data to obtain a characteristic set.
Further, the step (3) comprises the following steps:
(3-1) the prediction model comprises an encoding layer and a logistic regression layer, the feature set and the bus trip amount are input vectors of the prediction model, and the bus trip prediction amount is an output vector of the prediction model; in an encoding layer, carrying out unsupervised pre-training by using an input vector to obtain high-level representation of a feature set;
and (3-2) obtaining a bus trip prediction quantity by utilizing the high-level representation of the feature set in a logistic regression layer, establishing a loss function by utilizing the input vector, the bus trip prediction quantity and the parameters of the prediction model, and carrying out supervised fine tuning by using an encoding layer to adjust the parameters of the prediction model to minimize the loss function so as to obtain the trained prediction model.
Further, the specific implementation manner of the step (3-1) is as follows: the prediction model comprises a coding layer and a logistic regression layer, the feature set and the bus trip amount are input vectors of the prediction model, and the bus trip prediction amount is an output vector of the prediction model; the coding layer is composed of N layers of denoising automatic encoders, reconstruction errors of input vectors of each layer are minimized through the reconstruction denoising automatic encoders, and then unsupervised training is carried out on the denoising automatic encoders of each layer, so that high-level representation of the feature set is obtained.
Furthermore, the coding layer is composed of 4 layers of denoising automatic coders.
Further, the specific implementation manner of the step (3-2) is as follows: the logic layer is mainly composed of a logistic regression supervised learning algorithm, the high level representation of the feature set is input into the logic layer, and the predicted amount y of bus travel on the ith line is obtained(i)Using the predicted trip amount of the bus and the input vector x on the ith bus line(i)Establishing a loss function:
where β is a prediction model parameter, D(s)Representing a set of input vectors, P representing a function for predicting accuracy, and Y representing a set of bus trip prediction quantities;
and the coding layer is used for adjusting parameters of the prediction model to minimize the loss function, and carrying out supervised fine tuning to obtain the trained prediction model.
According to another aspect of the present invention, there is provided a system for predicting a bus traffic based on urban big data, comprising:
the sample data extracting module is used for extracting sample data from the urban big data, and the sample data comprises: bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data;
the input vector extraction module is used for calculating the bus traffic volume on the ith bus line according to the bus IC card data on the ith bus line, carrying out correlation analysis on the sample data, extracting a feature set related to the bus traffic volume on the ith bus line, and taking the bus traffic volume and the feature set as input vectors;
the training prediction model module is used for carrying out unsupervised pre-training and supervised fine tuning on the prediction model by utilizing the input vector to obtain a trained prediction model;
and the prediction module is used for acquiring bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data of the line to be predicted, extracting the feature set of the line to be predicted, and inputting the feature set of the line to be predicted into the trained prediction model to obtain the bus traffic of the line to be predicted.
Further, the training prediction model module comprises:
the system comprises an unsupervised pre-training submodule and a prediction model, wherein the unsupervised pre-training submodule is used for carrying out unsupervised pre-training, the prediction model comprises a coding layer and a logistic regression layer, a feature set and a bus trip prediction quantity are input vectors of the prediction model, and the bus trip prediction quantity is an output vector of the prediction model; in an encoding layer, carrying out unsupervised pre-training by using an input vector to obtain high-level representation of a feature set;
and the supervision fine-tuning sub-module is used for obtaining the bus trip prediction quantity by utilizing the high-level representation of the characteristic set in the logistic regression layer, establishing a loss function by utilizing the input vector, the bus trip prediction quantity and the parameters of the prediction model, and the coding layer is used for adjusting the parameters of the prediction model to minimize the loss function and carrying out supervision fine tuning to obtain the trained prediction model.
With supervisionThe specific execution mode of the tuner module is as follows: the logic layer is mainly composed of a logistic regression supervised learning algorithm, the high level representation of the feature set is input into the logic layer, and the predicted amount y of bus travel on the ith line is obtained(i)Using the predicted trip amount of the bus and the input vector x on the ith bus line(i)Establishing a loss function:
where β is a prediction model parameter, D(s)Representing a set of input vectors, P representing a function for predicting accuracy, and Y representing a set of bus trip prediction quantities;
and the coding layer is used for adjusting parameters of the prediction model to minimize the loss function, and carrying out supervised fine tuning to obtain the trained prediction model.
In general, compared with the prior art, the above technical solution contemplated by the present invention can achieve the following beneficial effects:
(1) extracting sample data such as bus IC card data, bus GPS data, taxi GPS data, city POI static data, city meteorological data and the like from a city, and being beneficial to deeply analyzing the traveling condition of urban residents; by analyzing the incidence relation among the sample data, the method is beneficial to extracting the characteristic set closely related to the traffic volume of the bus on the bus line; training a prediction model based on the extracted feature set and the bus traffic, and facilitating accurate prediction of future bus traffic of the line to be predicted;
(2) preferably, the invalid value and the missing value in the sample data can be effectively processed by cleaning the sample data, the quality of the sample data is improved, and the training efficiency and the training accuracy are further improved; through the analysis of the characteristics in each affected area on the bus line, the specific information of each road section on the bus line can be obtained, and a more representative characteristic set on the bus line can be obtained;
(3) preferably, the input vector has more essential representativeness to the sample data through the high-level representation generated by the coding layer through unsupervised pre-training, and is more beneficial to the classification prediction of the logistic regression layer; the method comprises the following steps of carrying out supervised fine adjustment on a prediction model after unsupervised pre-training, so that the prediction model has better prediction performance on the traffic of buses; the coding layer can obviously enhance the expression capability of the input vector by adding noise into the input vector and reconstructing the input vector to obtain a high-level representation with higher robustness; the coding layer adopts a 4-layer denoising automatic encoder to abstract the input vector to obtain high-level representation, has stronger expression capability and is more beneficial to improving the performance of classification prediction of a logistic regression layer;
(4) preferably, the logistic regression layer minimizes a loss function built according to the bus trip prediction quantity and the input vector by finely adjusting parameters of the whole prediction model, and improves the prediction performance of the bus trip quantity of the line to be predicted by the prediction model.
Drawings
FIG. 1 is a flow chart of a method for predicting bus traffic based on city big data.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
As shown in fig. 1, a method for predicting the traffic volume of a bus based on urban big data includes:
(1) extracting sample data from the urban big data, wherein the sample data comprises: bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data;
(2) calculating the bus traffic volume on each bus line according to the bus IC card data on each bus line, performing correlation analysis on sample data, extracting a feature set related to the bus traffic volume on each bus line, and taking the bus traffic volume and the feature set as input vectors;
(3) carrying out unsupervised pre-training and supervised fine tuning on the prediction model by using the input vector to obtain a trained prediction model;
(4) the method comprises the steps of collecting bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data of a line to be predicted, extracting a feature set of the line to be predicted, inputting the feature set of the line to be predicted into a trained prediction model, and obtaining the bus traffic of the line to be predicted.
Further, the step (1) further comprises the step of cleaning redundant and incomplete data in the sample data.
Further, because a bus route spans a large area, each area has unique characteristics, such as traffic characteristics, movement characteristics of residents, geography, and weather. In order to obtain the characteristics of each section of the whole bus line, the specific implementation mode of performing correlation analysis on the sample data in the step (2) is as follows: taking an influence area by taking each bus stop on a bus line as a center, and analyzing the association relation between the mobile characteristic, the geographic characteristic and the meteorological characteristic in each influence area and the resident trip by using sample data to obtain a characteristic set.
Preferably, a random sampling method is adopted to randomly select 15 areas from all the influence areas on each bus line, and the characteristics in the 15 influence areas are respectively mined to combine the characteristic set of the whole line.
Further, the step (3) comprises the following steps:
(3-1) the prediction model comprises an encoding layer and a logistic regression layer, the feature set and the bus trip amount are input vectors of the prediction model, and the bus trip prediction amount is an output vector of the prediction model; in an encoding layer, carrying out unsupervised pre-training by using an input vector to obtain high-level representation of a feature set;
and (3-2) obtaining a bus trip prediction quantity by utilizing the high-level representation of the feature set in a logistic regression layer, establishing a loss function by utilizing the input vector, the bus trip prediction quantity and the parameters of the prediction model, and carrying out supervised fine tuning by using an encoding layer to adjust the parameters of the prediction model to minimize the loss function so as to obtain the trained prediction model.
Further, the specific implementation manner of the step (3-1) is as follows: the prediction model comprises an input layer, an encoding layer, a logistic regression layer and an output layer, the feature set and the bus trip amount are input vectors of the input layer of the prediction model, and the bus trip prediction amount is an output vector of the output layer of the prediction model; the coding layer is composed of N layers of denoising automatic encoders, a reconstruction vector and reconstruction denoising automatic coding parameters are obtained through the reconstruction denoising automatic encoders, the reconstruction error of each layer of input vector is minimized, then unsupervised training is carried out on each layer of denoising automatic encoders, high-level representation of a characteristic set is obtained, and the denoising automatic coding parameters theta and the reconstruction denoising automatic coding parameters theta' are as follows:
wherein x is(i)For the predicted amount of bus travel and the input vector, x 'on the ith bus route'(i)For the prediction quantity of bus trip and the reconstruction vector on the ith line, n represents a total of n bus lines, L represents the reconstruction error, and L (x)(i)，x′(i))＝||x′(i)-x(i)||2。
Furthermore, the coding layer is composed of 4 layers of denoising automatic coders.
Further, the specific implementation manner of the step (3-2) is as follows: the logic layer is mainly composed of a logistic regression supervised learning algorithm, the high level representation of the feature set is input into the logic layer, and the predicted amount y of bus travel on the ith line is obtained(i)Using the predicted trip amount of the bus and the input vector x on the ith bus line(i)Establishing a loss function:
where β is a prediction model parameter, D(s)Representing a set of input vectors, P representing oneA function for prediction accuracy, Y representing a set of bus trip predictors;
and the coding layer is used for adjusting parameters of the prediction model to minimize the loss function, and carrying out supervised fine tuning to obtain the trained prediction model.
Furthermore, the prediction model parameters comprise denoising automatic coding parameters, reconstruction denoising automatic coding parameters and reconstruction denoising automatic coding function bias parameters.
P(Y＝y(i)|x(i)β) may be represented as P (Y ═ Y)(i)|x(i)，W，b)，
Wherein, W represents the de-noising automatic coding parameter and the reconstruction de-noising automatic coding parameter, and b represents the offset parameter of the reconstruction de-noising automatic coding function.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (4)
1. A method for predicting the traffic volume of a bus based on urban big data is characterized by comprising the following steps:
(1) extracting sample data from the urban big data, wherein the sample data comprises: bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data;
(2) calculating the bus traffic volume on each bus line according to the bus IC card data on each bus line, performing correlation analysis on sample data, extracting a feature set related to the bus traffic volume on each bus line, and taking the bus traffic volume and the feature set as input vectors;
(3) carrying out unsupervised pre-training and supervised fine tuning on the prediction model by using the input vector to obtain a trained prediction model;
(4) collecting bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data of a line to be predicted, extracting a feature set of the line to be predicted, inputting the feature set of the line to be predicted into a trained prediction model, and obtaining the bus traffic of the line to be predicted;
the step (3) comprises the following steps:
(3-1) the prediction model comprises an input layer, an encoding layer, a logistic regression layer and an output layer, the feature set and the bus trip amount are input vectors of the input layer of the prediction model, and the bus trip prediction amount is an output vector of the output layer of the prediction model; the coding layer is composed of 4 layers of denoising automatic encoders, a reconstruction vector and reconstruction denoising automatic coding parameters are obtained through the reconstruction denoising automatic encoders, the reconstruction error of each layer of input vector is minimized, then the unsupervised training is carried out on each layer of denoising automatic encoders, the high-level representation of the characteristic set is obtained, and the denoising automatic coding parameters theta and the reconstruction denoising automatic coding parameters theta' are as follows:
wherein x is(i)For the predicted amount of bus travel and the input vector, x 'on the ith bus route'(i)For the prediction quantity of bus trip and the reconstruction vector on the ith line, n represents a total of n bus lines, L represents the reconstruction error, and L (x)(i),x'(i))＝||x'(i)-x(i)||2；
(3-2) the logic layer is mainly composed of a logistic regression supervised learning algorithm, the high level of the feature set is expressed and input into the logic layer, and the predicted bus trip amount y on the ith line is obtained(i)Using the predicted trip amount of the bus and the input vector x on the ith bus line(i)Establishing a loss function:
where β is a prediction model parameter, D(s)Representing a set of input vectors, P representing a function for predicting accuracy, and Y representing a set of bus trip prediction quantities;
and the coding layer is used for adjusting parameters of the prediction model to minimize the loss function, and carrying out supervised fine tuning to obtain the trained prediction model.
2. The method for predicting the traffic volume of the bus based on the big city data as claimed in claim 1, wherein the step (1) further comprises cleaning redundant and incomplete data in the sample data.
3. The method for predicting the traffic volume of the bus based on the urban big data according to claim 1, wherein the specific implementation manner of the correlation analysis on the sample data in the step (2) is as follows: taking an influence area by taking each bus stop on a bus line as a center, and analyzing the association relation between the mobile characteristic, the geographic characteristic and the meteorological characteristic in each influence area and the resident trip by using sample data to obtain a characteristic set.
4. A system for predicting bus traffic based on urban big data, comprising:
the sample data extracting module is used for extracting sample data from the urban big data, and the sample data comprises: bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data;
the input vector extraction module is used for calculating the bus traffic volume on the ith bus line according to the bus IC card data on the ith bus line, carrying out correlation analysis on the sample data, extracting a feature set related to the bus traffic volume on the ith bus line, and taking the bus traffic volume and the feature set as input vectors;
the training prediction model module is used for carrying out unsupervised pre-training and supervised fine tuning on the prediction model by utilizing the input vector to obtain a trained prediction model;
the prediction module is used for acquiring bus IC card data, bus GPS data, taxi GPS data, city POI static data and city meteorological data of the line to be predicted, extracting a feature set of the line to be predicted, and inputting the feature set of the line to be predicted into a trained prediction model to obtain the bus traffic of the line to be predicted;
the training prediction model module comprises:
the prediction model comprises an input layer, an encoding layer, a logistic regression layer and an output layer, the feature set and the bus trip amount are input vectors of the input layer of the prediction model, and the bus trip prediction amount is an output vector of the output layer of the prediction model; the coding layer is composed of 4 layers of denoising automatic encoders, a reconstruction vector and reconstruction denoising automatic coding parameters are obtained through the reconstruction denoising automatic encoders, the reconstruction error of each layer of input vector is minimized, then the unsupervised training is carried out on each layer of denoising automatic encoders, the high-level representation of the characteristic set is obtained, and the denoising automatic coding parameters theta and the reconstruction denoising automatic coding parameters theta' are as follows:
wherein x is(i)For the predicted amount of bus travel and the input vector, x 'on the ith bus route'(i)For the prediction quantity of bus trip and the reconstruction vector on the ith line, n represents a total of n bus lines, L represents the reconstruction error, and L (x)(i),x'(i))＝||x'(i)-x(i)||2；
The logic layer is mainly composed of a logistic regression supervised learning algorithm, the high level representation of the feature set is input into the logic layer, and the predicted amount y of bus travel on the ith line is obtained(i)Using the predicted trip amount of the bus and the input vector x on the ith bus line(i)Establishing a loss function:
where β is a prediction model parameter, D(s)Representing a set of input vectors, P representing a function for predicting accuracy, and Y representing a set of bus trip prediction quantities;
and the coding layer is used for adjusting parameters of the prediction model to minimize the loss function, and carrying out supervised fine tuning to obtain the trained prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710324534.3A CN107194505B (en) | 2017-05-10 | 2017-05-10 | Method and system for predicting bus traffic based on urban big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710324534.3A CN107194505B (en) | 2017-05-10 | 2017-05-10 | Method and system for predicting bus traffic based on urban big data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107194505A CN107194505A (en) | 2017-09-22 |
| CN107194505B true CN107194505B (en) | 2020-05-19 |
Family
ID=59872869
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710324534.3A Active CN107194505B (en) | 2017-05-10 | 2017-05-10 | Method and system for predicting bus traffic based on urban big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107194505B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107944674A (en) * | 2017-11-10 | 2018-04-20 | 华中科技大学 | A kind of method using on-line off-line data assessment commercial planning |
| CN108182196B (en) * | 2017-11-27 | 2021-09-07 | 东南大学 | Urban traffic demand prediction method based on POI |
| CN109191846B (en) * | 2018-10-12 | 2021-03-09 | 国网浙江省电力有限公司温州供电公司 | Traffic travel flow prediction method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1407513A (en) * | 2001-09-13 | 2003-04-02 | 孔勇 | Monitoring bus punctual rate system |
| CN105243444A (en) * | 2015-10-09 | 2016-01-13 | 杭州尚青科技有限公司 | City monitoring station air quality prediction method based on online multi-core regression |
| CN106295906A (en) * | 2016-08-22 | 2017-01-04 | 万马联合新能源投资有限公司 | A kind of urban public bus lines querying method based on degree of depth study |
-
2017
- 2017-05-10 CN CN201710324534.3A patent/CN107194505B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1407513A (en) * | 2001-09-13 | 2003-04-02 | 孔勇 | Monitoring bus punctual rate system |
| CN105243444A (en) * | 2015-10-09 | 2016-01-13 | 杭州尚青科技有限公司 | City monitoring station air quality prediction method based on online multi-core regression |
| CN106295906A (en) * | 2016-08-22 | 2017-01-04 | 万马联合新能源投资有限公司 | A kind of urban public bus lines querying method based on degree of depth study |
Non-Patent Citations (2)
| Title |
|---|
| 时空路网交通拥堵预测与疏导决策方法研究;任其亮;《中国博士学位论文全文数据库 工程科技Ⅱ辑》;20080515(第5期);第2-5章 * |
| 物联网环境下基于改进蚁群的大城市公交快速响应、运能优化与评价研究;潘若愚;《中国博士学位论文全文数据库 工程科技Ⅱ辑》;20150715(第 07 期);第2-5章 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107194505A (en) | 2017-09-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lv et al. | Impacts of intelligent transportation systems on energy conservation and emission reduction of transport systems: A comprehensive review | |
| Iliashenko et al. | Big data in transport modelling and planning | |
| CN107194505B (en) | Method and system for predicting bus traffic based on urban big data | |
| WO2020093701A1 (en) | Vehicle accident risk prediction model based on adaboost-so in vanets | |
| CN108182800B (en) | Freight traffic information processing method and equipment | |
| Li et al. | Public bus arrival time prediction based on traffic information management system | |
| WO2023109720A1 (en) | Urban public transportation planning method | |
| CN112434854B (en) | Intelligent environmental sanitation management method and device, computer equipment and readable medium | |
| CN112036757B (en) | Mobile phone signaling and floating car data-based parking transfer parking lot site selection method | |
| CN113326974A (en) | Multi-source traffic flow prediction method based on multitask hypergraph convolution network | |
| CN110796317A (en) | Urban taxi scheduling method based on demand prediction | |
| Li et al. | Using smart card data trimmed by train schedule to analyze metro passenger route choice with synchronous clustering | |
| Li et al. | Influence of weather conditions on the intercity travel mode choice: A Case of Xi’an | |
| CN116663742A (en) | Regional capacity prediction method based on multi-factor and model fusion | |
| CN105335758A (en) | Model identification method based on video Fisher vector descriptors | |
| Ku et al. | Real-time taxi demand prediction using recurrent neural network | |
| Roxas et al. | Estimating the environmental effects of the car shifting behavior along EDSA | |
| CN112559909B (en) | Business area discovery method based on GCN embedded spatial clustering model | |
| CN112949939B (en) | Taxi passenger carrying hotspot prediction method based on random forest model | |
| Yi et al. | Internet+ | |
| CN112990360A (en) | Online booking vehicle-mounted passenger characteristic analysis method based on K-means | |
| Han et al. | Research on Travel Time Prediction of Multiple Bus Trips Based on MDARNN | |
| Saranya | Significance of artificial intelligence in the development of sustainable transportation | |
| Li et al. | Using Random Undersampling Boosting Classifier to Estimate Mode Shift Response to Bus Local Network Expansion and Bus Rapid Transit Services | |
| CN111062589B (en) | Urban taxi scheduling method based on destination prediction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |