CN107705556A - A kind of traffic flow forecasting method combined based on SVMs and BP neural network - Google Patents
A kind of traffic flow forecasting method combined based on SVMs and BP neural network Download PDFInfo
- Publication number
- CN107705556A CN107705556A CN201710779559.2A CN201710779559A CN107705556A CN 107705556 A CN107705556 A CN 107705556A CN 201710779559 A CN201710779559 A CN 201710779559A CN 107705556 A CN107705556 A CN 107705556A
- Authority
- CN
- China
- Prior art keywords
- data
- traffic flow
- mrow
- neural network
- svms
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
Landscapes
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of Short-time Traffic Flow Forecasting Methods combined based on SVMs and BP neural network, the collection of historical traffic flow data first, traffic flow data is pre-processed using method for normalizing, data set after being normalized, the data set after normalization is divided into training dataset and test data set；Then analysis is predicted to test set using SVM models, obtains prediction result, residual sequence is analyzed using BP neural network model, obtain revised residual sequence；Prediction result obtained by SVM models is added with the amendment residual sequence obtained by BP neural network model, obtains final prediction data；Test data set and prediction data are compared, analytical error.The traffic flow forecasting method that the present invention is combined using SVMs and BP neural network, is analyzed sample data by supporting vector machine model, obtains higher prediction accuracy using less data collection, reduce amount of calculation and difficulty in computation.
Description
Technical field
The present invention relates to the technical fields such as machine learning method and forecasting traffic flow, and in particular to one kind is based on supporting vector
The traffic flow forecasting method that machine and BP neural network combine.
Prior art and background technology
With the development of China's automobile industry, city and Expressway Road congestion problems are increasingly serious, accurately, timely
Telecommunication flow information extremely closes to the successful application of intelligent transportation system (Intelligent Transportation System, ITS)
It is important.It can help road user to make more preferable trip decision-making, alleviate traffic congestion, reduce carbon emission, and improve friendship
Logical operational efficiency.And the premise and key that these are realized are that Short-Term Traffic Flow accurately can be predicted, the standard of prediction
Exactness directly determines the efficiency of traffic circulation.
Forecasting traffic flow model is broadly divided into time series models, nonparametric Regression Model and neural network model.In early days
Researchers predict traffic using based on the parameterized model of time series analysis.The model is unlike other time sequence side
Method equally needs the initialization simulation of fixation.It regards the magnitude of traffic flow at a certain moment as more in general non-stationary stochastic ordering
Row, typically carry 3 or 6 model parameters.Non parametric regression is a kind of the non-of suitable uncertain, nonlinear dynamical system
Parameter model method.It is not required to priori, only needs enough historical datas, finds similar to current point in historical data
" neighbour ", and predict subsequent time value with those " neighbours ".Therefore, particularly when there is special event generation, prediction result will
It is more accurate than parameter model.U.S.Federal office points out in the report about developing intelligent transportation system:Advanced traffic control
System processed should not only possess the advantages of existed system, and what is more important will can make full use of the experience constantly accumulated, effectively
Ground produces control strategy, model is had the ability for carrying out study and experience accumulation according to historical data.Eleni etc. is reviewed closely
The achievement in research of traffic flow forecasting method over 10 years, ten challenge directions in forecasting traffic flow field are summarized, are then pointed out
Researcher needs statistics and artificial intelligence approach composition complementary and to provide unified common data sets pre- to solve traffic
These challenges of survey field, so as to improve forecasting traffic flow performance.In existing technology, such as entitled " one kind is based on deep learning
The traffic flow forecasting method of neural network structure " (Application No. 201510478215.9) is using depth encoder model to collection
Traffic flow data be trained, depth autocoder model is adjusted in the training process, finally using adjustment after
Depth autocoder model Short-term Traffic Flow is predicted, improve the performance and accuracy of short-term traffic flow forecasting.
In forecasting traffic flow field, present most popular method is predicted in short-term based on the nonparametric model of machine learning
Traffic flow, this method possess higher precision of prediction, but need complicated parameter to estimate on the basis of a large amount of uninterrupted datas
Meter, and the parameter calculated can not transplant.In a practical situation, data something lost is often easily caused due to various
Leakage, causes model accuracy to reduce.Current forecasting traffic flow algorithm still largely relies on historical data, in the base of a large amount of historical datas
Model is established on plinth causes amount of calculation huge, causes forecasting traffic flow efficiency low.
The content of the invention
The technical problem to be solved in the present invention is that solve in short-time traffic flow forecast, because the traffic flow data of magnanimity is led
SVMs (SVM) the training time length of cause, big to computer resource usage, estimated performance is poor, and BP (Back
Propagation) Generalization Ability of Neural Network difference is i.e. poor to the adaptability of fresh sample, and prediction result easily produces over-fitting
The problems such as.
To solve the above problems, the present invention is proposed by the method for being combined SVMs and BP neural network come pre-
Short-term traffic flow is surveyed, reduces the dependence to a large amount of historical datas, improves estimated performance and the adaptability to fresh sample.Specifically
A kind of Short-time Traffic Flow Forecasting Methods combined based on SVMs and BP neural network of technical scheme, including following step
Suddenly：
Step 1:Historical traffic flow data is gathered, traffic flow data is pre-processed using method for normalizing, returned
Data set after one change, training dataset and test data set are divided into by the data set after normalization；
Step 2:Analysis is predicted to test data set using SVM models, prediction result is obtained, uses BP neural network
Model is analyzed residual sequence, obtains revised residual sequence；
Step 3:By the prediction result obtained by SVM models and the amendment residual sequence phase obtained by BP neural network model
Add, obtain final prediction data；
Step 4:Test data set and final prediction data are compared, and analytical error.
Further, the normalization detailed process in above-mentioned steps 1 is as follows：
Minimum value min and maximum max in some sample of calculating historical traffic flow data respectively, uses min-max
Data are normalized for standardized method so that the traffic flow data result after normalizing is mapped between [0-1], i.e. root
According to traffic flow data set F={ ft| t=1,2 ..., T } try to achieve set in maximum max and minimum value min, in set
Each data calculate：
X in formula*The traffic flow data after normalized is represented, min represents the minimum value in sample data, and max is represented
Sample data maximum, x represent to treat the historical traffic flow data of normalized.
In step 1, using in historical traffic flow data percent 80 data as training set, percentage after normalized
20 data as test set.
Further, step 2 specifically includes following steps：
2.1：The forecasting traffic flow model based on SVMs and BP neural network is established, utilizes the training after normalization
Parameter C, γ, parameter γ after collection sample training SVM models find out optimization with cross validation are the parameters of Sigmoid kernel functions,
Parameter C is the parameter of SVM penalties；
2.2：Analysis is predicted to initial data using SVM models, prediction result is obtained, is designated asOriginal series and
The difference of prediction result sequence is new sequence, is designated as eiSequence, with BP neural network model to eiSequence i.e. residual sequence enter
Row analysis, obtains revised residual sequence, is designated as ei′。
Prediction result obtained by SVM models is added with the amendment residual sequence obtained by BP neural network model, obtained
To final pre- data, i.e.
Step 4 includes：Error analysis, calculation formula are carried out to prediction data by mean absolute percentage error MAPE
It is as follows：
In formulaThe mean absolute percentage error of test data set and prediction data is represented,The root-mean-square error of test data set and prediction data is represented, f represents the observation of traffic flow,Represent traffic
The predicted value of stream, N represent that forecasting traffic flow is worth quantity, fiI-th of traffic flow parameter that test data is concentrated is represented,Represent
I-th of predicting traffic flow parameter that prediction data is concentrated, i span is 1,2...N.
Compared with prior art, the beneficial effects of the present invention are：
1st, the traffic flow forecasting method being combined using SVMs and BP neural network, passes through supporting vector machine model
Sample data is analyzed, higher prediction accuracy is obtained using less data collection, reduces amount of calculation and calculates hardly possible
Degree.
2nd, solve conventional method generalization ability difference it is i.e. poor to the adaptability of fresh sample, prediction result easily produces
The problems such as over-fitting.The accuracy of forecasting traffic flow is largely further increased, improves the stability of prediction, no
Easily there is apparent error.
Brief description of the drawings
Fig. 1 is SVMs structure chart.
Fig. 2 is BP neural network structure chart.
Fig. 3 is the key step flow chart of the present invention.
Fig. 4 is SVMs and BP neural network built-up pattern structure chart.
Fig. 5 is the prediction result and existing method prediction result comparison diagram of the inventive method.
Embodiment
The present invention is described further with example below in conjunction with the accompanying drawings, it should be noted that described embodiment
The understanding of the present invention is intended merely to facilitate, and does not play any restriction effect to it.
The traffic flow forecasting method being combined based on SVMs and BP neural network, main flow and its structure chart are such as
Shown in Fig. 3 and Fig. 4, comprise the following steps:
Step 1:The collection of historical traffic flow data, traffic flow data is pre-processed using method for normalizing, obtained
Data set after normalization, the data set after normalization is divided into training dataset and test data set；
Step 2:Analysis is predicted to test set using SVM models, obtains prediction result, uses BP neural network mould
Type is analyzed residual sequence, obtains revised residual sequence.
Step 3:By the prediction result obtained by SVM models and the amendment residual sequence phase obtained by BP neural network model
Add, obtain final prediction data.
Step 4:Test data set and prediction data are compared, analytical error.
The collection of traffic flow data：
The method of traffic flow data sampling has a lot, mainly there is the methods of ultrasound examination, infrared detection, microwave remote sensor
Obtain.
The vehicle number that the historical traffic flows data of acquisition are specific observation station or section is passed through in a certain time interval.
The time interval of the formulation can be specified (such as 5 minutes) according to forecast demand.
History observation data acquisition system is expressed as F={ ft| t=1,2 ..., T }, wherein ftRepresent that the specific observation station of road network obtains
T-th of the historical traffic stream parameter obtained.Difference between T moment and T+1 moment is predicted time interval of delta t (such as 5 minutes).
Minimum value min and maximum max in some sample of calculating historical traffic flow data respectively, uses min-max
Data are normalized standardization (Min-Max Normalization) method so that the traffic flow data after normalizing
As a result it is mapped between [0-1].
Specifically, according to traffic flow data set F={ ft| t=1,2 ..., T try to achieve set in maximum max and
Minimum value min, each data in set are calculated：
So as to the historical traffic flow data after being normalized, wherein x* represents the traffic fluxion after normalized
According to min represents the minimum value in sample data, and max represents sample data maximum, and x represents to treat that the history of normalized is handed over
Through-flow data.
And using in historical traffic flow data percent 80 data as training set, percent 20 data are as test
Collection.
Above-mentioned steps 2 establish the establishment step of the forecasting traffic flow model based on SVMs and BP neural network such as
Under：
Combination forecasting is actually a kind of residual GM type built-up pattern, is instructed using the training set sample after normalization
It is Sigmoid core letters to practice parameter C, γ, parameter γ after SVM models find out optimization with cross validation (cross validation)
Several parameters, parameter C are the parameter i.e. penalty factors of SVM penalties；
The structure of SVMs is as shown in Figure 1.In Fig. 1, function k is kernel function, and its species mainly has：
1 linear kernel function：k(x,xi)=xTxi；
2 Polynomial kernel functions：k(x,xi)=(γ xTxi+r)p, γ ＞ 0；
3RBF Radial basis kernel functions：k(x,xi)=exp (- γ | | x-xi||2), γ ＞ 0
4Sigmoid kernel functions:k(x,xi)=tanh (γ xTxi+r)
Kernel function of the Sigmoid kernel functions as SVMs is used in this example.
Penalty factor determines the degree size for the loss that outlier is brought, and C is bigger, and the loss to object function is also got over
Greatly.
Utilize LIBSVM MATLAB tool boxes training sample (this example completes parameter optimization using LIBSVM tool boxes)
SVM is based on structural risk minimization, and whole solution procedure is converted into a convex quadratic programming problem, and it is solved
Global optimum and unique.
Analysis is predicted to initial data using SVM models, prediction result is obtained, is designated asOriginal series and prediction are tied
The difference of infructescence row is new sequence, is designated as eiSequence, with BP neural network model to eiSequence i.e. residual sequence are divided
Analysis, obtains revised residual sequence, is designated as ei′。
Residual error is referring to the difference between actual observation value and estimate.
BP neural network is one of widest neutral net of current application.It is a kind of Multi-layered Feedforward Networks, and it is by mistake
The inverse propagation algorithm of difference is trained.BP neural network learns and the substantial amounts of input of storage and output mode mapping relations.Its
It is to use steepest descent method to practise rule, the weights and threshold value of network is constantly adjusted by backpropagation so that the error of network
Quadratic sum is minimum.The structure of BP neural network model includes input layer, hidden layer and output layer.As shown in Figure 2.
The mapping that x represents h to hidden layer is inputted in BP neural network, is expressed as：H (x)=f (x)=σf(w+bn), σfFor
Nonlinear activation function, generally Sigmoid functions, its expression formula are：
σ (x)=1/1+e-x
Prediction result obtained by SVM models is added with the amendment residual sequence obtained by BP neural network model, obtained
To final prediction data.That is,
Test data set and prediction data are compared, carry out error analysis.Specifically, error can be commented by two indices
Estimate, i.e. mean absolute percentage error (Mean Absolute Percentage Error, MAPE) and root-mean-square error (Root
Mean Square Error, RMSE), their definition is as follows:
Wherein f is the observation of traffic flow,For the predicted value of traffic flow, n represents the quantity of forecasting traffic flow value.This hair
The prediction result of bright method and existing method prediction result comparison diagram are as shown in Figure 5.
Claims (6)
1. a kind of Short-time Traffic Flow Forecasting Methods combined based on SVMs and BP neural network, it is characterised in that including such as
Lower step：
Step 1:Historical traffic flow data is gathered, traffic flow data is pre-processed using method for normalizing, normalized
Data set afterwards, the data set after normalization is divided into training dataset and test data set；
Step 2:Analysis is predicted to test data set using SVM models, obtains prediction result, uses BP neural network model
Residual sequence is analyzed, obtains revised residual sequence；
Step 3:Prediction result obtained by SVM models is added with the amendment residual sequence obtained by BP neural network model,
Obtain final prediction data；
Step 4:Test data set and final prediction data are compared, and analytical error.
2. the Short-time Traffic Flow Forecasting Methods according to claim 1 combined based on SVMs and BP neural network,
It is characterized in that the normalization detailed process in step 1 is as follows：
Minimum value min and maximum max in some sample of calculating historical traffic flow data respectively, uses min-max standards
Data are normalized change method so that the traffic flow data result after normalizing is mapped between [0-1], i.e., according to friendship
Through-flow data acquisition system F={ ft| t=1,2 ..., T } try to achieve set in maximum max and minimum value min, to every in set
Individual data calculate：
<mrow>
<mi>x</mi>
<mo>*</mo>
<mo>=</mo>
<mfrac>
<mrow>
<mi>x</mi>
<mo>-</mo>
<mi>min</mi>
</mrow>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
<mo>-</mo>
<mi>min</mi>
</mrow>
</mfrac>
</mrow>
X in formula*The traffic flow data after normalized is represented, min represents the minimum value in sample data, and max represents sample number
According to maximum, x represents to treat the historical traffic flow data of normalized.
3. the Short-time Traffic Flow Forecasting Methods according to claim 1 combined based on SVMs and BP neural network,
It is characterized in that using in historical traffic flow data percent 80 data as training set after normalized in step 1, hundred
/ 20 data are as test set.
4. the Short-time Traffic Flow Forecasting Methods according to claim 1 combined based on SVMs and BP neural network,
It is characterized in that step 2 specifically includes following steps：
4.1：The forecasting traffic flow model based on SVMs and BP neural network is established, utilizes the training set sample after normalization
Parameter C, γ, parameter γ after this training SVM models find out optimization with cross validation are the parameters of Sigmoid kernel functions, parameter C
It is the parameter of SVM penalties；
4.2：Analysis is predicted to initial data using SVM models, prediction result is obtained, is designated asOriginal series and prediction are tied
The difference of infructescence row is new sequence, is designated as eiSequence, with BP neural network model to eiSequence i.e. residual sequence are divided
Analysis, obtains revised residual sequence, is designated as e 'i。
5. the Short-time Traffic Flow Forecasting Methods according to claim 4 combined based on SVMs and BP neural network,
It is characterized in that the prediction result obtained by SVM models is added with the amendment residual sequence obtained by BP neural network model,
Obtain final pre- data, i.e.
6. the Short-time Traffic Flow Forecasting Methods according to claim 1 combined based on SVMs and BP neural network,
It is characterized in that the step 4 includes：Error analysis, meter are carried out to prediction data by mean absolute percentage error MAPE
It is as follows to calculate formula：
<mrow>
<mi>M</mi>
<mi>A</mi>
<mi>P</mi>
<mi>E</mi>
<mrow>
<mo>(</mo>
<mi>f</mi>
<mo>,</mo>
<mover>
<mi>f</mi>
<mo>^</mo>
</mover>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>N</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<mo>|</mo>
<mfrac>
<mrow>
<msub>
<mi>f</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msubsup>
<mi>f</mi>
<mi>i</mi>
<mo>&prime;</mo>
</msubsup>
</mrow>
<msub>
<mi>f</mi>
<mi>i</mi>
</msub>
</mfrac>
<mo>|</mo>
</mrow>
<mrow>
<mi>R</mi>
<mi>M</mi>
<mi>S</mi>
<mi>E</mi>
<mrow>
<mo>(</mo>
<mi>f</mi>
<mo>,</mo>
<mover>
<mi>f</mi>
<mo>^</mo>
</mover>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msup>
<mrow>
<mo>&lsqb;</mo>
<mfrac>
<mn>1</mn>
<mi>n</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<mrow>
<mo>(</mo>
<msup>
<mrow>
<mo>|</mo>
<mrow>
<msub>
<mi>f</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mover>
<mi>f</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
</mrow>
<mo>|</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>)</mo>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
</msup>
</mrow>
In formulaThe mean absolute percentage error of test data set and prediction data is represented,Table
Showing the root-mean-square error of test data set and prediction data, f represents the observation of traffic flow,Represent the predicted value of traffic flow, N
Represent that forecasting traffic flow is worth quantity, fiI-th of traffic flow parameter that test data is concentrated is represented,Represent that prediction data is concentrated
I-th of predicting traffic flow parameter, i span is 1,2...N.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710779559.2A CN107705556A (en) | 2017-09-01 | 2017-09-01 | A kind of traffic flow forecasting method combined based on SVMs and BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710779559.2A CN107705556A (en) | 2017-09-01 | 2017-09-01 | A kind of traffic flow forecasting method combined based on SVMs and BP neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107705556A true CN107705556A (en) | 2018-02-16 |
Family
ID=61171569
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710779559.2A Pending CN107705556A (en) | 2017-09-01 | 2017-09-01 | A kind of traffic flow forecasting method combined based on SVMs and BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107705556A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109146156A (en) * | 2018-08-03 | 2019-01-04 | 大连理工大学 | A method of for predicting charging pile system charge volume |
| CN109377754A (en) * | 2018-10-29 | 2019-02-22 | 东南大学 | A kind of short-term traffic flow speed predicting method under car networking environment |
| CN109460727A (en) * | 2018-10-31 | 2019-03-12 | 中国矿业大学 | A kind of examination hall monitoring system and method based on Human bodys' response |
| CN109635246A (en) * | 2018-12-06 | 2019-04-16 | 西南交通大学 | A kind of multiattribute data modeling method based on deep learning |
| CN109840628A (en) * | 2019-01-17 | 2019-06-04 | 湖南大学 | A kind of multizone speed prediction method and system in short-term |
| WO2020010717A1 (en) * | 2018-07-13 | 2020-01-16 | 南京理工大学 | Short-term traffic flow prediction method based on spatio-temporal correlation |
| CN111435469A (en) * | 2019-01-11 | 2020-07-21 | 中国长江电力股份有限公司 | Method for predicting warehousing flow of downstream power station in cascade power station |
| CN112820120A (en) * | 2020-12-30 | 2021-05-18 | 杭州趣链科技有限公司 | Multi-party traffic flow space-time cross validation method based on alliance chain |
| CN112862164A (en) * | 2021-01-22 | 2021-05-28 | 桂林电子科技大学 | Dry clutch temperature prediction method based on dynamic neural network time sequence prediction |
| CN113553350A (en) * | 2021-05-27 | 2021-10-26 | 四川大学 | Traffic flow partition model for similar evolution mode clustering and dynamic time zone partitioning |
| CN115358475A (en) * | 2022-08-29 | 2022-11-18 | 河南农业大学 | Disaster prediction method and system based on support vector machine and gray BP neural network |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105389980A (en) * | 2015-11-09 | 2016-03-09 | 上海交通大学 | Short-time traffic flow prediction method based on long-time and short-time memory recurrent neural network |
| KR101638368B1 (en) * | 2015-01-02 | 2016-07-11 | 경희대학교 산학협력단 | Prediction System And Method of Urban Traffic Flow Using Multifactor Pattern Recognition Model |
-
2017
- 2017-09-01 CN CN201710779559.2A patent/CN107705556A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101638368B1 (en) * | 2015-01-02 | 2016-07-11 | 경희대학교 산학협력단 | Prediction System And Method of Urban Traffic Flow Using Multifactor Pattern Recognition Model |
| CN105389980A (en) * | 2015-11-09 | 2016-03-09 | 上海交通大学 | Short-time traffic flow prediction method based on long-time and short-time memory recurrent neural network |
Non-Patent Citations (2)
| Title |
|---|
| 傅贵等: "基于支持向量机回归的短时交通流预测模型", 《华南理工大学学报》 * |
| 成云等: "基于ARIMA和小波神经网络组合模型的交通流预测", 《计算机技术与发展》 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020010717A1 (en) * | 2018-07-13 | 2020-01-16 | 南京理工大学 | Short-term traffic flow prediction method based on spatio-temporal correlation |
| CN109146156A (en) * | 2018-08-03 | 2019-01-04 | 大连理工大学 | A method of for predicting charging pile system charge volume |
| CN109146156B (en) * | 2018-08-03 | 2021-12-03 | 大连理工大学 | Method for predicting charging amount of charging pile system |
| CN109377754A (en) * | 2018-10-29 | 2019-02-22 | 东南大学 | A kind of short-term traffic flow speed predicting method under car networking environment |
| CN109377754B (en) * | 2018-10-29 | 2021-07-02 | 东南大学 | Short-term traffic flow speed prediction method in Internet of vehicles environment |
| CN109460727A (en) * | 2018-10-31 | 2019-03-12 | 中国矿业大学 | A kind of examination hall monitoring system and method based on Human bodys' response |
| CN109635246B (en) * | 2018-12-06 | 2021-07-13 | 西南交通大学 | Multi-attribute data modeling method based on deep learning |
| CN109635246A (en) * | 2018-12-06 | 2019-04-16 | 西南交通大学 | A kind of multiattribute data modeling method based on deep learning |
| CN111435469A (en) * | 2019-01-11 | 2020-07-21 | 中国长江电力股份有限公司 | Method for predicting warehousing flow of downstream power station in cascade power station |
| CN111435469B (en) * | 2019-01-11 | 2022-04-19 | 中国长江电力股份有限公司 | Method for predicting warehousing flow of downstream power station in cascade power station |
| CN109840628A (en) * | 2019-01-17 | 2019-06-04 | 湖南大学 | A kind of multizone speed prediction method and system in short-term |
| CN109840628B (en) * | 2019-01-17 | 2023-03-24 | 湖南大学 | Short-time multi-zone vehicle speed prediction method and system |
| CN112820120A (en) * | 2020-12-30 | 2021-05-18 | 杭州趣链科技有限公司 | Multi-party traffic flow space-time cross validation method based on alliance chain |
| CN112862164A (en) * | 2021-01-22 | 2021-05-28 | 桂林电子科技大学 | Dry clutch temperature prediction method based on dynamic neural network time sequence prediction |
| CN112862164B (en) * | 2021-01-22 | 2022-07-12 | 桂林电子科技大学 | Dry clutch temperature prediction method based on dynamic neural network time sequence prediction |
| CN113553350A (en) * | 2021-05-27 | 2021-10-26 | 四川大学 | Traffic flow partition model for similar evolution mode clustering and dynamic time zone partitioning |
| CN113553350B (en) * | 2021-05-27 | 2023-07-18 | 四川大学 | Traffic flow partition model for similar evolution mode clustering and dynamic time zone division |
| CN115358475A (en) * | 2022-08-29 | 2022-11-18 | 河南农业大学 | Disaster prediction method and system based on support vector machine and gray BP neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107705556A (en) | A kind of traffic flow forecasting method combined based on SVMs and BP neural network | |
| Zhang et al. | Modeling and simulating of reservoir operation using the artificial neural network, support vector regression, deep learning algorithm | |
| Yang et al. | A new air quality monitoring and early warning system: Air quality assessment and air pollutant concentration prediction | |
| Wu et al. | Evolving RBF neural networks for rainfall prediction using hybrid particle swarm optimization and genetic algorithm | |
| Kisi et al. | Forecasting daily lake levels using artificial intelligence approaches | |
| CN101480143B (en) | Method for predicating single yield of crops in irrigated area | |
| Shiri et al. | Estimation of daily suspended sediment load by using wavelet conjunction models | |
| Li et al. | Evaluation of urban green space landscape planning scheme based on PSO-BP neural network model | |
| Yu et al. | A neural network ensemble method for precision fertilization modeling | |
| CN110751318A (en) | IPSO-LSTM-based ultra-short-term power load prediction method | |
| CN106022954A (en) | Multiple BP neural network load prediction method based on grey correlation degree | |
| CN106529818A (en) | Water quality evaluation prediction method based on fuzzy wavelet neural network | |
| CN110363349A (en) | A kind of LSTM neural network hydrologic(al) prognosis method and system based on ASCS | |
| CN103886395A (en) | Reservoir optimal operation method based on neural network model | |
| CN109871622A (en) | A kind of low-voltage platform area line loss calculation method and system based on deep learning | |
| Li et al. | A novel combined prediction model for monthly mean precipitation with error correction strategy | |
| CN104732091A (en) | Cellular automaton river bed evolution prediction method based on natural selection ant colony algorithm | |
| Zhang et al. | Study on water quality prediction of urban reservoir by coupled CEEMDAN decomposition and LSTM neural network model | |
| CN113420868A (en) | Traveling salesman problem solving method and system based on deep reinforcement learning | |
| CN110969312A (en) | Short-term runoff prediction coupling method based on variational modal decomposition and extreme learning machine | |
| CN106570594A (en) | Similar day photovoltaic power generation short period prediction method based on TMBP | |
| Robati et al. | Inflation rate modeling: Adaptive neuro-fuzzy inference system approach and particle swarm optimization algorithm (ANFIS-PSO) | |
| CN109408896B (en) | Multi-element intelligent real-time monitoring method for anaerobic sewage treatment gas production | |
| Guo et al. | A combined model based on sparrow search optimized BP neural network and Markov chain for precipitation prediction in Zhengzhou City, China | |
| CN108764523A (en) | Predictive Methods of Road Accidents based on unbiased nonhomogeneous gray model and geneva model |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180216 |
|
| RJ01 | Rejection of invention patent application after publication |