CN111754025A - Public transport short-time passenger flow prediction method based on CNN + GRU - Google Patents
Public transport short-time passenger flow prediction method based on CNN + GRU Download PDFInfo
- Publication number
- CN111754025A CN111754025A CN202010449431.1A CN202010449431A CN111754025A CN 111754025 A CN111754025 A CN 111754025A CN 202010449431 A CN202010449431 A CN 202010449431A CN 111754025 A CN111754025 A CN 111754025A
- Authority
- CN
- China
- Prior art keywords
- cnn
- gru
- passenger flow
- time
- bus
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 61
- 239000011159 matrix material Substances 0.000 claims abstract description 23
- 238000011176 pooling Methods 0.000 claims description 11
- 230000015654 memory Effects 0.000 claims description 10
- 210000002569 neuron Anatomy 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 238000002474 experimental method Methods 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 7
- 238000005457 optimization Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000002779 inactivation Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 3
- 210000004027 cell Anatomy 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 239000013598 vector 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
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G06Q50/40—
Abstract
The invention discloses a public transport short-time passenger flow prediction method based on CNN + GRU. The bus stop passenger flow prediction is a main basis of bus scheduling, and for the problem that only time characteristics are considered in the conventional bus short-time passenger flow prediction, a prediction model based on a Convolutional Neural Network (CNN) and a gated cycle unit (GRU) is provided, and the CNN and the GRU are respectively used for extracting the space and time characteristics of the bus passenger flow to construct a bus passenger flow prediction model. And (3) building a station passenger flow space-time matrix by using the IC card swiping data of the buses in Suzhou city and the GPS data of the buses, and predicting the passenger flow of the bus station by using a model. The experiment result shows that the CNN + GRU prediction model can effectively predict the short-time passenger flow of the public transport and has better accuracy than other models.
Description
Technical Field
The invention relates to the field of bus prediction, in particular to a short-time bus passenger flow prediction method based on CNN + GRU.
Background
The Intelligent Traffic System (ITS) is used as an important component of a smart city, can effectively relieve traffic jam, saves travel time of urban residents, improves traffic safety, and contributes to urban energy conservation and environmental protection. The intelligent bus is an important part of the intelligent traffic, needs to predict the passenger flow of a bus stop, is an important basis for bus operation decision, and is effectively applied to bus scheduling, scheduling plans and the like so as to improve the operation efficiency of the bus.
The bus passenger flow prediction means that the passenger flow at the next moment or even a plurality of later moments is predicted in real time at the current moment. The short-time passenger flow prediction mainly serves dynamic scheduling and instant information service, quick response is required, time interval division usually takes minutes as a unit, and actual operation usually takes 5-30 min.
The traditional technology has the following technical problems:
at present, there are many methods for predicting passenger flow at home and abroad, and the methods can be generally divided into two types: the first category is parametric methods including autoregressive moving average (ARIMA) models, kalman filter models, Multivariate Linear Regression (MLR) models, etc.; the second category is non-parametric methods such as support vector machines, neural networks, combinatorial models, etc. The parameter model is relatively simple, the requirement on data is low, the interpretability is strong, and the operation speed is high; however, due to the fact that the passenger flow is nonlinear and random, the characteristic properties of the parameter model cannot be effectively represented by the parameter model, and the prediction error is large. The non-parametric model is relatively complex, the algorithm convergence is slow, the optimization difficulty is high, and the over-fitting problem is easy to generate.
Disclosure of Invention
The invention aims to provide a public transport short-time passenger flow prediction method based on CNN + GRU.
In order to solve the technical problem, the invention provides a public transport short-time passenger flow prediction method based on CNN + GRU, which comprises the following steps: carrying out time and station matching pretreatment on the original public transportation data to construct a space-time matrix; inputting the matrix data into a CNN network to extract the spatial characteristics of site traffic; and inputting the data processed by the CNN into a GRU layer, and predicting the flow of the site at the next moment.
In one embodiment, the forward calculation formula of the GRU is:
zt＝σ(Wz·[ht-1,xt]) (7)
rt＝σ(Wr·[ht-1,xt]) (8)
in one embodiment, the matrix structure is as follows: the row of the matrix is the passenger flow of a certain station at the past n times, and the column of the matrix represents the flow of the adjacent m stations at a certain time;
in the ith row, Si is all traffic of the ith station at the past n times; si (t-n) is the traffic volume at the ith station, before time n.
In one embodiment, the convolutional neural network is composed of several convolutional layers (Convolution Layer), Pooling layers (Pooling Layer), and fully-connected layers.
In one embodiment, the Convolution Layer (Convolution Layer) performs Convolution operation on the input, and performs local Feature extraction to obtain a Feature matrix Feature Map. Each layer of neurons is connected with the local part of the previous layer, and the weight of the convolution kernel is shared by all the neurons.
In one embodiment, the Pooling Layer (Pooling Layer) is used for down-sampling Feature maps, and removing unimportant samples from the Feature maps to obtain a plurality of smaller Feature maps.
In one embodiment, the last is a fully connected layer; the output of the whole network is obtained by connecting with the neurons in the Feature Map, and the last full connection layer is the output layer.
Based on the same inventive concept, the present application also provides a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of any of the methods when executing the program.
Based on the same inventive concept, the present application also provides a computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of any of the methods.
Based on the same inventive concept, the present application further provides a processor for executing a program, wherein the program executes to perform any one of the methods.
The invention has the beneficial effects that:
the invention provides a deep neural network prediction model combining CNN and GRU, which utilizes CNN and GRU to extract space and time characteristics of bus passenger flow and predict bus short-time passenger flow. Probability inactivation is carried out by adopting a dropout technology in the training process, overfitting is effectively prevented, optimization of a back propagation process is carried out by adopting an adam algorithm with a self-adaptive learning rate, and the performance of the model is improved.
Drawings
FIG. 1 is an LSTM memory block structure of the public transit short-time passenger flow prediction method based on CNN + GRU of the present invention.
Fig. 2 is a GRU structure diagram of the public transit short-time passenger flow prediction method based on CNN + GRU of the present invention.
FIG. 3 is a structural diagram of a CNN plus GRU model of the public transportation short-term passenger flow prediction method based on CNN + GRU of the present invention.
FIG. 4 is a comparison graph of experimental prediction results in the public transportation short-term passenger flow prediction method based on CNN + GRU of the present invention.
Detailed Description
The present invention is further described below in conjunction with the following figures and specific examples so that those skilled in the art may better understand the present invention and practice it, but the examples are not intended to limit the present invention.
The bus stop passenger flow prediction is a main basis of bus scheduling, and for the problem that only time characteristics are considered in the conventional bus short-time passenger flow prediction, a prediction model based on a Convolutional Neural Network (CNN) and a gated cycle unit (GRU) is provided, and the CNN and the GRU are respectively used for extracting the space and time characteristics of the bus passenger flow to construct a bus passenger flow prediction model. And (3) building a station passenger flow space-time matrix by using the IC card swiping data of the buses in Suzhou city and the GPS data of the buses, and predicting the passenger flow of the bus station by using a model. The experiment result shows that the CNN + GRU prediction model can effectively predict the short-time passenger flow of the public transport and has better accuracy than other models.
Deep learning is a characterization learning method with a multilayer structure, and is successfully applied in a plurality of fields.
Most of the existing deep learning prediction methods are based on the time sequence data of the site, the time characteristic of the passenger flow is considered, and the space characteristic of the passenger flow is not fully mined. The invention provides a CNN + GRU bus short-time passenger flow prediction model, which is applied to short-time passenger flow prediction of bus stops.
CNN + GRU prediction model
CNN
Convolutional Neural Networks (CNN) is a multi-layer perceptual deep learning network, and is mainly used for processing data with similar network structures, such as time series data and image data, and obtaining significant results in the fields of image classification, natural language processing, and the like. The CNN excavates spatial correlation characteristics through a local connection mode, so the method uses the CNN to excavate the spatial characteristics of the public transport passenger flow.
A convolutional neural network is composed of a plurality of convolutional layers (Convolution Layer), pooling layers (PoolingLayer) and full-connection layers. And a Convolution Layer (Convolution Layer) for performing Convolution operation on the input and performing local Feature extraction to obtain a Feature matrix Feature Map. Each layer of neurons is connected with the local part of the previous layer, and the weight of the convolution kernel is shared by all the neurons. The pool Layer (Pooling Layer) is mainly used for downsampling the Feature Map, and removing unimportant samples in the Feature Map to obtain a plurality of smaller Feature maps. And finally, a full connection layer. The output of the whole network is obtained by connecting with the neurons in the Feature Map, and the last full connection layer is the output layer.
GRU
LSTM
The Long Short Term Memory Network (LSTM) is an improved Recurrent Neural Network (RNN). While original RNNs are difficult to train in real applications to deeper-ranked networks, LSTM networks have "memory" and are good at handling multiple variable problems, LSTM can perform better in longer sequences than normal RNNs.
The basic unit of the LSTM hidden layer is a memory block structure, which is composed of an input gate, a forgetting gate, and an output gate, and the memory block structure is shown in fig. 1.
A forgetting gate (forget gate) for controlling the information reserved to the current previous moment; an input gate (input gate) controls the current input; the output gate (output gate) is used to control the current output value of the LSTM. The updating of the history information is achieved by the control of the three gates.
The method for processing a memory block at time t can be divided into the following steps
The LSTM unit update procedure is as follows:
ft＝σ(Wf·[ht-1,xt]+bf) (1)
it＝σ(Wi·[ht-1,xt]+bi) (2)
ot＝σ(W0·[ht-1,xt]+b0) (5)
f represents a forgetting gate; i denotes an input gate; o is an output gate; c represents a cell state;
is the current transmissionAn incoming cell state; h is the final output; symbol
Multiplying elements representing corresponding positions of the two vectors; "σ" denotes sigmoid activation function;
GRU
gated Recycling Unit (GRU) is one of the most successful variants of LSTM. It simplifies the LSTM even more, but maintains the same effect of the LSTM. The GRU makes two major changes to the LSTM: one is to use two gates: update Gate (Update Gate) ztAnd a Reset Gate (Reset Gate) rtThe original three gates of the forgetting gate, the input gate and the output gate are replaced. The second is to replace the cell state and output with a state h.
The forward calculation formula of the GRU is:
zt＝σ(Wz·[ht-1,xt]) (7)
rt＝σ(Wr·[ht-1,xt]) (8)
the GRU structure is shown in fig. 2.
CNN+GRU
The traffic volume of a site is not only related to the traffic of the site at the historical time, but also affected by the traffic of other sites.
The Convolutional Neural Network (CNN) has the characteristics of local perception and weight sharing, so that the invention uses the CNN to extract spatial features related to positions of passenger flow data. A spatio-temporal matrix containing position and time information is constructed. The row of the matrix is the passenger flow of a certain station at the past n times, and the column of the matrix represents the flow of the adjacent m stations at a certain time. The matrix structure is as follows:
in the ith row, Si is all traffic of the ith station at the past n times; si (t-n) is the traffic volume at the ith station, before time n.
And combining the CNN model and the GRU model, and mining the time and space characteristics of the passenger flow of the bus stop according to the spatial perception characteristic of the CNN and the memory characteristic of the GRU to realize short-time flow prediction. The short-time passenger flow prediction process of the CNN + GRU bus is as follows:
the model structure is shown in fig. 3:
a specific application scenario of the present invention is given below:
data pre-processing
The data of the invention is from a bus system from 5 months to 7 months in 2017 of New district of Suzhou city, the passenger flow statistics is carried out on the stops of a plurality of lines through IC card swiping data and bus GPS data, and 32 buses, the garden stops of the riverside and the adjacent stops thereof are selected for passenger flow prediction.
The IC card data structure is shown in table 1, and the GPS data structure is shown in table 2.
TABLE 1 IC card field description
| Name of field | Type (B) | Description of the invention |
| CARD_ID | char(20) | Consumption card number |
| XF_DATE | DATETIME | Consumption date |
| XF_TIME | DATETIME | Consumption time |
| LINE_NO | char(10) | Line |
| BUS_NO | char(20) | Vehicle number |
| XF_JE | DECIMAL(10,2) | Amount of consumption |
| POS_NO | char(16) | POS machine number |
| CARD_TYPE | char(10) | Data source |
TABLE 2 bus GPS data field description
Because the card swiping data does not contain the station information, the GPS data is independent of the card swiping data, and the passenger card swiping station cannot be directly obtained. In view of this situation, the present invention first solves the passenger boarding station identification problem: according to the information of the circuit, the license plate number, the card swiping time and the like in the IC card swiping data, the information is matched with the license plate number, the driving direction and the time in the GPS data, and the passenger card swiping station is determined according to the GPS station.
After the station value is supplemented to the card swiping data by the method, the following data preprocessing is required to be carried out:
firstly, dividing the bus running time of 6:00-21:00 every day by one time period of every 30 minutes;
secondly, carrying out time clustering on the IC card swiping data of 2 months, and dividing all the card swiping data into corresponding time periods;
and then, performing data statistical analysis, namely counting the number of card swipes of each time period of a certain station to obtain the passenger flow volume of each time period, wherein passenger flow data samples obtained after counting are shown in a table 3.
And finally, constructing a space-time matrix according to the station and time interval passenger flow, and preparing data for flow prediction.
TABLE 3 passenger flow data samples
| Date | Site | Time period | Passenger flow volume |
| 2017-07-16 | 11 | 06:00:00 | 8 |
| 2017-07-16 | 11 | 06:30:00 | 12 |
| 2017-07-16 | 11 | 07:00:00 | 10 |
| 2017-07-16 | 11 | 08:00:00 | 16 |
| 2017-07-16 | 11 | 08:30:00 | 21 |
| 2017-07-16 | 11 | 09:00:00 | 17 |
Model construction
Taking a 32-way bus riverside garden site in Suzhou new district as an example, the CNN + GRU model is used for prediction.
The number of CNN convolution layers in the CNN + GRU prediction model is 1, the number of convolution kernels is 32, the size of the convolution kernels is 3x3, and the number of CNN pooling layers is 1; the GRU comprises two layers, the number of nodes in each layer is 100, the time step is 5, and the traffic of a station at the next moment is predicted according to the passenger traffic of 5 previous moments of the station.
In order to prevent overfitting, a dropout technology is adopted in the training process to carry out random probability inactivation, and the random inactivation probability is set to be 0.2.
Since the learning rate has a significant influence on the performance of the model, and too large or too small of the learning rate causes oscillation of the network and cannot converge to an optimal solution, an optimization method that can update the learning rate by using the general information in the gradient descent is required to improve the model training speed. The invention adopts the adaptive learning rate Adam algorithm to carry out the optimization of the back propagation process, the batch processing size is 128, the iteration times are 300, the relu activation function is used, and the learning rate is set to be 0.001.
Evaluation index
In order to measure the performance of the prediction model, the method adopts the average absolute error (MAPE) and the Root Mean Square Error (RMSE) as performance evaluation indexes to carry out model evaluation. The calculation formulas of MAPE and RMSE are shown in formulas (11) and (12):
wherein N represents the number of samples, xiRepresenting the actual passenger traffic at a site at a time, yiThe delegate model predicts passenger flow. MAPE reflects the total deviation of passenger flow predicted values, and the prediction accuracy of the model can be measured; the RMSE reflects the error between the predicted value and the true value, which can measure the accuracy of the predicted value.
Analysis of Experimental results
The experiment predicts the traffic of the site at the next moment according to the traffic of the riverside garden site and the past 5 moments of the previous 4 sites. In order to verify the effectiveness of the CNN + GRU model, the invention uses the LSTM model, the CNN model and the GRU model to carry out prediction comparison on the site passenger flow, and the prediction results of the four models are shown in FIG. 4. These models use the same data set.
The experimental results are shown in table 4, and it can be seen from the results that the CNN + GRU model and other 3 deep learning models can effectively predict the short-sight passenger flow of the bus, while the CNN + GRU model of the present invention comprehensively considers the time and space characteristics for prediction, and has higher accuracy.
TABLE 4 comparison of the results
The invention provides a deep neural network prediction model combining CNN and GRU, which utilizes CNN and GRU to extract space and time characteristics of bus passenger flow and predict bus short-time passenger flow. Probability inactivation is carried out by adopting a dropout technology in the training process, overfitting is effectively prevented, optimization of a back propagation process is carried out by adopting an adam algorithm with a self-adaptive learning rate, and the performance of the model is improved. The method comprises the steps of using IC card swiping data and bus GPS data of the Suzhou public transportation system to construct a space-time matrix as input, using a model to predict station passenger flow, and comparing the station passenger flow with other prediction models, wherein the result shows that the CNN + GRU prediction model can effectively predict the bus passenger flow in a short time, can be applied to the urban intelligent public transportation system, and has higher practical value.
The above-mentioned embodiments are merely preferred embodiments for fully illustrating the present invention, and the scope of the present invention is not limited thereto. The equivalent substitution or change made by the technical personnel in the technical field on the basis of the invention is all within the protection scope of the invention. The protection scope of the invention is subject to the claims.
Claims (10)
1. A public transport short-time passenger flow prediction method based on CNN + GRU is characterized by comprising the following steps: and carrying out time and station matching pretreatment on the original bus data to construct a space-time matrix. Inputting the matrix data into a CNN network to extract the spatial characteristics of site traffic; and inputting the data processed by the CNN into a GRU layer, and predicting the flow of the site at the next moment.
2. The CNN + GRU-based bus short-time passenger flow prediction method as claimed in claim 1, wherein the forward calculation formula of the GRU is as follows:
zt＝σ(Wz·[ht-1,xt]) (7)
rt＝σ(Wr·[ht-1,xt]) (8)
3. the CNN + GRU-based bus short-time passenger flow prediction method as claimed in claim 1, wherein the matrix structure is as follows: the row of the matrix is the passenger flow of a certain station at the past n times, and the column of the matrix represents the flow of the adjacent m stations at a certain time;
in the ith row, Si is all traffic of the ith station at the past n times; si (t-n) is the traffic volume at the ith station, before time n.
4. The CNN + GRU-based bus short-time passenger flow prediction method of claim 1, wherein the convolutional neural network is composed of a plurality of convolutional layers (convolutional layers), Pooling layers (Pooling layers), and fully-connected layers.
5. The CNN + GRU-based short-time bus traffic prediction method as claimed in claim 4, wherein the Convolution Layer (Convolution Layer) performs Convolution operation on the input to perform local feature extraction, thereby obtaining a feature matrix FeatureMap. Each layer of neurons is connected with the local part of the previous layer, and the weight of the convolution kernel is shared by all the neurons.
6. The CNN + GRU-based short transit time passenger flow prediction method of claim 4, wherein the Pooling Layer (Pooling Layer) is mainly used for downsampling the Feature Map, and removing unimportant samples in the Feature Map to obtain a plurality of smaller Feature maps.
7. The CNN + GRU-based bus short-time passenger flow prediction method of claim 4, wherein a full connection layer is last; the output of the whole network is obtained by connecting with the neurons in the Feature Map, and the last full connection layer is the output layer.
8. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1 to 7 are implemented when the program is executed by the processor.
9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
10. A processor, characterized in that the processor is configured to run a program, wherein the program when running performs the method of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010449431.1A CN111754025A (en) | 2020-05-25 | 2020-05-25 | Public transport short-time passenger flow prediction method based on CNN + GRU |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010449431.1A CN111754025A (en) | 2020-05-25 | 2020-05-25 | Public transport short-time passenger flow prediction method based on CNN + GRU |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111754025A true CN111754025A (en) | 2020-10-09 |
Family
ID=72673985
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010449431.1A Pending CN111754025A (en) | 2020-05-25 | 2020-05-25 | Public transport short-time passenger flow prediction method based on CNN + GRU |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111754025A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112418547A (en) * | 2020-12-03 | 2021-02-26 | 北京工业大学 | Bus stop passenger flow prediction method based on GCN-LSTM combined model |
| CN112862187A (en) * | 2021-02-05 | 2021-05-28 | 北京百度网讯科技有限公司 | Bus passenger flow prediction method and device, electronic equipment and storage medium |
| CN112905659A (en) * | 2021-02-05 | 2021-06-04 | 希盟泰克(重庆)实业发展有限公司 | Urban rail transit data analysis method based on BIM and artificial intelligence |
| WO2021189950A1 (en) * | 2020-10-29 | 2021-09-30 | 平安科技（深圳）有限公司 | Short-time bus station passenger flow prediction method and apparatus, and computer device and storage medium |
| CN113537569A (en) * | 2021-06-18 | 2021-10-22 | 华南理工大学 | Short-term bus passenger flow prediction method and system based on weight stack decision tree |
| CN113688770A (en) * | 2021-09-02 | 2021-11-23 | 重庆大学 | Long-term wind pressure missing data completion method and device for high-rise building |
| CN114141020A (en) * | 2022-02-08 | 2022-03-04 | 北京城建设计发展集团股份有限公司 | Pedestrian simulation method and system based on rail transit and electronic equipment |
| CN115100907A (en) * | 2022-06-17 | 2022-09-23 | 南京航空航天大学 | Terminal area airspace flight flow prediction method for meteorological scene classification |
-
2020
- 2020-05-25 CN CN202010449431.1A patent/CN111754025A/en active Pending
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021189950A1 (en) * | 2020-10-29 | 2021-09-30 | 平安科技（深圳）有限公司 | Short-time bus station passenger flow prediction method and apparatus, and computer device and storage medium |
| CN112418547A (en) * | 2020-12-03 | 2021-02-26 | 北京工业大学 | Bus stop passenger flow prediction method based on GCN-LSTM combined model |
| CN112418547B (en) * | 2020-12-03 | 2024-04-09 | 北京工业大学 | Bus stop passenger flow prediction method based on GCN-LSTM combination model |
| CN112862187A (en) * | 2021-02-05 | 2021-05-28 | 北京百度网讯科技有限公司 | Bus passenger flow prediction method and device, electronic equipment and storage medium |
| CN112905659A (en) * | 2021-02-05 | 2021-06-04 | 希盟泰克(重庆)实业发展有限公司 | Urban rail transit data analysis method based on BIM and artificial intelligence |
| CN113537569A (en) * | 2021-06-18 | 2021-10-22 | 华南理工大学 | Short-term bus passenger flow prediction method and system based on weight stack decision tree |
| CN113537569B (en) * | 2021-06-18 | 2023-11-28 | 华南理工大学 | Short-term bus passenger flow prediction method and system based on weight stacking decision tree |
| CN113688770A (en) * | 2021-09-02 | 2021-11-23 | 重庆大学 | Long-term wind pressure missing data completion method and device for high-rise building |
| CN113688770B (en) * | 2021-09-02 | 2024-02-06 | 重庆大学 | Method and device for supplementing long-term wind pressure missing data of high-rise building |
| CN114141020A (en) * | 2022-02-08 | 2022-03-04 | 北京城建设计发展集团股份有限公司 | Pedestrian simulation method and system based on rail transit and electronic equipment |
| CN114141020B (en) * | 2022-02-08 | 2022-09-16 | 北京城建设计发展集团股份有限公司 | Pedestrian simulation method and system based on rail transit and electronic equipment |
| CN115100907A (en) * | 2022-06-17 | 2022-09-23 | 南京航空航天大学 | Terminal area airspace flight flow prediction method for meteorological scene classification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111754025A (en) | Public transport short-time passenger flow prediction method based on CNN + GRU | |
| Zhao et al. | Deep temporal convolutional networks for short-term traffic flow forecasting | |
| CN109919358B (en) | Real-time station flow prediction method based on neural network space-time attention mechanism | |
| Singh et al. | Short term electricity price forecast based on environmentally adapted generalized neuron | |
| CN104662526B (en) | Apparatus and method for efficiently updating spiking neuron network | |
| Dai et al. | Deeptrend: A deep hierarchical neural network for traffic flow prediction | |
| CN111899510A (en) | Intelligent traffic system flow short-term prediction method and system based on divergent convolution and GAT | |
| CN108280551A (en) | A kind of photovoltaic power generation power prediction method using shot and long term memory network | |
| CN109886444A (en) | A kind of traffic passenger flow forecasting, device, equipment and storage medium in short-term | |
| CN110942637B (en) | SCATS system road traffic flow prediction method based on airspace map convolutional neural network | |
| He et al. | Research on a novel combination system on the basis of deep learning and swarm intelligence optimization algorithm for wind speed forecasting | |
| Zhang et al. | A Traffic Prediction Method of Bicycle-sharing based on Long and Short term Memory Network. | |
| CN111160622A (en) | Scenic spot passenger flow prediction method and device based on hybrid neural network model | |
| Lu et al. | A hybrid model based on convolutional neural network and long short-term memory for short-term load forecasting | |
| Li et al. | A Kriging-based bi-objective constrained optimization method for fuel economy of hydrogen fuel cell vehicle | |
| Modi et al. | Multistep traffic speed prediction: A deep learning based approach using latent space mapping considering spatio-temporal dependencies | |
| Costa | Convolutional-LSTM networks and generalization in forecasting of household photovoltaic generation | |
| CN110210656A (en) | Shared bicycle method for predicting and system based on website behavioural analysis | |
| Lv et al. | An improved long short-term memory neural network for stock forecast | |
| CN112766603A (en) | Traffic flow prediction method, system, computer device and storage medium | |
| Brahimi et al. | Modelling on car-sharing serial prediction based on machine learning and deep learning | |
| Zou et al. | Short-term traffic flow prediction based on PCC-BiLSTM | |
| Mi et al. | A dynamic ensemble deep deterministic policy gradient recursive network for spatiotemporal traffic speed forecasting in an urban road network | |
| CN116227180A (en) | Data-driven-based intelligent decision-making method for unit combination | |
| Papageorgiou et al. | Application of fuzzy cognitive maps to electricity consumption 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 |