CN108985475B - Network taxi appointment and taxi calling demand prediction method based on deep neural network - Google Patents
Network taxi appointment and taxi calling demand prediction method based on deep neural network Download PDFInfo
- Publication number
- CN108985475B CN108985475B CN201810609232.5A CN201810609232A CN108985475B CN 108985475 B CN108985475 B CN 108985475B CN 201810609232 A CN201810609232 A CN 201810609232A CN 108985475 B CN108985475 B CN 108985475B
- Authority
- CN
- China
- Prior art keywords
- taxi
- calling
- demand
- network
- calling demand
- 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
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
- G06Q10/02—Reservations, e.g. for tickets, services or events
-
- 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
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a method for predicting taxi booking and calling requirements based on a deep neural network, which comprises the following steps: s1, carrying out regional division on the urban road network, and calculating and counting the taxi appointment and taxi calling demand of each region; s2, designing a network car booking and calling demand prediction model: learning and training the taxi calling demand time-space characteristics based on a deep neural network, and predicting the regional taxi calling demand in the next time period by combining with the influence factors of weather influence factors on taxi calling willingness of taxi calling users; and S3, training the regional car calling demand prediction model according to the historical data, and predicting the net appointment car calling demand quantity of each region by using the trained model. According to the method, by fusing external factors and the space-time correlation, the complex nonlinear space-time correlation characteristics of the net appointment car calling demand are expressed by utilizing the deep neural network, and higher prediction precision can be obtained.
Description
Technical Field
The invention relates to the field of cross technology application of deep learning and taxi calling demand prediction, in particular to a network taxi appointment taxi calling demand prediction method based on a deep neural network.
Background
With the rapid development of location-based services and mobile internet technologies, network taxi appointment is gradually becoming an important alternative for urban travel. However, the taxi calling requirements and supply capacity of the taxi appointment have the problem of mismatching in the space-time dimension. For example, the urban center has a phenomenon of 'difficult driving' during off-duty rush hours, and the suburban area has a phenomenon of excess transportation capacity. In addition, external factors such as weather influence the travel desire of people, and for example, the demand for calling a car in rainy days is obviously increased. Therefore, how to utilize mass cross-domain data to predict taxi calling requirements and provide big data decision support for a network taxi appointment platform becomes a hotspot problem in the field of traffic big data application research gradually.
Earlier prediction models have relied primarily on techniques such as time series prediction, for example: autoregressive model (AR), historical average model (HA), autoregressive moving average model (ARIMA), and the like. The drawback of these techniques is the difficulty in expressing the complex non-linear spatiotemporal correlation between the summons requirements and external factors. With the excellent performance of the deep neural network in learning the correlation of the complex features of the mass data, people are motivated to apply a deep learning method to predict the demand, but the current research method considers single factors, such as: only the Convolutional Neural Network (CNN) is used for considering spatial factors, and only the long-short term memory network (LSTM) is used for considering temporal factors, so that the prediction accuracy needs to be improved.
Disclosure of Invention
The invention aims to provide a method for predicting vehicle calling demand of a network appointment based on a deep neural network. Based on massive network car booking order data, the method can accurately predict the car calling demand of the network car booking in the city and guide the dynamic scheduling of the network car booking, and can improve the satisfaction degree of users and effectively relieve urban traffic jam.
The method for predicting the taxi calling demand of the online taxi appointment based on the deep neural network comprises the following steps:
s1, carrying out grid division on the urban roads to form grid areas, and calculating and counting the taxi calling demand in each area according to the taxi appointment order data;
s2, designing a network car booking and calling demand prediction model: learning and training the space-time characteristics of the taxi calling demands based on a deep neural network, and predicting taxi calling demands of each region in the next time period by combining the influence of weather factors on taxi calling willingness of taxi booking users;
and S3, training the regional car calling demand prediction model according to the historical data, and predicting the net appointment car calling demand quantity of each region by using the trained model.
Further, step S1 specifically includes:
s11, dividing the urban road network into M multiplied by N grid areas according to the longitude and latitude, so that the area (M, N) represents the grid area of the mth row and the nth column;
s12, extracting networked taxi appointment order data o (o.t, o.l and o.u), wherein o.t represents taxi calling time stamps, o.l represents taxi calling places, and o.u represents taxi calling order ids, and the taxi calling order data o is used for filtering repeated and invalid orders;
s13, mapping the network taxi appointment order data into the grid areas according to the longitude and latitude, and calculating the order quantity recorded in each area in each time period as the taxi calling demand:
wherein the content of the first and second substances,
representing the taxi-calling demand of the area (m, n) in the time period t;
s14, normalizing the taxi calling demand in each area:
wherein y is0Demand for calling, yminAnd ymaxRespectively representing the minimum value and the maximum value of the car calling demand, wherein y is a normalized value;
s15, representing the normalized taxi calling demand quantity in the time period t obtained by the method in the step S14 as a 3-dimensional matrix of M rows, N columns and 1 channel, and recording the matrix as yt∈SM*N*1The historical h time periods of the taxi calling demand yhistory＝{yt1,2, h, and the vehicle-calling demand y in the next time periodt+1As output, construct the sample { yhistroy,yt+1}；
And S16, constructing samples of all order data according to the method in the step S15, and dividing the samples into a training set and a testing set according to a certain proportion.
Further, step S2 specifically includes:
s21, extracting each time interval by using a residual error network
The result is denoted as y't∈SM*N*1；
S22, fusing spatial features of h time periodsThe result is denoted as y ″)t∈SM*N*h；
S23, inputting the result obtained in the S22 into a convolution duration memory network to extract space-time characteristics;
s24, converting the weather condition into vector representation of a certain dimensionality by using One-Hot coding, and normalizing other numerical weather data;
s25, inputting the results obtained in S23 and S24 into a two-layer full-connection network to obtain a predicted value
Further, in step S21, the number of layers of the residual network is 6, the convolutional neural network CNN uses zero padding, the size of the convolutional kernel is set to 3 × 3 × 3, the CNN convolutional kernel is set to 64, and the output layer uses 1 convolutional kernel to extract spatial features.
Further, in step S23, the number of layers of the convolution long and short term memory network is 2, the size of the convolution kernel is set to 5 × 5 using zero padding, 32 convolution kernels are used, and 1 convolution kernel is used for the output layer.
Further, step S3 specifically includes:
s31, training the car calling demand prediction model in the step S2 by utilizing a training set, wherein the loss function is
Wherein y ist+1The actual amount of demand is represented,
representing a predicted demand amount;
s32, selecting a model with the minimum Loss function Loss as an optimal model according to the test set;
and S33, normalizing the vehicle calling demand data, inputting the normalized vehicle calling demand data into the optimal model obtained in S32, obtaining an output result, performing inverse normalization to obtain a final prediction result, and realizing prediction of vehicle calling demand in each area of the city.
After adopting the technical scheme, compared with the background technology, the invention has the following advantages:
according to the invention, a residual error network is adopted for extracting spatial features, the spatial-temporal features are extracted through a convolution long-term memory network, and the influence of weather factors on the taxi calling willingness of taxi calling users is fully considered, so that the taxi calling demand of taxi calling in each area of a city can be accurately predicted, the dynamic dispatching of taxi calling in each area can be guided, the satisfaction degree of users can be improved, and the urban traffic jam can be effectively relieved.
Drawings
FIG. 1 is a flow chart of a network appointment car-calling demand prediction method with a deep neural network fusing external factors and space-time correlations;
FIG. 2 is a schematic diagram of a model for predicting a demand for a taxi call;
FIG. 3 is a schematic diagram of model training and taxi calling demand prediction;
FIG. 4 is a schematic diagram of urban road network grid area division;
FIG. 5 is a sample diagram of the data of the network car booking order in 2016 and 11 months in metropolis;
FIG. 6 is a sample graph of weather data for 2016 and 11 months in metropolis;
fig. 7 is a comparison of the results of various algorithms.
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.
Example 1
The method for predicting the taxi calling demand of the online taxi appointment based on the deep neural network comprises the following steps:
s1, carrying out grid division on the urban roads to form grid areas, and calculating and counting the taxi calling demand in each area according to the taxi appointment order data;
s2, designing a network car booking and calling demand prediction model: and (3) learning and training the space-time characteristics of the taxi calling demands based on a deep neural network, and predicting the taxi calling demands of all areas in the next time period by combining the influence of weather factors on taxi calling willingness of taxi booking users.
And S3, training the regional car calling demand prediction model according to the historical data, and predicting the net appointment car calling demand quantity of each region by using the trained model.
Wherein, step S1 specifically includes:
s11, the city road network is divided into M × N grid areas according to the longitude and latitude, so that the area (M, N) represents the grid area of the M-th row and the N-th column.
And S12, extracting networked taxi appointment order data o (o.t, o.l and o.u), wherein o.t represents a taxi calling timestamp, o.l represents a taxi calling place, and o.u represents a taxi calling order id, and the data are used for filtering repeated and invalid orders.
S13, mapping the network taxi appointment order data into the grid areas according to the longitude and latitude, and calculating the order quantity recorded in each area in each time period as the taxi calling demand:
wherein the content of the first and second substances,
indicating the summoning demand for the area (m, n) over the time period t.
S14, normalizing the taxi calling demand in each area:
wherein y is0Demand for calling, yminAnd ymaxRespectively represents the minimum value and the maximum value of the taxi calling demand, and y is a normalized value.
S15, representing the normalized taxi calling demand quantity in the time period t obtained by the method in the step S14 as a 3-dimensional matrix of M rows, N columns and 1 channel, and recording the matrix as yt∈SM*N*1The historical h time periods of the taxi calling demand yhistory＝{yt1,2, h, and the vehicle-calling demand y in the next time periodt+1As output, construct the sample { yhistroy,yt+1}。
And S16, constructing samples of all order data according to the method in the step S15, and dividing the samples into a training set and a testing set according to a certain proportion.
Wherein, step S2 specifically includes:
s21, extracting each time interval by using a residual error network
The result is denoted as y't∈SM*N*1。
Because the urban-level demand has spatial correlation, namely the current regional taxi calling demand and the surrounding environment are mutually influenced, and the problem of gradient disappearance in the feature extraction process is caused by only overlapping the number of CNN network layers, the method adopts a residual error network to extract the spatial features.
In step S21, the number of layers of the residual error network is 6, the convolutional neural network CNN uses zero padding, the size of the convolutional kernel is set to 3 × 3 × 3, the CNN convolutional kernel is set to 64, and the output layer uses 1 convolutional kernel to extract spatial features.
S22, fusing the spatial characteristics of the h time periods, and marking the result as y ″t∈SM*N*h；
Because the demand has a time series relationship and forms space-time correlation with the space characteristics, the space characteristic results of h time periods are combined, and therefore the space-time characteristics are extracted by utilizing a convolution long-time memory network in the next step.
And S23, inputting the result obtained in the S22 into a convolution long-short term memory network (ConvLSTM) to extract space-time characteristics.
In the invention, the number of layers of the convolution long-time memory network is 2, zero padding operation is used, the size of a convolution kernel is set to be 5 multiplied by 5, 32 convolution kernels are used, and 1 convolution kernel is used in an output layer.
S24, converting weather conditions (sunny, cloudy and rainy) into vector representation with certain dimensionality by using One-Hot coding, and normalizing other numerical weather data (temperature, visibility and wind speed);
s25, inputting the results obtained in S23 and S24 into a two-layer full-connection network to obtain a predicted value
Finally, a schematic diagram of the network appointment car calling demand prediction model is shown in fig. 2.
As shown in fig. 3, step S3 specifically includes:
s31, training the car calling demand prediction model in the step S2 by utilizing a training set, wherein the loss function is
Wherein, yt+1The actual amount of demand is represented,
representing a predicted demand amount;
s32, selecting a model with the minimum Loss function Loss as an optimal model according to the test set;
and S33, normalizing the vehicle calling demand data, inputting the normalized vehicle calling demand data into the optimal model obtained in S32, obtaining an output result, performing inverse normalization to obtain a final prediction result, and realizing prediction of vehicle calling demand in each area of the city.
Example 2
The method disclosed by the invention is used for predicting the car calling demand.
1. First, mesh division is performed, and as shown in fig. 4, a schematic diagram of city road network mesh region division in a metropolis is shown.
2. And acquiring the order data of the online taxi appointment.
Precision sample order data from 2016, 11 months, 1 days to 11 months, 30 days, 2-4s in a metropolitan area opened to academia by a network appointment platform company [ data source: https:// gaia. dichuxing. com ], 7,065,937 pieces of data. The order data includes five fields (all string types): order ID, start billing time, end billing time, boarding location longitude, boarding location latitude, alighting location longitude, and boarding location latitude. The order ID is used as unique identification information, the time information field is a unix timestamp, and the unit is second. The latitude and longitude information field is a GCJ-02 coordinate system. The data sample is shown in FIG. 5.
3. And acquiring weather data.
In WEATHER UNDERGROUND [ data source: https:// www.wunderground.com ] and the national weather website [ data sources: https:// rp5.ru ] collects 720 pieces of data for collating the weather data for one hour from 2016 at the time interval of 11 months 1 day to 11 months 30 days. Weather data includes five main fields: time period information, air temperature, visibility, wind speed, weather conditions. Wherein, the fields of air temperature, visibility and wind speed are continuous data (int type), and the time section information and the weather condition are classified data (string type). The data sample is shown in FIG. 6.
In summary, the data sets used in the experiment are 7,065,937 pieces of network car booking and car calling order data and 720 pieces of weather data of 30 days in 2016, 11 and month. The first 23 days were taken as training set and the last 7 days as test set.
The final experimental results are shown in fig. 7, where FAST _1 and FAST _2 are the prediction methods disclosed in the present invention, and the difference is that FAST _1 is not incorporated into the weather factor, and FAST _2 is incorporated into the weather factor. It can be seen that adding weather, an external factor, improves the model accuracy, and FAST _2 improves the accuracy by 1.4% over FAST _ 1.
In fig. 7, ARIMA and ANN algorithms are classic algorithms for machine learning, and CNN and LSTM are classic algorithms for deep learning spatial and temporal feature extraction, respectively. As can be seen from the experimental results, the algorithm of the invention is superior to the comparison algorithm, wherein (FAST _2) is improved by 75.65% compared with ARIMA algorithm, 17.50% compared with ANN algorithm, 1.89% compared with CNN algorithm, and 11.70% compared with LSTM algorithm.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (5)
1. The method for predicting the taxi calling demand of the online taxi appointment based on the deep neural network is characterized by comprising the following steps of:
s1, carrying out grid division on the urban roads to form grid areas, and calculating and counting the taxi calling demand in each area according to the taxi appointment order data;
s2, designing a network car booking and calling demand prediction model: learning and training the space-time characteristics of the taxi calling demands based on a deep neural network, and predicting taxi calling demands of each region in the next time period by combining the influence of weather factors on taxi calling willingness of taxi booking users; step S2 specifically includes:
s21, extracting each time interval by using a residual error network
The result is denoted as y't∈SM *N*1；
S22, fusing the spatial characteristics of the h time periods, and marking the result as y ″t∈SM*N*h；
S23, inputting the result obtained in the S22 into a convolution duration memory network to extract space-time characteristics;
s24, converting the weather condition into vector representation of a certain dimensionality by using One-Hot coding, and normalizing other numerical weather data;
s25, inputting the results obtained in S23 and S24 into a two-layer full-connection network to obtain a predicted value
And S3, training the regional car calling demand prediction model according to the historical data, and predicting the net appointment car calling demand quantity of each region by using the trained model.
2. The method for predicting vehicle calling demand for car booking based on deep neural network as claimed in claim 1, wherein step S1 specifically comprises:
s11, dividing the urban road network into M multiplied by N grid areas according to the longitude and latitude, so that the area (M, N) represents the grid area of the mth row and the nth column;
s12, extracting networked taxi appointment order data o (o.t, o.l and o.u), wherein o.t represents taxi calling time stamps, o.l represents taxi calling places, and o.u represents taxi calling order ids, and the taxi calling order data o is used for filtering repeated and invalid orders;
s13, mapping the network taxi appointment order data into the grid areas according to the longitude and latitude, and calculating the order quantity recorded in each area in each time period as the taxi calling demand:
wherein the content of the first and second substances,
representing the taxi-calling demand of the area (m, n) in the time period t;
s14, normalizing the taxi calling demand in each area:
wherein y is0Demand for calling, yminAnd ymaxRespectively representing the minimum value and the maximum value of the car calling demand, wherein y is a normalized value;
s15, representing the normalized taxi calling demand quantity in the time period t obtained by the method in the step S14 as a 3-dimensional matrix of M rows, N columns and 1 channel, and recording the matrix as yt∈SM*N*1The historical h time periods of the taxi calling demand yhistory＝{yt1,2, h, and the vehicle-calling demand y in the next time periodt+1As output, construct the sample { yhistroy,yt+1}；
And S16, constructing samples of all order data according to the method in the step S15, and dividing the samples into a training set and a testing set according to a certain proportion.
3. The deep neural network-based online taxi appointment and taxi calling demand prediction method according to claim 1, characterized in that: in step S21, the number of layers of the residual error network is 6, the convolutional neural network CNN uses zero padding, the size of the convolutional kernel is set to 3 × 3 × 3, the CNN convolutional kernel is set to 64, and the output layer uses 1 convolutional kernel to extract spatial features.
4. The deep neural network-based online taxi appointment and taxi calling demand prediction method according to claim 1, characterized in that: in step S23, the number of layers of the convolution duration memory network is 2, zero padding is used, the size of the convolution kernel is set to 5 × 5, 32 convolution kernels are used, and 1 convolution kernel is used in the output layer.
5. The deep neural network-based online taxi appointment and taxi calling demand prediction method according to claim 1, characterized in that: step S3 specifically includes:
s31, training the car calling demand prediction model in the step S2 by utilizing a training set, wherein the loss function is
Wherein, yt+1The actual amount of demand is represented,
representing a predicted demand amount;
s32, selecting a model with the minimum Loss function Loss as an optimal model according to the test set;
and S33, normalizing the vehicle calling demand data, inputting the normalized vehicle calling demand data into the optimal model obtained in S32, obtaining an output result, performing inverse normalization to obtain a final prediction result, and realizing prediction of vehicle calling demand in each area of the city.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810609232.5A CN108985475B (en) | 2018-06-13 | 2018-06-13 | Network taxi appointment and taxi calling demand prediction method based on deep neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810609232.5A CN108985475B (en) | 2018-06-13 | 2018-06-13 | Network taxi appointment and taxi calling demand prediction method based on deep neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108985475A CN108985475A (en) | 2018-12-11 |
| CN108985475B true CN108985475B (en) | 2021-07-23 |
Family
ID=64540335
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810609232.5A Active CN108985475B (en) | 2018-06-13 | 2018-06-13 | Network taxi appointment and taxi calling demand prediction method based on deep neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108985475B (en) |
Families Citing this family (39)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110046787A (en) * | 2019-01-15 | 2019-07-23 | 重庆邮电大学 | A kind of urban area charging demand for electric vehicles spatio-temporal prediction method |
| CN109829649A (en) * | 2019-01-31 | 2019-05-31 | 北京首汽智行科技有限公司 | A kind of vehicle dispatching method |
| CN111612122B (en) * | 2019-02-25 | 2023-08-08 | 北京嘀嘀无限科技发展有限公司 | Real-time demand prediction method and device and electronic equipment |
| CN111612183A (en) * | 2019-02-25 | 2020-09-01 | 北京嘀嘀无限科技发展有限公司 | Information processing method, information processing device, electronic equipment and computer readable storage medium |
| CN111784018A (en) * | 2019-04-03 | 2020-10-16 | 北京嘀嘀无限科技发展有限公司 | Resource scheduling method and device, electronic equipment and storage medium |
| CN110246329A (en) * | 2019-04-07 | 2019-09-17 | 武汉理工大学 | A kind of taxi quantitative forecasting technique |
| CN110119838A (en) * | 2019-04-17 | 2019-08-13 | 成都信息工程大学 | A kind of shared bicycle demand forecast system, method and device |
| CN110264248A (en) * | 2019-05-30 | 2019-09-20 | 阿里巴巴集团控股有限公司 | The prediction technique and device of user experience information |
| CN110415013A (en) * | 2019-06-12 | 2019-11-05 | 河海大学 | A kind of combination forecasting method of net about vehicle trip requirements in short-term |
| CN110322064B (en) * | 2019-06-27 | 2022-04-15 | 清华大学 | Urban trip demand prediction method |
| CN110458336B (en) * | 2019-07-23 | 2020-06-09 | 内蒙古工业大学 | Online appointment vehicle supply and demand prediction method based on deep learning |
| CN110599767A (en) * | 2019-09-04 | 2019-12-20 | 广东工业大学 | Long-term and short-term prediction method based on network taxi appointment travel demands |
| CN111062589B (en) * | 2019-12-02 | 2022-08-16 | 武汉理工大学 | Urban taxi scheduling method based on destination prediction |
| CN111160747B (en) * | 2019-12-23 | 2022-07-22 | 北京百度网讯科技有限公司 | Scheduling method and device for unmanned robot taxi and electronic equipment |
| CN111340543B (en) * | 2020-02-25 | 2024-03-22 | 上海大学 | Fine granularity taxi taking demand prediction method based on attention mechanism |
| CN110991775B (en) * | 2020-03-02 | 2020-06-26 | 北京全路通信信号研究设计院集团有限公司 | Deep learning-based rail transit passenger flow demand prediction method and device |
| CN111507762B (en) * | 2020-04-15 | 2023-10-31 | 中国科学院自动化研究所 | Urban taxi demand prediction method based on multitasking co-prediction neural network |
| CN111242394B (en) * | 2020-04-26 | 2021-03-23 | 北京全路通信信号研究设计院集团有限公司 | Method and system for extracting spatial correlation characteristics |
| CN111507541B (en) * | 2020-04-30 | 2021-01-29 | 南京福佑在线电子商务有限公司 | Goods quantity prediction model construction method, goods quantity measurement device and electronic equipment |
| CN111862579B (en) * | 2020-06-10 | 2021-07-13 | 深圳大学 | Taxi scheduling method and system based on deep reinforcement learning |
| CN111882171A (en) * | 2020-07-06 | 2020-11-03 | 湖北合纵科创能源科技有限公司 | Intelligent scheduling system and method for scenic spot sightseeing vehicle based on demand prediction |
| CN111915081B (en) * | 2020-08-03 | 2023-10-17 | 东北大学秦皇岛分校 | Peak sensitive travel demand prediction method based on deep learning |
| CN112150207A (en) * | 2020-09-30 | 2020-12-29 | 武汉大学 | Online taxi appointment order demand prediction method based on space-time context attention network |
| CN112291807B (en) * | 2020-10-15 | 2022-09-23 | 山东科技大学 | Wireless cellular network traffic prediction method based on deep migration learning and cross-domain data fusion |
| CN112330215B (en) * | 2020-11-26 | 2024-02-02 | 长沙理工大学 | Urban vehicle demand prediction method, equipment and storage medium |
| CN112488574B (en) * | 2020-12-16 | 2022-11-18 | 东南大学 | Travel demand prediction method based on space-time feature extraction |
| CN112561187B (en) * | 2020-12-22 | 2022-06-03 | 武汉理工大学 | Network taxi booking target order prediction method based on CNN-LSTM |
| CN112734101B (en) * | 2021-01-04 | 2022-06-03 | 北京瀚文智远信息技术有限公司 | Sharing bicycle intelligent allocation method based on vehicle demand prediction |
| CN112861925B (en) * | 2021-01-18 | 2023-04-07 | 中国科学院自动化研究所 | Deep learning network-based multi-region vehicle demand prediction method and system |
| CN113034195B (en) * | 2021-04-07 | 2023-05-02 | 西北工业大学 | Network vehicle-approaching demand prediction method based on convolution network and non-local network |
| CN113222287A (en) * | 2021-06-01 | 2021-08-06 | 平安科技（深圳）有限公司 | Network appointment demand prediction method, device, equipment and storage medium |
| CN113672846A (en) * | 2021-06-18 | 2021-11-19 | 中国科学院自动化研究所 | Network appointment scheduling method and device, electronic equipment and storage medium |
| CN113538067B (en) * | 2021-08-17 | 2023-05-05 | 华侨大学 | Inter-city network vehicle-closing demand prediction method and system based on machine learning |
| CN113570004B (en) * | 2021-09-24 | 2022-01-07 | 西南交通大学 | Riding hot spot area prediction method, device, equipment and readable storage medium |
| CN114418606B (en) * | 2021-12-01 | 2024-05-28 | 武汉大学 | Network vehicle order demand prediction method based on space-time convolution network |
| CN114386880B (en) * | 2022-03-22 | 2022-07-08 | 北京骑胜科技有限公司 | Model training method for determining order quantity, order quantity determining method and device |
| CN114819414B (en) * | 2022-06-24 | 2022-10-28 | 北京阿帕科蓝科技有限公司 | Block demand prediction method, system and computer storage medium |
| CN116029407A (en) * | 2022-07-29 | 2023-04-28 | 大连海事大学 | Taxi travel demand prediction method based on ConvLSTM |
| CN116468469A (en) * | 2023-04-24 | 2023-07-21 | 北京化工大学 | Network taxi demand prediction method based on improved EMD-LSTM |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10558750B2 (en) * | 2016-11-18 | 2020-02-11 | Salesforce.Com, Inc. | Spatial attention model for image captioning |
| CN107103758B (en) * | 2017-06-08 | 2019-06-21 | 厦门大学 | A kind of city area-traffic method for predicting based on deep learning |
| CN107844853A (en) * | 2017-10-23 | 2018-03-27 | 重庆大学 | It is a kind of that the commending system by bus for reducing net about fare is predicted using dynamic price |
| CN107967532B (en) * | 2017-10-30 | 2020-07-07 | 厦门大学 | Urban traffic flow prediction method fusing regional vitality |
| CN108053080B (en) * | 2017-12-30 | 2021-05-11 | 中国移动通信集团江苏有限公司 | Method, device, equipment and medium for predicting regional user number statistics |
-
2018
- 2018-06-13 CN CN201810609232.5A patent/CN108985475B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108985475A (en) | 2018-12-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108985475B (en) | Network taxi appointment and taxi calling demand prediction method based on deep neural network | |
| WO2021212866A1 (en) | Vehicle travel volume prediction model construction method, and prediction method and system | |
| Yi et al. | Electric vehicle charging demand forecasting using deep learning model | |
| CN109658695B (en) | Multi-factor short-term traffic flow prediction method | |
| Chen et al. | Fine-grained prediction of urban population using mobile phone location data | |
| Kim et al. | Graph convolutional network approach applied to predict hourly bike-sharing demands considering spatial, temporal, and global effects | |
| CN112037506B (en) | Vehicle track prediction model construction method, vehicle track prediction method and system | |
| CN108960476B (en) | AP-TI clustering-based shared bicycle flow prediction method and device | |
| CN110766280A (en) | Vehicle scheduling method and generation method and device of target order prediction model | |
| CN112561187B (en) | Network taxi booking target order prediction method based on CNN-LSTM | |
| CN110210664B (en) | Deep learning method for short-term prediction of using behaviors of multiple individual vehicles | |
| CN110956807A (en) | Highway flow prediction method based on combination of multi-source data and sliding window | |
| CN110889092A (en) | Short-time large-scale activity peripheral track station passenger flow volume prediction method based on track transaction data | |
| CN112579922A (en) | Interest point recommendation method based on user track | |
| Xiong et al. | An integrated and personalized traveler information and incentive scheme for energy efficient mobility systems | |
| CN115412857A (en) | Resident travel information prediction method | |
| Ma et al. | Investigating the impact of spatial-temporal grid size on the microscopic forecasting of the inflow and outflow gap in a free-floating bike-sharing system | |
| CN115034429A (en) | Urban taxi pick-up point recommendation method based on depth map neural network | |
| Zhou et al. | An attention-based deep learning model for citywide traffic flow forecasting | |
| CN112801340B (en) | Crowd density prediction method based on multi-level city information unit portraits | |
| Huang et al. | Central station-based demand prediction for determining target inventory in a bike-sharing system | |
| Nicoletta et al. | Bayesian spatio-temporal modelling and prediction of areal demands for ambulance services | |
| Pramanik et al. | Modeling traffic congestion in developing countries using google maps data | |
| Lin et al. | Predictions of taxi demand based on neural network algorithms | |
| CN113344290B (en) | Method for correcting sub-season rainfall weather forecast based on U-Net network |
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 |