CN112242060A - Traffic flow prediction method and apparatus, computer device, and readable storage medium - Google Patents
Traffic flow prediction method and apparatus, computer device, and readable storage medium Download PDFInfo
- Publication number
- CN112242060A CN112242060A CN202011107026.8A CN202011107026A CN112242060A CN 112242060 A CN112242060 A CN 112242060A CN 202011107026 A CN202011107026 A CN 202011107026A CN 112242060 A CN112242060 A CN 112242060A
- Authority
- CN
- China
- Prior art keywords
- traffic
- neural network
- data
- network model
- traffic data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
Abstract
The application discloses a traffic flow prediction device, a computer device and a readable storage medium. The method comprises the following steps: acquiring original traffic data of a target road section within a preset time; processing the original traffic data to obtain an influence parameter of an external factor on the traffic flow of the target road section; processing the original traffic data according to the influence parameters to obtain corrected traffic data; dividing the corrected traffic data to obtain training set data and test set data; training an initial causal convolution-cyclic neural network model using training set data to obtain an intermediate causal convolution-cyclic neural network model; testing the intermediate causal convolution-cyclic neural network model by using the test set data to obtain an evaluation result, and confirming the intermediate causal convolution-cyclic neural network model as a target causal convolution-cyclic neural network model when the evaluation result meets a preset condition; and predicting the traffic flow of the target road section at the time point to be predicted by using the target causal convolution-cyclic neural network model.
Description
Technical Field
The present application relates to the field of traffic flow prediction technologies, and in particular, to a traffic flow prediction method, a traffic flow prediction apparatus, a computer device, and a non-volatile computer-readable storage medium.
Background
Traffic plays a crucial role in everyone's daily life, and everyone needs to spend a lot of time on traffic going out every day, in which case accurate real-time traffic situation prediction is very important for road users, private departments and governments. In addition, the traffic services widely used by people at present, such as flow control, route planning, navigation and the like, also depend on high-quality traffic condition assessment seriously, so that the accurate prediction of the traffic state is significant, and the purpose of traffic prediction is to predict the future traffic state of the connected road sections according to historical traffic data in the basic road network structure. However, the conventional traffic flow prediction method has the problem of low prediction accuracy.
Disclosure of Invention
The embodiment of the application provides a traffic flow prediction method, a traffic flow prediction device, computer equipment and a nonvolatile computer readable storage medium, so as to solve the problem that the existing traffic flow prediction method is low in prediction precision.
The traffic flow prediction method according to the embodiment of the present application includes: acquiring original traffic data of a target road section within a preset time, wherein the original traffic data comprises traffic flow data of the target road section and external factor information influencing the traffic flow of the target road section; processing the original traffic data to obtain an influence parameter of an external factor on the traffic flow of the target road section; processing the original traffic data according to the influence parameters to obtain corrected traffic data from which the influence of the external factors is removed; dividing the corrected traffic data to obtain training set data and test set data; training an initial causal convolution-recurrent neural network model using the training set data to obtain an intermediate causal convolution-recurrent neural network model; testing the intermediate causal convolution-cyclic neural network model by using the test set data to obtain an evaluation result, and confirming the intermediate causal convolution-cyclic neural network model as a target causal convolution-cyclic neural network model when the evaluation result meets a preset condition; and predicting the traffic flow of the target road section at the time point to be predicted by utilizing the target causal convolution-cyclic neural network model.
In certain embodiments, the traffic flow data includes at least one of traffic speed and traffic density; the external factor information includes information of at least one external factor of low temperature, high temperature, normal, light rain, heavy rain, light snow, heavy wind, traffic condition and road repair.
In some embodiments, the processing the raw traffic data to obtain an influence parameter of an external factor on the traffic flow of the target road segment includes: processing the original traffic data to obtain normal traffic data which are not influenced by the external factors and influenced traffic data which are influenced by the external factor information in the original traffic data; and calculating the influence parameter according to the influenced traffic data and the normal traffic data.
In some embodiments, each of the external factors corresponds to one of the influence parameters, and the calculating the influence parameter according to the influenced traffic data and the normal traffic data includes: for each time node within the predetermined time length, calculating a mean value of the normal traffic data belonging to the same time node to obtain a plurality of mean values corresponding to a plurality of the time nodes; for each external factor, calculating initial parameters of the external factor under each time node according to the influenced traffic data influenced by the external factor and the average values; and calculating the influence parameters according to the plurality of initial parameters corresponding to the plurality of time nodes.
In some embodiments, the training of the initial causal convolutional-cyclic neural network model with the training set data to obtain an intermediate causal convolutional-cyclic neural network model comprises: inputting the training set data into the causal convolution unit in units of a predetermined period of time to obtain a causal convolution output result; inputting the causal convolution output result into the recurrent neural network element to obtain a predicted value; calculating a loss value according to the real value and the predicted value; and confirming that the trained initial causal convolution-cyclic neural network model is the intermediate initial causal convolution-cyclic convolution neural network model when the loss value is smaller than a preset threshold value; continuing to train the trained initial causal convolutional-cyclic neural network model when the loss value is greater than the predetermined threshold.
In some embodiments, the causal convolution unit comprises a plurality of causal convolution layers, and the cyclic neural network unit is a cyclic gate unit.
In some embodiments, the predicting, with the target causal convolution-cyclic neural network model, traffic flow of the target road segment at the time point to be predicted includes: acquiring original traffic data in a preset period, wherein the preset period comprises a time point to be predicted; processing the original traffic data in the preset period according to the influence parameters to obtain corrected traffic data in the preset period; inputting the corrected traffic data in the preset period into the target causal convolution-circulation neural network model to obtain an initial predicted value of the time point to be predicted; and calculating a target predicted value according to the initial predicted value and the influence parameter.
The traffic flow prediction device comprises an acquisition module, a first processing module, a second processing module, a dividing module, a training module, a testing module and a prediction module. The acquisition module is used for acquiring original traffic data of a target road section within a preset time period, wherein the original traffic data comprises traffic flow data of the target road section and external factor information influencing the traffic flow of the target road section. The first processing module is used for processing the original traffic data to obtain an influence parameter of an external factor on the traffic flow of the target road section. And the second processing module is used for processing the original traffic data according to the influence parameters to obtain corrected traffic data with the influence of the external factors removed. The dividing module is used for dividing the corrected traffic data to obtain training set data and test set data. The training module is used for training an initial causal convolution-cyclic neural network model by using the training set data to obtain an intermediate causal convolution-cyclic neural network model. The test module is used for testing the intermediate causal convolution-cyclic neural network model by using the test set data to obtain an evaluation result, and confirming that the intermediate causal convolution-cyclic neural network model is a target causal convolution-cyclic neural network model when the evaluation result meets a preset condition. The prediction module is used for predicting the traffic flow of the target road section at the time point to be predicted by utilizing the target causal convolution-circulation neural network model.
The computer device of the embodiments of the present application includes a processor, a memory, and one or more programs. The one or more programs are stored in the memory, and the one or more programs are executable by the processor to implement the traffic flow prediction method according to any one of the above embodiments.
The nonvolatile computer-readable storage medium of the embodiments of the present application contains a computer program. The computer program is executed by a processor to implement the traffic flow prediction method according to any one of the above embodiments.
The traffic flow prediction method, the traffic flow prediction device, the electronic device, and the nonvolatile computer-readable storage medium according to the embodiments of the present application take into account the influence of external factors on the traffic flow, and can make the traffic flow prediction result more accurate. In addition, the causal convolution-circulation neural network model is adopted to predict the traffic flow, and the causal convolution-circulation neural network model can sense a time sequence with a larger time width on the premise of reducing the calculated amount, so that the model can predict the traffic flow of a time point to be predicted depending on historical traffic data of a large number of time nodes, and the model is favorable for further improving the accuracy of traffic flow prediction and improving the real-time property of traffic flow prediction.
Additional aspects and advantages of embodiments of the present application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the present application.
Drawings
The above and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a schematic flow diagram of a traffic flow prediction method according to certain embodiments of the present application;
FIG. 2 is a schematic view of a traffic flow prediction device according to certain embodiments of the present application;
FIG. 3 is a schematic flow diagram of a traffic flow prediction method according to certain embodiments of the present application;
FIG. 4 is a schematic flow diagram of a traffic flow prediction method according to certain embodiments of the present application;
FIG. 5 is a schematic diagram of a first processing module of certain embodiments of the present application;
FIG. 6 is a flow diagram illustrating a traffic flow prediction method according to certain embodiments of the present application;
FIG. 7 is a schematic diagram of a training module of certain embodiments of the present application;
FIG. 8 is a schematic diagram of a causal convolution unit of certain embodiments of the present application;
FIG. 9 is a schematic illustration of a traffic flow prediction method according to certain embodiments of the present application;
FIG. 10 is a schematic diagram of a cyclic gating cell in accordance with certain embodiments of the present application;
FIG. 11 is a flow diagram illustrating a traffic flow prediction method according to certain embodiments of the present application;
FIG. 12 is a schematic diagram of a prediction module of certain embodiments of the present application;
FIG. 13 is a schematic diagram of a computer device of certain embodiments of the present application;
FIG. 14 is a schematic diagram of the interaction of a non-volatile computer readable storage medium and a processor of certain embodiments of the present application.
Detailed Description
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below by referring to the drawings are exemplary only for the purpose of explaining the embodiments of the present application, and are not to be construed as limiting the embodiments of the present application.
Referring to fig. 1, the present application discloses a traffic flow prediction method. The traffic flow prediction method comprises the following steps:
01: acquiring original traffic data of a target road section within a preset time, wherein the original traffic data comprises traffic flow data of the target road section and external factor information influencing the traffic flow of the target road section;
02: processing the original traffic data to obtain an influence parameter of an external factor on the traffic flow of the target road section;
03: processing the raw traffic data to obtain corrected traffic data from which the influence of the external factors is removed;
04: dividing the corrected traffic data to obtain training set data and test set data;
05: training an initial causal convolution-cyclic neural network model using training set data to obtain an intermediate causal convolution-cyclic neural network model;
06: testing the intermediate causal convolution-cyclic neural network model by using the test set data to obtain an evaluation result, and confirming the intermediate causal convolution-cyclic neural network model as a target causal convolution-cyclic neural network model when the evaluation result meets a preset condition; and
07: and predicting the traffic flow of the target road section at the time point to be predicted by using the target causal convolution-cyclic neural network model.
Referring to fig. 2, the present application further discloses a traffic flow prediction apparatus 10. The traffic flow prediction method according to the embodiment of the present application can be realized by the traffic flow prediction apparatus 10 according to the embodiment of the present application. The traffic flow prediction device 10 comprises an acquisition module 11, a first processing module 12, a second processing module 13, a dividing module 14, a training module 15 and a testing module 16-level prediction module 17. Wherein step 01 can be implemented by the obtaining module 11. Step 02 may be implemented by the first processing module 12. Step 03 may be implemented by the second processing module 13. Step 04 may be implemented by the partitioning module 14. Step 05 may be implemented by training module 15. Step 06 may be implemented by test module 16. Step 07 may be implemented by prediction module 17.
That is, the obtaining module 11 may be configured to obtain original traffic data of the target road segment within a predetermined time period, where the original traffic data includes traffic flow data of the target road segment and external factor information affecting traffic flow of the target road segment. The first processing module 12 may be configured to process the raw traffic data to obtain parameters of influence of external factors on the traffic flow of the target road segment. The second processing module 13 may be used to process the raw traffic data to obtain modified traffic data from which the influence of external factors is removed. The partitioning module 14 may be used to partition the modified traffic data to obtain training set data and test set data. The training module 15 may be configured to train the initial causal convolution-cyclic neural network model with training set data to obtain an intermediate causal convolution-cyclic neural network model. The test module 16 may be configured to test the intermediate causal convolutional-recurrent neural network model using the test set data to obtain an evaluation result, and to identify the intermediate causal convolutional-recurrent neural network model as a target causal convolutional-recurrent neural network model when the evaluation result satisfies a predetermined condition. The prediction module 17 may be configured to predict the traffic flow of the target road segment at the time point to be predicted by using the target causal convolution-cyclic neural network model.
Wherein, the traffic flow data in the original traffic data comprises at least one of traffic speed and traffic density. That is, the traffic flow data may include only traffic speed or traffic density, or the traffic flow data may include both traffic speed and traffic density. In particular embodiments of the present application, traffic flow data includes both traffic speed and traffic density. The traffic speed and the traffic density may be acquired by a sensor provided in the target section.
The external factor information in the original traffic data comprises information of at least one external factor in low temperature, high temperature, normal, light rain, heavy rain, light snow, heavy wind, traffic conditions and road repair. Illustratively, the external factor information may simultaneously include information of six external factors of high temperature, normal, light rain, heavy rain, and traffic conditions, or the external factor information may simultaneously include information of twelve external factors of low temperature, high temperature, normal, light rain, heavy rain, light snow, heavy wind, traffic conditions, road repair, and the like. The criteria for evaluating high temperature, light rain, heavy rain, light snow, heavy wind, etc. may be determined by referring to the relevant division criteria of weather departments for weather phenomena such as temperature, rainfall, snow mass, wind force, etc.
The predetermined period of time may be one week, half month, one month, three months, six months, nine months, one year, two years, three years, five years, etc., without limitation. The raw traffic data may be collected at intervals of a predetermined duration, for example, at ten minute intervals. Of course, in other embodiments, the interval time may be one minute, five minutes, fifteen minutes, twenty minutes, thirty minutes, etc., without limitation. Here, assuming that the predetermined time period is one week, and the original traffic data is collected every ten minutes, the original traffic data of the target link in the predetermined time period includes data collected 6 × 24 × 7 times in step 01.
In one embodiment of the present application, the traffic flow data is represented as a vector Xd,tWhere d denotes the day, t denotes the time of day and t is spaced apart by ten minutes. External factor Pd,tHigh temperature, normal, light rain, heavy rain, traffic conditions],Pd,tIt can be represented by a 0, 1 vector, where 0 indicates no external factor and 1 indicates the presence of the external factor. Illustratively, Pd,t(0, 1, 0, 0, 0, 0) indicates that the traffic flow data is normal traffic flow data not affected by external factors; pd,tThe term (0, 0, 0, 0, 1) indicates that the traffic flow data is affected traffic flow data affected by an external factor of "traffic condition".
After the original traffic data is obtained, the original traffic data can be further processed, the processed data is used for model training and testing to obtain a prediction model of the target, and finally, a traffic flow prediction result can be obtained by using the prediction model of the target.
It is understood that, in the related art, the traffic flow prediction may be implemented using a statistical analysis method, a nonlinear theory method, a deep learning method, and the like. However, these methods have problems such as low accuracy of the traffic flow prediction result or too large calculation amount.
The traffic flow prediction method and the traffic flow prediction apparatus 10 according to the embodiment of the present application can make the prediction result of the traffic flow more accurate by considering the influence of the external factors on the traffic flow. In addition, the causal convolution-circulation neural network model is adopted to predict the traffic flow, and the causal convolution-circulation neural network model can sense a time sequence with a larger time width on the premise of reducing the calculated amount, so that the model can predict the traffic flow of a time point to be predicted depending on historical data of a large number of time nodes, and the model is favorable for further improving the accuracy of traffic flow prediction and improving the real-time property of traffic flow prediction.
Referring to fig. 3 and 4, in some embodiments, step 02 processes raw traffic data to obtain parameters of influence of external factors on traffic flow of a target road segment, including:
021: processing the original traffic data to obtain normal traffic data which are not influenced by external factors and influenced traffic data which are influenced by the external factor information in the original traffic data; and
022: and calculating the influence parameters according to the influenced traffic data and the normal traffic data.
Wherein, every external factor corresponds an influence parameter, and step 022 calculates the influence parameter according to the influenced traffic data and the normal traffic data, including:
0221: calculating a mean value of normal traffic data belonging to the same time node for each time node within a predetermined length of time to obtain a plurality of mean values corresponding to a plurality of time nodes; and
0222: for each external factor, calculating an initial parameter of the external factor under each time node according to the influenced traffic data influenced by the external factor and a plurality of average values; and
0223: and calculating influence parameters according to a plurality of initial parameters corresponding to the time nodes.
Referring to fig. 5, in some embodiments, the first processing module 12 includes a first processing unit 121 and a first calculating unit 122. Step 021 may be implemented by the first processing unit 121, and step 022, step 0221, step 0222, and step 0223 may all be implemented by the first computing unit 122.
That is, the first processing unit 121 may be configured to process the raw traffic data to obtain normal traffic data that is not affected by the external factor and affected traffic data that is affected by the external factor information in the raw traffic data. The first calculation unit 122 may be configured to calculate the impact parameter based on the affected traffic data and the normal traffic data. When the first calculating unit 122 is configured to calculate the influence parameter according to the influenced traffic data and the normal traffic data, the first calculating unit 122 is specifically configured to: calculating a mean value of normal traffic data belonging to the same time node for each time node within a predetermined length of time to obtain a plurality of mean values corresponding to a plurality of time nodes; for each external factor, calculating an initial parameter of the external factor under each time node according to the influenced traffic data influenced by the external factor and a plurality of average values; and calculating influence parameters according to a plurality of initial parameters corresponding to the time nodes.
Specifically, the first processing unit 121 may filter the raw traffic data acquired in step 01, and the first processing unit 121 may filter the raw traffic data according to Pd,tTo determine which data is normal traffic data and which data is affected traffic data. By external factors Pd,tHigh temperature, normal, light rain, heavy rain, traffic conditions]The influence parameter ω of the external factor on the traffic flow of the target link is [ ω ═ ω1,ω2,ω3,ω4,ω5,ω6]For example, the first processing unit 121 extracts all P in the original traffic datad,t=(0,1, 0, 0, 0, 0) is screened out as normal traffic data, and P is divided from the original traffic datad,tAll data other than the data of (0, 1, 0, 0, 0, 0) are merged into affected traffic data.
Subsequently, the first computing unit 122 can classify the affected traffic data as Xd,t,i(wherein i is 1, 2, 3, 4, 5, 6, Pd,tNot (0, 1, 0, 0, 0, 0)) affected traffic data on different days (i.e., d is different) but with the same time node (i.e., t is the same) and affected by the same external factor (i is the same) are merged into the same category of data; similarly, the first calculating unit 122 can classify the normal traffic data to classify Xd,t,2(wherein, Pd,tNormal traffic data on different days (i.e., different d) but the same time node (i.e., same t)) in (0, 1, 0, 0, 0, 0)) are merged into the same category of data.
Then, for each time node, the first calculating subunit may calculate, according to the normal traffic data at the time node, an average value of the normal traffic data at the time node(where 2 denotes i-2, which corresponds to the case where the first term of the sequence is denoted by 1; if the first term of the sequence is denoted by 0, i-1 here), for example, assuming that the predetermined period of time is 7 days, for the time node where t is 10 (i.e., the zero-point ten of each day), the first calculation unit 122 may calculate the average value of the normal traffic data at t-10, that is, the average value of the normal traffic data at t-10, from the normal traffic data at 7 t-10 of the 7 days, that is, from the normal traffic data at 7 t-10 Similarly, for the time node of t-60 (i.e., a whole point of each day), the first calculating unit 122 may calculate the average value of the normal traffic data at t-10 according to the normal traffic data at 7 t-60 in the 7 days, that is, the average value of the normal traffic data at t-10 And so on.
Subsequently, for each external factor, the first calculating unit 122 may calculate the initial parameter ω of each external factor at each time node according to the affected traffic data under the same time node and affected by the same external factor and the plurality of average valuest,iThat is to sayWherein, d is 1, 2, 3. For example, for the external factor "traffic condition", the first calculating unit 122 may calculate the affected traffic data X affected by the external factor "traffic condition" at the time node when t is 10 (i.e., zero-point ten times per day)d,10,6And t is the mean value of normal traffic data at 10To calculate an initial parameter omega of the influence of the external factor of 'traffic condition' on the traffic flow at the time node of t-1010,6Similarly, the first calculation unit 122 may calculate the affected traffic data X affected by the external factor of "traffic condition" at the time node of t 60 (i.e., at one time each day)d,60,6And t is the mean value of normal traffic data at 60To calculate an initial parameter omega of the influence of the external factor of 'traffic condition' on the traffic flow at the time node of t-6060,6And so on.
Subsequently, the first calculation unit 122 may calculate an influence parameter according to a plurality of initial parameters corresponding to a plurality of time nodes, for example, an influence parameter of an external factor of "traffic condition" on the traffic flowWhere n represents the number of time nodes t. Therefore, the influence parameters of each external factor on the traffic flow can be calculated. The influence parameters of external factors on the traffic flow can be uniformly recorded
Referring to fig. 1, in some embodiments, after obtaining the influence parameter, the second processing module 13 processes the original traffic data within the predetermined time period according to the influence parameter to obtain the corrected traffic data within the predetermined time period. In particular, correcting traffic dataThus, by performing the external dependency removal processing on the original traffic data, the part of the data weakened by the external factors can be supplemented to obtain the corrected traffic data from which the influence of the external factors is removed.
Referring to fig. 1, in some embodiments, the dividing module 14 may correct the traffic data into the training set data and the test set data by using a K-fold intersection algorithm, a leave-out method, a leave-one-out method, a self-help method, and the like, which is not limited herein. In a specific embodiment of the present application, the dividing module 14 divides the corrected traffic data by using a K-fold intersection algorithm, where a value of K is 10, and certainly, the value of K may also be other values, which is not limited herein. Specifically, the partitioning module 14 partitions the entire modified traffic data into K disjoint subsets S1,S2,…,SKAnd if the number of all samples in the corrected traffic data is M, M/K samples exist in each subset. And taking one sample from the divided subsets as a test set every time, and taking other K-1 samples as a training set, so that training set data and test set data can be obtained. The K-fold intersection operation method is adopted to divide and correct traffic data, all data in the sample can be well utilized, and subsequent model training and testing are facilitated.
Referring to FIG. 6, in some embodiments, the initial causal convolutional-cyclic neural network model includes a causal convolutional unit 151 and a cyclic neural network unit 152, and step 05 trains the initial causal convolutional-cyclic neural network model using training set data to obtain an intermediate causal convolutional-cyclic neural network model, including:
051: inputting training set data into the causal convolution unit 151 in units of a predetermined period to obtain a causal convolution output result;
052: inputting the causal convolution output result into the recurrent neural network unit 152 to obtain a predicted value;
053: calculating a loss value according to the real value and the predicted value; and
054: confirming that the trained initial causal convolution-circulation neural network model is an intermediate initial causal convolution-circulation convolution neural network model when the loss value is smaller than a preset threshold value;
055: and continuing to train the trained initial causal convolution-circulation neural network model when the loss value is larger than the preset threshold value.
Referring to fig. 7, in some embodiments, the training module 15 includes a causal convolution unit 151, a recurrent neural network unit 152, a second calculation unit 153, and a confirmation unit 154. Step 051 may be implemented by the causal convolution unit 151. Step 052 may be implemented by recurrent neural network unit 152. Step 053 may be implemented by the second calculation unit 153. Step 054 may be implemented by the validation unit 154. Step 055 may be implemented by causal convolution unit 151 and recurrent neural network unit 152. That is, the causal convolution unit 151 may be configured to calculate training set data input in units of a predetermined period of time to obtain a causal convolution output result. The recurrent neural network element 152 may be used to calculate the input causal convolution output result to obtain a predicted value. The second calculation unit 153 may be configured to calculate the loss value based on the real value and the predicted value. The validation unit 154 may be configured to validate the trained initial causal convolutional-cyclic neural network model as an intermediate initial causal convolutional-cyclic convolutional neural network model when the loss value is less than a predetermined threshold. The causal convolution unit 151 and the recurrent neural network unit 152 may be configured to continue training the trained initial causal convolution-recurrent neural network model when the loss value is greater than a predetermined threshold.
The predetermined time period may be 6 hours, 12 hours, 18 hours, 24 hours, 36 hours, 72 hours, etc., and is not limited herein.
For example, please refer to fig. 7 to 10, training set data is input to the causal convolution unit 151 in units of days (24 hours) for extracting the dependence of the traffic flow of each time point to be predicted on the traffic flow of a plurality of time points associated therewith. The causal convolution unit 151 may include one or more causal convolution layers, which is not limited herein.
Illustratively, when the causal convolution unit 151 includes only one causal convolution layer, the corrected traffic data Y obtained in step 04d,tWhen the data is input into the cause and effect convolution layer, the processing formula for correcting the traffic data is as follows:
wherein, wd,t,pIs the convolution result; q is the expansion ratio; phi is ak,m,pAll of the elements of (a) represent a convolution kernel,Kris the nucleus length; m is a parameter set for the expandability of the causal convolution unit 151, and in one embodiment of the present application, M is 1, but in other examples, M may also be a numerical value such as 2, 3, 4, 5, etc., which is not limited herein; p is the number of convolution kernel channels. The above formula can be written as: w ═ Φ -qY。
Illustratively, when the causal convolution unit 151 includes multiple layers of causal convolution layers, as shown in fig. 8, each causal convolution layer includes four layers, and 15 time nodes may be sensed, where, under the second layer, the time point to be predicted may sense the modified traffic data at the first 3 time instants, after one layer (i.e., the third layer) is superimposed, the time point to be predicted may sense the modified traffic data at the first 7 time instants, and after one layer (i.e., the fourth layer) is further superimposed, the time point to be predicted may sense the data at the first 15 time instants. With the superposition of the layer number, the time which can be sensed at the time point to be predicted rises exponentially. At this time, the processing formula for correcting the traffic data is as follows:
whereinqThe causal convolution operation with an expansion rate q is shown, f is the nonlinear activation function, and 1 is the number of layers.
The four layers shown in fig. 8 are merely examples, and in other examples, the number of layers of the causal convolution layer in the causal convolution unit 151 may be two, three, five, ten, or the like, which is not limited herein.
Causal convolution output result W output by causal convolution unit 151dWill be input to the recurrent neural network element 152 for training. The Recurrent Neural Network Unit 152 may be a Recurrent Neural Network (RNN) model or a Gate Recurrent Unit (GRU) model of the RNN model. Because the GRU model is computationally efficient, in one embodiment of the present application, the GRU model is used to output the result W for causal convolutiondAnd (6) processing. As shown in fig. 9 and 10, specifically, the number of iterations is first set for initializing the parameters of the GRU model using the standard normal distribution. Subsequently, the causal convolution is output as a result WdInput into the GRU model. Wherein, the calculation process of the GRU model is as follows:
rd=σ(υr·[hd-1,Wd])
zd=σ(vz·[hd-1,Wd])
Od=σ(Wo·hd)
wherein upsilon isr,υz,Is a parameter to be trained, WdFor a certain time point to be predicted, rdIndicating a reset gate, zdTo renew the door, hd-1Hidden state for the previous day, hdIndicating the status output on the d-th day,as an intermediate parameter, sigma is a sigmoid activation function, and tanh is a trigonometric tangent function; []Representing vector concatenation, representing matrix multiplication, OdIs a predicted value.
Thus, the causal convolution is output as a result WdInputting the predicted value O to the recurrent neural network unit 152d。
wherein loss is a loss value; realdWhich is the true value, can be obtained from the raw traffic data. It should be noted that the formula is exemplified by the predetermined period of 1 day and the predetermined period of 7 days.
The second calculation unit 153 may determine whether the trained initial causal convolutional-cyclic neural network model satisfies the requirement according to whether the loss value is less than or equal to a predetermined threshold. If the loss value is less than or equal to the predetermined threshold, the determining unit 154 determines that the trained initial causal convolutional-cyclic neural network model is an intermediate causal convolutional-cyclic neural network model (i.e., the trained initial causal convolutional-cyclic neural network model meets the requirement); and if the loss value is larger than the preset threshold value, the trained initial causal convolution-circulation neural network model is considered to be not satisfied with the requirements, and the trained initial causal convolution-circulation neural network model needs to be trained continuously.
For example, when the trained initial causal convolutional-cyclic neural network model is further trained, parameters in the trained initial causal convolutional-cyclic neural network model can be changed by using a time back propagation algorithm, wherein each layer of weight parameters form a vector IcRm. Specifically, the gradient may be calculated first:
since the parameters are shared at every moment, the gradient of the current time point to be predicted also depends on the accumulated gradients at all the previous moments, and the accumulated square gradient is used in consideration of the accumulated positive and negative problems:
the parameter update equation is then:
Ij=Ij-1-αΔloss
where α is the learning rate. In one embodiment of the present application, α is 0.002, but in other examples, α may be other values, and is not limited herein.
According to the embodiment of the application, a great amount of time dependencies of the day are extracted by using the overlapping-level causal convolution, and then the time dependencies of the time points to be predicted in each day of the previous week are extracted by using the GRU model, so that a time sequence sensing range with a larger width can be extracted, model prediction can depend on historical traffic data of a great amount of time nodes, and the accuracy of traffic flow prediction is favorably improved.
With continued reference to fig. 1, in some embodiments, the training effectiveness of the intermediate causal convolutional-cyclic neural network model may be evaluated using the Mean Square Error (MSE) of the intermediate causal convolutional-cyclic neural network model. And if the evaluation result meets the preset condition (namely, the expected effect is achieved), the intermediate causal convolution-circulation neural network model is confirmed to be the target convolution-circulation neural network model, and if the evaluation result does not meet the preset condition (namely, the expected effect is not achieved), the training of the intermediate causal convolution-circulation neural network model is continued. It should be noted that, other methods than Mean Square Error (MSE) may also be used to evaluate the training effect of the intermediate causal convolutional-cyclic neural network model, and is not limited herein.
Referring to fig. 11, in some embodiments, the step 07 of predicting the traffic flow of the target road segment at the time point to be predicted by using the target causal convolution-recurrent neural network model includes:
071: acquiring original traffic data in a preset period, wherein the preset period comprises a time point to be predicted;
072: processing the original traffic data in the preset period according to the influence parameters to obtain corrected traffic data in the preset period;
073: inputting the corrected traffic data in a preset period into a target causal convolution-cyclic neural network model to obtain an initial predicted value of a time point to be predicted; and
074: and calculating a target predicted value according to the initial predicted value and the influence parameter.
Referring to fig. 12, in some embodiments, the prediction module 17 includes an obtaining unit 171, a second processing unit 172, a prediction unit 173, and a third computing unit 174. Step 071 may be implemented by the obtaining unit 171. Step 072 may be implemented by the second processing unit 172. Step 073 may be implemented by the prediction unit 173. Step 074 may be implemented by the third calculation unit 174. That is, the acquisition unit 171 may be configured to acquire raw traffic data within a predetermined period, the predetermined period including a time point to be predicted. The second processing unit 172 may be configured to process the raw traffic data within the predetermined period according to the influencing parameter to obtain the corrected traffic data within the predetermined period. The prediction unit 173 may be used to input the modified traffic data within a predetermined period to the target causal convolutional-cyclic neural network model to obtain an initial prediction value of a time point to be predicted. The third calculating unit 174 may be configured to calculate the target predicted value according to the initial predicted value and the influence parameter.
Specifically, for a time point to be predicted, original traffic data in a predetermined period of a target road section is acquired, wherein the predetermined period comprises the time point to be predicted. For example, the acquisition unit 171 may acquire all the raw traffic data of a week (including the current day) before the time point to be predicted, the raw traffic data including the traffic flow and the external factor information. It should be noted that the predetermined period may be three days, five days, ten days, etc., besides one week, and is not limited herein.
Subsequently, the second processing unit 172 may process all the original traffic data of the week before the time point to be predicted according to the influence parameter calculated in step 02 to obtain the corrected traffic data in the predetermined period. The specific processing procedure is the same as the processing procedure in step 021, and is not described herein again.
Subsequently, the prediction unit 173 may input the corrected traffic data within a predetermined period into the causal convolution unit 151 in units of a predetermined period (e.g., one day) to obtain a causal convolution output result for each day corresponding to the time point to be predicted. Subsequently, the prediction unit 173 further inputs the causal convolution output result for each day corresponding to the time point to be predicted into the cyclic convolution unit to output the initial prediction value O for the time point to be predicted.
Finally, the third calculating unit 174 adds the external dependency of the time point to be predicted to the initial predicted value O, that is, reduces the initial predicted value with the influence parameter to obtain the target predicted value, and the processing formula is as follows:
S=(1-Pd,t*ωT)O
thus, the target predicted value S can be obtained.
As described above, the traffic flow prediction method and the traffic flow prediction apparatus 10 (shown in fig. 2) according to the present embodiment can make the traffic flow prediction result more accurate, taking into account the influence of external factors on the traffic flow. In addition, the causal convolution-circulation neural network model is adopted to predict the traffic flow, and the causal convolution-circulation neural network model can sense a time sequence with a larger time width on the premise of reducing the calculated amount, so that the model can predict the traffic flow of a time point to be predicted depending on historical traffic data of a large number of time nodes, and the model is favorable for further improving the accuracy of traffic flow prediction and improving the real-time property of traffic flow prediction.
Referring to fig. 12, the present application further provides a computer apparatus 20. The computer device 20 includes a processor 21, a memory 22, and one or more programs. One or more programs are stored in the memory 22, and the one or more programs are executable by the processor 21 to implement the traffic flow prediction method of any of the above embodiments.
For example, referring to fig. 1 and 12 in conjunction, one or more programs can be executed by the processor 21 to implement the following steps:
01: acquiring original traffic data of a target road section within a preset time, wherein the original traffic data comprises traffic flow data of the target road section and external factor information influencing the traffic flow of the target road section;
02: processing the original traffic data to obtain an influence parameter of an external factor on the traffic flow of the target road section;
03: processing the raw traffic data to obtain corrected traffic data from which the influence of the external factors is removed;
04: dividing the corrected traffic data to obtain training set data and test set data;
05: training an initial causal convolution-cyclic neural network model using training set data to obtain an intermediate causal convolution-cyclic neural network model;
06: testing the intermediate causal convolution-cyclic neural network model by using the test set data to obtain an evaluation result, and confirming the intermediate causal convolution-cyclic neural network model as a target causal convolution-cyclic neural network model when the evaluation result meets a preset condition; and
07: and predicting the traffic flow of the target road section at the time point to be predicted by using the target causal convolution-cyclic neural network model.
As another example, referring to fig. 3 and 12, one or more programs can be executed by the processor 21 to implement the steps of:
021: processing the original traffic data to obtain normal traffic data which are not influenced by external factors and influenced traffic data which are influenced by the external factor information in the original traffic data; and
022: and calculating the influence parameters according to the influenced traffic data and the normal traffic data.
Referring to fig. 13, the present application further provides a non-volatile computer readable storage medium 30. The non-volatile computer-readable storage medium 30 includes a computer program. The computer program is executed by the processor 21 to implement the traffic flow prediction method according to any one of the above embodiments.
For example, referring to fig. 1 and 13 in combination, the computer program when executed by the processor 21 implements the steps of:
01: acquiring original traffic data of a target road section within a preset time, wherein the original traffic data comprises traffic flow data of the target road section and external factor information influencing the traffic flow of the target road section;
02: processing the original traffic data to obtain an influence parameter of an external factor on the traffic flow of the target road section;
03: processing the raw traffic data to obtain corrected traffic data from which the influence of the external factors is removed;
04: dividing the corrected traffic data to obtain training set data and test set data;
05: training an initial causal convolution-cyclic neural network model using training set data to obtain an intermediate causal convolution-cyclic neural network model;
06: testing the intermediate causal convolution-cyclic neural network model by using the test set data to obtain an evaluation result, and confirming the intermediate causal convolution-cyclic neural network model as a target causal convolution-cyclic neural network model when the evaluation result meets a preset condition; and
07: and predicting the traffic flow of the target road section at the time point to be predicted by using the target causal convolution-cyclic neural network model.
As another example, please refer to fig. 3 and 13 in conjunction, the computer program is executed by the processor 21 to implement the following steps:
021: processing the original traffic data to obtain normal traffic data which are not influenced by external factors and influenced traffic data which are influenced by the external factor information in the original traffic data; and
022: and calculating the influence parameters according to the influenced traffic data and the normal traffic data.
In the description herein, reference to the description of the terms "one embodiment," "some embodiments," "an illustrative embodiment," "an example," "a specific example" or "some examples" or the like means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and the scope of the preferred embodiments of the present application includes other implementations in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present application.
Although embodiments of the present application have been shown and described above, it is to be understood that the above embodiments are exemplary and not to be construed as limiting the present application, and that changes, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present application.
Claims (10)
1. A traffic flow prediction method characterized by comprising:
acquiring original traffic data of a target road section within a preset time, wherein the original traffic data comprises traffic flow data of the target road section and external factor information influencing the traffic flow of the target road section;
processing the original traffic data to obtain an influence parameter of an external factor on the traffic flow of the target road section;
processing the original traffic data according to the influence parameters to obtain corrected traffic data from which the influence of the external factors is removed;
dividing the corrected traffic data to obtain training set data and test set data;
training an initial causal convolution-recurrent neural network model using the training set data to obtain an intermediate causal convolution-recurrent neural network model;
testing the intermediate causal convolution-cyclic neural network model by using the test set data to obtain an evaluation result, and confirming the intermediate causal convolution-cyclic neural network model as a target causal convolution-cyclic neural network model when the evaluation result meets a preset condition; and
and predicting the traffic flow of the target road section at the time point to be predicted by utilizing the target causal convolution-circulation neural network model.
2. The traffic-flow prediction method according to claim 1, characterized in that the traffic-flow data includes at least one of traffic speed and traffic density;
the external factor information includes information of at least one external factor of low temperature, high temperature, normal, light rain, heavy rain, light snow, heavy wind, traffic condition and road repair.
3. The traffic-flow prediction method according to claim 2, wherein the processing the raw traffic data to obtain an influence parameter of an external factor on the traffic flow of the target link includes:
processing the original traffic data to obtain normal traffic data which are not influenced by the external factors and influenced traffic data which are influenced by the external factor information in the original traffic data; and
and calculating the influence parameters according to the influenced traffic data and the normal traffic data.
4. The traffic-flow prediction method according to claim 3, wherein each of the external factors corresponds to one of the influence parameters, and the calculating the influence parameter from the influenced traffic data and the normal traffic data includes:
for each time node within the predetermined time length, calculating a mean value of the normal traffic data belonging to the same time node to obtain a plurality of mean values corresponding to a plurality of the time nodes; and
for each external factor, calculating an initial parameter of the external factor under each time node according to the influenced traffic data influenced by the external factor and the plurality of average values; and
and calculating the influence parameters according to the plurality of initial parameters corresponding to the plurality of time nodes.
5. The traffic flow prediction method of claim 1, wherein the initial causal convolution-cyclic neural network model comprises a causal convolution unit and a cyclic neural network unit, and wherein training the initial causal convolution-cyclic neural network model with the training set data to obtain an intermediate causal convolution-cyclic neural network model comprises:
inputting the training set data into the causal convolution unit in units of a predetermined period of time to obtain a causal convolution output result;
inputting the causal convolution output result into the recurrent neural network element to obtain a predicted value;
calculating a loss value according to the real value and the predicted value; and
confirming that the trained initial causal convolutional-cyclic neural network model is the intermediate initial causal convolutional-cyclic convolutional neural network model when the loss value is smaller than a preset threshold value;
continuing to train the trained initial causal convolutional-cyclic neural network model when the loss value is greater than the predetermined threshold.
6. The traffic flow prediction method according to claim 5, wherein the causal convolution unit includes a plurality of causal convolution layers, and the cyclic neural network unit is a cyclic gate control unit.
7. The traffic flow prediction method according to claim 1, wherein the predicting the traffic flow of the target road segment at the time point to be predicted by using the target causal convolution-cyclic neural network model comprises:
acquiring original traffic data in a preset period, wherein the preset period comprises a time point to be predicted;
processing the original traffic data in the preset period according to the influence parameters to obtain corrected traffic data in the preset period;
inputting the corrected traffic data in the preset period into the target causal convolution-circulation neural network model to obtain an initial predicted value of the time point to be predicted; and
and calculating a target predicted value according to the initial predicted value and the influence parameter.
8. A traffic flow prediction apparatus characterized by comprising:
the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring original traffic data of a target road section within a preset time, and the original traffic data comprises traffic flow data of the target road section and external factor information influencing the traffic flow of the target road section;
the first processing module is used for processing the original traffic data to acquire an influence parameter of an external factor on the traffic flow of the target road section;
the second processing module is used for processing the original traffic data according to the influence parameters to obtain corrected traffic data without the influence of the external factors;
the dividing module is used for dividing the corrected traffic data to obtain training set data and test set data;
a training module for training an initial causal convolution-recurrent neural network model using the training set data to obtain an intermediate causal convolution-recurrent neural network model;
a test module for testing the intermediate causal convolutional-cyclic neural network model using the test set data to obtain an evaluation result, and confirming the intermediate causal convolutional-cyclic neural network model as a target causal convolutional-cyclic neural network model when the evaluation result satisfies a predetermined condition; and
and the prediction module is used for predicting the traffic flow of the target road section at the time point to be predicted by utilizing the target causal convolution-circulation neural network model.
9. A computer device, comprising:
a processor;
a memory; and
one or more programs stored in the memory, the one or more programs executable by the processor to implement the traffic flow prediction method of any of claims 1-7.
10. A non-transitory computer-readable storage medium containing a computer program, wherein the computer program, when executed by a processor, implements the traffic flow prediction method according to any one of claims 1 to 7.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011107026.8A CN112242060B (en) | 2020-10-16 | 2020-10-16 | Traffic flow prediction method and apparatus, computer device, and readable storage medium |
PCT/CN2020/139299 WO2022077767A1 (en) | 2020-10-16 | 2020-12-25 | Traffic flow prediction method and apparatus, computer device, and readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011107026.8A CN112242060B (en) | 2020-10-16 | 2020-10-16 | Traffic flow prediction method and apparatus, computer device, and readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112242060A true CN112242060A (en) | 2021-01-19 |
CN112242060B CN112242060B (en) | 2022-06-21 |
Family
ID=74169201
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011107026.8A Active CN112242060B (en) | 2020-10-16 | 2020-10-16 | Traffic flow prediction method and apparatus, computer device, and readable storage medium |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN112242060B (en) |
WO (1) | WO2022077767A1 (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112700072A (en) * | 2021-03-24 | 2021-04-23 | 同盾控股有限公司 | Traffic condition prediction method, electronic device, and storage medium |
CN113139646A (en) * | 2021-05-17 | 2021-07-20 | 中国水利水电科学研究院 | Data correction method and device, electronic equipment and readable storage medium |
CN114664090A (en) * | 2022-04-14 | 2022-06-24 | 山东大学 | Traffic data filling method and system based on recurrent neural network |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115240418B (en) * | 2022-07-20 | 2023-07-25 | 浙江科技学院 | Short-time traffic flow prediction method based on causal gate control-low-pass graph convolution network |
CN115457764B (en) * | 2022-08-24 | 2023-07-18 | 华南理工大学 | Road section traffic density estimation method, device and medium based on vehicle track data |
CN116362714B (en) * | 2023-02-21 | 2023-11-03 | 中国公路工程咨询集团有限公司 | Prediction method and device for pavement maintenance period |
CN116703008B (en) * | 2023-08-02 | 2023-10-31 | 山东高速股份有限公司 | Traffic volume prediction method, equipment and medium for newly built highway |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20150072471A (en) * | 2013-12-19 | 2015-06-30 | 건국대학교 산학협력단 | Traffic flow prediction system using spatiotemporal stochastic model |
CN109147324A (en) * | 2018-09-04 | 2019-01-04 | 东南大学 | A method of the traffic congestion probability forecast based on user feedback mechanisms |
CN110516846A (en) * | 2019-07-26 | 2019-11-29 | 浙江工业大学 | Intersection turning rate prediction technique based on LSTM neural network |
CN110969854A (en) * | 2019-12-13 | 2020-04-07 | 深圳先进技术研究院 | Traffic flow prediction method, system and terminal equipment |
CN111027686A (en) * | 2019-12-26 | 2020-04-17 | 杭州鲁尔物联科技有限公司 | Landslide displacement prediction method, device and equipment |
CN111047078A (en) * | 2019-11-25 | 2020-04-21 | 中山大学 | Traffic characteristic prediction method, system and storage medium |
CN111126680A (en) * | 2019-12-11 | 2020-05-08 | 浙江大学 | Road section traffic flow prediction method based on time convolution neural network |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10868829B2 (en) * | 2018-10-10 | 2020-12-15 | Northrop Grumman Systems Corporation | Predicted network traffic |
CN111027673B (en) * | 2019-11-14 | 2023-05-26 | 中山大学 | Subway passenger flow prediction method |
-
2020
- 2020-10-16 CN CN202011107026.8A patent/CN112242060B/en active Active
- 2020-12-25 WO PCT/CN2020/139299 patent/WO2022077767A1/en active Application Filing
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20150072471A (en) * | 2013-12-19 | 2015-06-30 | 건국대학교 산학협력단 | Traffic flow prediction system using spatiotemporal stochastic model |
CN109147324A (en) * | 2018-09-04 | 2019-01-04 | 东南大学 | A method of the traffic congestion probability forecast based on user feedback mechanisms |
CN110516846A (en) * | 2019-07-26 | 2019-11-29 | 浙江工业大学 | Intersection turning rate prediction technique based on LSTM neural network |
CN111047078A (en) * | 2019-11-25 | 2020-04-21 | 中山大学 | Traffic characteristic prediction method, system and storage medium |
CN111126680A (en) * | 2019-12-11 | 2020-05-08 | 浙江大学 | Road section traffic flow prediction method based on time convolution neural network |
CN110969854A (en) * | 2019-12-13 | 2020-04-07 | 深圳先进技术研究院 | Traffic flow prediction method, system and terminal equipment |
CN111027686A (en) * | 2019-12-26 | 2020-04-17 | 杭州鲁尔物联科技有限公司 | Landslide displacement prediction method, device and equipment |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112700072A (en) * | 2021-03-24 | 2021-04-23 | 同盾控股有限公司 | Traffic condition prediction method, electronic device, and storage medium |
CN113139646A (en) * | 2021-05-17 | 2021-07-20 | 中国水利水电科学研究院 | Data correction method and device, electronic equipment and readable storage medium |
CN113139646B (en) * | 2021-05-17 | 2023-10-31 | 中国水利水电科学研究院 | Data correction method and device, electronic equipment and readable storage medium |
CN114664090A (en) * | 2022-04-14 | 2022-06-24 | 山东大学 | Traffic data filling method and system based on recurrent neural network |
CN114664090B (en) * | 2022-04-14 | 2023-07-04 | 山东大学 | Traffic data filling method and system based on cyclic neural network |
Also Published As
Publication number | Publication date |
---|---|
WO2022077767A1 (en) | 2022-04-21 |
CN112242060B (en) | 2022-06-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112242060B (en) | Traffic flow prediction method and apparatus, computer device, and readable storage medium | |
CN109658695B (en) | Multi-factor short-term traffic flow prediction method | |
CN107909206B (en) | PM2.5 prediction method based on deep structure recurrent neural network | |
CN109035761B (en) | Travel time estimation method based on auxiliary supervised learning | |
CN111767517B (en) | BiGRU multi-step prediction method, system and storage medium applied to flood prediction | |
CN112949828B (en) | Graph convolution neural network traffic prediction method and system based on graph learning | |
KR20200023697A (en) | Apparatus for predicting traffic status information, method thereof and recoding medium for predicting traffic status information | |
US11416007B2 (en) | Computer-implemented method and system for evaluating uncertainty in trajectory prediction | |
Chen et al. | Probabilistic forecasting of drought: a hidden Markov model aggregated with the RCP 8.5 precipitation projection | |
CN111127879B (en) | Urban traffic flow prediction method based on generative countermeasure network | |
Muralidharan et al. | Probabilistic graphical models of fundamental diagram parameters for simulations of freeway traffic | |
CN114723188A (en) | Water quality prediction method, device, computer equipment and storage medium | |
Wang et al. | Probabilistic inference coupled with possibilistic reasoning for robust estimation of hydrologic parameters and piecewise characterization of interactive uncertainties | |
CN114564487B (en) | Meteorological raster data updating method combining forecast prediction | |
Chu et al. | Traffic flow prediction model based on LSTM with Finnish Dataset | |
CN113053536A (en) | Infectious disease prediction method, system and medium based on hidden Markov model | |
Hema et al. | Reconstructing missing hourly real-time precipitation data using a novel intermittent sliding window period technique for automatic weather station data | |
Tewari et al. | Application of LSTMs and HAND in rapid flood inundation mapping | |
Abotaleb et al. | Forecasting and modeling on average rainwater and vapor pressure in Chelyabinsk Russia using deep learning models | |
AL–Ruwashdi et al. | Simulating and predicting of urban expansion in Al Najaf city utilizing a Ca-Markov model | |
CN112365705B (en) | Method for determining road traffic volume | |
CN111524354B (en) | Method, system, medium and device for predicting urban traffic network path selection behavior based on language model | |
AYAR et al. | Long-term traffic flow estimation: a hybrid approach using location-basedtraffic characteristic | |
Ištoka Otković et al. | Analysis of neural network responses in calibration of microsimulation traffic model | |
CN111061966B (en) | Missing target searching method based on reinforcement learning algorithm |
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 |