CN109614742A - A kind of sea level height duration prediction algorithm - Google Patents
A kind of sea level height duration prediction algorithm Download PDFInfo
- Publication number
- CN109614742A CN109614742A CN201811593352.7A CN201811593352A CN109614742A CN 109614742 A CN109614742 A CN 109614742A CN 201811593352 A CN201811593352 A CN 201811593352A CN 109614742 A CN109614742 A CN 109614742A
- Authority
- CN
- China
- Prior art keywords
- model
- prediction
- formula
- node
- sea level
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Geometry (AREA)
- Game Theory and Decision Science (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of sea level height duration prediction algorithms, carry out the prediction of sea level height duration using ARMA model；The prediction of sea level height duration is carried out using artificial nerve network model；Establish sea level height duration Statistical Prediction Model.The beneficial effects of the invention are as follows with global tidal station information products be main data source, sea level height duration prediction experiment is carried out by statistical methods, establish sea level height duration Statistical Prediction Model, and precision of prediction can satisfy Service assurance demand, provide technical support for Marine Environmental Security guarantee.
Description
Technical field
The invention belongs to oceanography technical fields, are related to a kind of sea level height duration prediction algorithm.
Background technique
The prediction prediction business of meteorological department, China temporally scale can be divided into 1~3h close in short-term prediction, 1~
It is more than the Short-term Weather prediction of 3d, the medium-range weather prediction of 4~10d, the duration weather forecasting of 10~30d and moon scale short
Phase climatic prediction.In synoptic climate prediction business, the duration prediction of 10~30d is the difficult point in " seamless prediction ".Limit
In the technical level of current weather climatic prediction, the significant weather process of (moon scale) within 10d or more and 30d is also difficult to do
It more accurately predicts out.Wind speed forecasting method is roughly divided into three classes at present: the mutually knot of statistical method, physical method and the two
It closes.Statistical method does not consider the physical process of wind speed variation, using certain mathematical statistics method, historical data and wind speed it
Between establish a kind of mapping relations, wind speed is predicted with this, it is contemplated that the advantage and disadvantage of single statistical method are different, by
The combination statistical prediction methods that statistical methods combine also gradually are adopted.For marine mobile platform, due to that can not configure
Mainframe computer is limited by computing resource, it is difficult to realize businessization prediction using the method for numerical model.And statistical forecast
Although equally having certain limitation, such as excessive dependence observational data, prediction timeliness is shorter etc..Artificial neural network
It (ANNs) is a kind of very strong nonlinear function approximation device, it has stronger computing capability.Due to ANNs have parallel organization, compared with
Strong self-learning capability and generalization, make it through setting inputoutput pair can preferably solve science and engineering in it is arbitrarily multiple
The modeling problem of miscellaneous Nonlinear Mapping.ANNs has been widely used in pattern-recognition, function approximation, control, optimization, prediction
Equal fields.It is predicted in conclusion carrying out sea level height duration using statistical theory, Marine Environmental Security is ensured with important
Value and significance.
Summary of the invention
The purpose of the present invention is to provide a kind of sea level height duration prediction algorithms, and the beneficial effects of the invention are as follows with complete
Ball tidal station information products (being an ocean big data) are main data source, carry out sea level height by statistical methods and hold
Continuous property prediction experiment, establishes sea level height duration Statistical Prediction Model, and precision of prediction can satisfy Service assurance demand, be
Marine Environmental Security, which ensures, provides technical support.
The technical scheme adopted by the invention is that following the steps below:
(1) prediction of sea level height duration is carried out using ARMA model；
(2) prediction of sea level height duration is carried out using artificial nerve network model；
(3) sea level height duration Statistical Prediction Model is established.
Further, ARMA model in step (1):
When being predicted with arma modeling, { X is allowedt(t=1,2,3 ..., N) expression time series, { αtIt is zero-mean, side
Difference isNormal white noise process,Indicate the auto-regressive parameter of model, θj(j=1,2 ... p) indicate mould
The sliding average parameter of type, B expression move back difference operator.Then ARMA (p.q) model is as follows
Above-mentioned model: model identification, parameter Estimation and model testing is established by three steps, model identification refers to judgement prediction
Model is AR, MA or ARMA, then determines model order, that is, determines p and q, parameter Estimation refers to after identification model, passes through
Suitable method calculates the unknown parameter in (1), that is, determinesθjWith
1. determining the order of model
By sample autocorrelation coefficient and the form of sample partial correlation coefficient come identification model classification, to time series data
After carrying out smoothing preprocessing, calculating original series auto-correlation function ACF, deviation―related function PACF, for stationary time first
Sequence Xt(t=1,2,3 ..., N), specific judgement are as follows:
Mean value are as follows:
Variance are as follows:
Covariance are as follows:
Correlation function are as follows:
Deviation―related function are as follows:
K=1,2 ...；I=1,2 ... k-1.
The order of model is by judging that the truncation of sum primarily determines；
2. parameter Estimation
Least-squares algorithm is selected to carry out parameter Estimation；
3. model testing
It is tested first with the model being fitted, while calculating residual sequence, if the residual sequence of model is white
Noise sequence, then model is effectively, otherwise to need to increase model order, then re-start parameter Estimation and model testing,
Until model is effective.
Further, artificial nerve network model in step (2):
Assuming that there is N number of training sample (Xk, Yk*), k=1,2 ..., N, to some sample (Xk, Yk* for), net is first passed through
Network model forward-propagating, if Xk=(X1k, X2k..., Xnk) be sample K input node, successively pass through input layer, hidden layer by
Layer processing, is finally exported by output layer, show that the BP network training output of training sample K is Yk=(Y1k, Y2k..., Ynk), sample
The network of this K exports YkY is exported with expectationk* the difference between, as network error, then, by error amount from output layer to defeated
Enter layer inversely to propagate, in reverse communication process, constantly corrects the threshold values of each layer neuron；
Error is inversely propagated in calculating process, if
Input vector XkThe input of l node layer j is traveled to after input；
The output of l node layer j；
The threshold values of l-1 layers of node i connection l node layer j；
n(l-1): l-1 layers of node number；
F: excitation function；
For the neuron excitation function of BP network usually using Sigmoid type function, the input according to BP network neural member is defeated
Relationship out has:
Neuron j exports the network query function of sample KWith sample K to the desired output of neuron jError
Are as follows:
If the l layers of output layer for BP network, i.e. neuron j is output node, then Sample
Error between the calculating output and expectation output of this K is:
If in N number of training sample, the output error of the m output node of any sample K within the scope of defined,
HaveJ=1,2 ..., m, then training process leaves it at that；Otherwise, continue inversely to propagate error, by error
Negative gradient constantly is corrected threshold values W, it may be assumed that
In formula, η is training rate, 0 < η < 1；
By (5) formula, (6) formula, (7) formula is obtained:
Wherein,
In order to obtainCalculation formula, discuss in two kinds of situation below.
(1) it if l layers are output layer, then can be obtained by (8) formula:
(13) formula can be obtained by (9) formula, (10) formula:
(14) formula can be obtained by (9) formula, (10) formula, (12) formula:
If l is not output layer, then defining the error of l node layer j according to the error back propagation of BP networkOutput to node jRate of change δ be l+1 layers of n(l+1)The sum of the change rate that a node error exports it.That
The formula of δ are as follows:
And
By (14), (16) formula, when can calculate error and inversely propagating, each node of output layer and each node of hidden layer change
Variability δ, (13) formula give the δ value calculating method of each node of output layer, reversely can successively be calculated with (15) formula each implicit
The δ value of all nodes of layer.After the δ value for obtaining each node, so that it may be repaired with the threshold values that (14) formula, (16) formula calculate each node
The amount of changing Δ W, to be modified to threshold values.
Further, it is to be with global tidal station information products that step (3), which establishes sea level height duration Statistical Prediction Model,
Sea level height is carried out using recurrence moving average model(MA model) prediction technique and artificial nerve network model prediction technique in key data source
Duration Predicting Technique research, and the prediction result obtained to two kinds of prediction techniques is tested assessment, Optimization Prediction technology, with
Based on this, sea level height duration Statistical Prediction Model is established, and precision of prediction can satisfy Service assurance demand.
Detailed description of the invention
Fig. 1 is the method for the present invention flow diagram.
Specific embodiment
The present invention is described in detail With reference to embodiment.
(1) prediction of sea level height duration is carried out using ARMA model；
ARMA model of the present invention:
When being predicted with arma modeling, { X is allowedt(t=1,2,3 ..., N) expression time series, { αtIt is zero-mean, side
Difference isNormal white noise process,Indicate the auto-regressive parameter of model, θj(j=1,2 ... p) indicate mould
The sliding average parameter of type, B expression move back difference operator.Then ARMA (p.q) model is as follows
Above-mentioned model: model identification, parameter Estimation and model testing is established by three steps, model identification refers to judgement prediction
Model is AR, MA or ARMA, then determines model order, that is, determines p and q, parameter Estimation refers to after identification model, passes through
Suitable method calculates the unknown parameter in (1), that is, determinesθjWith
1. determining the order of model
By sample autocorrelation coefficient and the form of sample partial correlation coefficient come identification model classification, to time series data
After carrying out smoothing preprocessing, calculating original series auto-correlation function ACF, deviation―related function PACF, for stationary time first
Sequence Xt(t=1,2,3 ..., N), specific judgement are as follows:
Mean value are as follows:
Variance are as follows:
Covariance are as follows:
Correlation function are as follows:
Deviation―related function are as follows:
K=1,2 ...；I=1,2 ... k-1.
The order of model is by judging that the truncation of sum primarily determines；
2. parameter Estimation
The method of common model parameter estimation has: the square of maximal possibility estimation, minimum variance estimate, model parameter is estimated
Meter, Least Square Method, Maximum entropy estimation etc., Least Square Method as most common optimal estimation algorithm, relative to
Moments estimation, it belongs to one of essence estimation, and the present invention selects least-squares algorithm to carry out parameter Estimation.
3. model testing
It is tested first with the model being fitted, while calculating residual sequence, if the residual sequence of model is white
Noise sequence, then model is effectively, otherwise to need to increase model order, then re-start parameter Estimation and model testing,
Until model is effective.
(2) prediction of sea level height duration is carried out using artificial nerve network model；
The present inventor's artificial neural networks model:
Assuming that there is N number of training sample (Xk, Yk*), k=1,2 ..., N.To some sample (Xk, Yk* for), net is first passed through
Network model forward-propagating, if Xk=(X1k, X2k..., Xnk) be sample K input node, successively pass through input layer, hidden layer by
Layer processing, is finally exported by output layer, show that the BP network training output of training sample K is Yk=(Y1k, Y2k..., Ynk), sample
The network of this K exports YkY is exported with expectationk* the difference between, as network error, then, by error amount from output layer to defeated
Enter layer inversely to propagate, in reverse communication process, constantly corrects the threshold values of each layer neuron；
Error is inversely propagated in calculating process, if
Input vector XkThe input of l node layer j is traveled to after input；
The output of l node layer j；
The threshold values of l-1 layers of node i connection l node layer j；
n(l-1): l-1 layers of node number；
F: excitation function；
For the neuron excitation function of BP network usually using Sigmoid type function, the input according to BP network neural member is defeated
Relationship out has:
Neuron j exports the network query function of sample KWith sample K to the desired output of neuron jError
Are as follows:
If the l layers of output layer for BP network, i.e. neuron j is output node, then
Error between the calculating output and expectation output of sample K is:
If in N number of training sample, the output error of the m output node of any sample K within the scope of defined,
HaveJ=1,2 ..., m, then training process leaves it at that；Otherwise, continue inversely to propagate error, by error
Negative gradient constantly is corrected threshold values W, it may be assumed that
In formula, η is training rate, 0 < η < 1；
By (5) formula, (6) formula, (7) formula is obtained:
Wherein,
In order to obtainCalculation formula, discuss in two kinds of situation below.
(1) it if l layers are output layer, then can be obtained by (8) formula:
(13) formula can be obtained by (9) formula, (10) formula:
(14) formula can be obtained by (9) formula, (10) formula, (12) formula:
If l is not output layer, then defining the error of l node layer j according to the error back propagation of BP networkOutput to node jRate of change δ be l+1 layers of n(l+1)The sum of the change rate that a node error exports it.That δ
Formula are as follows:
And
By (14), (16) formula, when can calculate error and inversely propagating, each node of output layer and each node of hidden layer change
Variability δ, (13) formula give the δ value calculating method of each node of output layer, reversely can successively be calculated with (15) formula each implicit
The δ value of all nodes of layer.After the δ value for obtaining each node, so that it may be repaired with the threshold values that (14) formula, (16) formula calculate each node
The amount of changing Δ W, to be modified to threshold values.
(3) sea level height duration Statistical Prediction Model is established；
As shown in Figure 1, utilizing recurrence moving average model(MA model) prediction side with global tidal station information products for main data source
Method and artificial nerve network model prediction technique carry out the Predicting Technique research of sea level height duration, and obtain to two kinds of prediction techniques
To prediction result test assessment, Optimization Prediction technology establishes sea level height duration statistical forecast mould based on this
Type, and precision of prediction can satisfy Service assurance demand.
The above is only not to make limit in any form to the present invention to better embodiment of the invention
System, any simple modification that embodiment of above is made according to the technical essence of the invention, equivalent variations and modification,
Belong in the range of technical solution of the present invention.
Claims (4)
1. a kind of sea level height duration prediction algorithm, it is characterised in that follow the steps below:
(1) prediction of sea level height duration is carried out using ARMA model；
(2) prediction of sea level height duration is carried out using artificial nerve network model；
(3) sea level height duration Statistical Prediction Model is established.
2. according to a kind of sea level height duration prediction algorithm described in claim 1, it is characterised in that: in the step (1) certainly
Return moving average model(MA model):
When being predicted with arma modeling, { X is allowedt(t=1,2,3 ..., N) expression time series, { αtIt is that zero-mean, variance areNormal white noise process,Indicate the auto-regressive parameter of model, θj(j=1,2 ... p) indicate the cunning of model
Dynamic mean parameter, B expression move back difference operator, then ARMA (p.q) model is as follows
Above-mentioned model: model identification, parameter Estimation and model testing is established by three steps, model identification refers to judgement prediction model
It is AR, MA or ARMA, then determines model order, that is, determines p and q, parameter Estimation refers to after identification model, by suitable
Method calculate the unknown parameter in (1), that is, determineθjWith
1. determining the order of model
By sample autocorrelation coefficient and the form of sample partial correlation coefficient come identification model classification, time series data is carried out
After smoothing preprocessing, calculating original series auto-correlation function ACF, deviation―related function PACF, for stationary time series first
Xt(t=1,2,3 ..., N), specific judgement are as follows:
Mean value are as follows:
Variance are as follows:
Covariance are as follows:
Correlation function are as follows:
Deviation―related function are as follows:
K=1,2 ...；I=1,2 ... k-1.
The order of model is by judging that the truncation of sum primarily determines；
2. parameter Estimation
Least-squares algorithm is selected to carry out parameter Estimation；
3. model testing
It is tested first with the model being fitted, while calculating residual sequence, if the residual sequence of model is white noise
Sequence, then model is effectively, otherwise to need to increase model order, then re-start parameter Estimation and model testing, until
Model is effective.
3. according to a kind of sea level height duration prediction algorithm described in claim 1, it is characterised in that: people in the step (2)
Artificial neural networks model:
Assuming that there is N number of training sample (Xk, Yk*), k=1,2 ..., N, to some sample (Xk, Yk* for), network model is first passed through
Forward-propagating, if Xk=(X1k, X2k..., Xnk) be sample K input node, successively pass through input layer, hidden layer successively handles,
It is finally exported by output layer, show that the BP network training output of training sample K is Yk=(Y1k, Y2k..., Ynk), the net of sample K
Network exports YkY is exported with expectationk* the difference between, as network error, it is then, error amount is reverse from output layer to input layer
It propagates, in reverse communication process, constantly corrects the threshold values of each layer neuron；
Error is inversely propagated in calculating process, if
Input vector XkThe input of l node layer j is traveled to after input；
The output of l node layer j；
The threshold values of l-1 layers of node i connection l node layer j；
n(l-1): l-1 layers of node number；
F: excitation function；
The neuron excitation function of BP network is closed usually using Sigmoid type function according to the input and output of BP network neural member
System, has:
Neuron j exports the network query function of sample KWith sample K to the desired output of neuron jError are as follows:
If the l layers of output layer for BP network, i.e. neuron j is output node, then Sample K
Calculating output expect output between error be:
If in N number of training sample the output error of the m output node of any sample K has within the scope of definedSo training process leaves it at that；Otherwise, continue inversely to propagate error, by error
Negative gradient constantly is corrected threshold values W, it may be assumed that
In formula, η is training rate, 0 < η < 1；
By (5) formula, (6) formula, (7) formula is obtained:
Wherein,
In order to obtainCalculation formula, discuss in two kinds of situation below.
(1) it if l layers are output layer, then can be obtained by (8) formula:
(13) formula can be obtained by (9) formula, (10) formula:
(14) formula can be obtained by (9) formula, (10) formula, (12) formula:
If l is not output layer, then defining the error of l node layer j according to the error back propagation of BP networkIt is right
The output of node jRate of change δ be l+1 layers of n(l+1)The sum of the change rate that a node error exports it.The public affairs of that δ
Formula are as follows:
And
By (14), (16) formula, when can calculate error and inversely propagating, the rate of change of each node of output layer and each node of hidden layer
δ, (13) formula give the δ value calculating method of each node of output layer, reversely can successively calculate each hidden layer with (15) formula
The δ value of all nodes.After the δ value for obtaining each node, so that it may calculate the threshold values modification amount of each node with (14) formula, (16) formula
Δ W, to be modified to threshold values.
4. according to a kind of sea level height duration prediction algorithm described in claim 1, it is characterised in that: the step (4) is established
Sea level height duration Statistical Prediction Model is with global tidal station information products for main data source, utilizes recurrence rolling average
Model prediction method and artificial nerve network model prediction technique carry out the Predicting Technique research of sea level height duration, and to prediction
The prediction result that method obtains is tested assessment, Optimization Prediction technology, based on this, establishes sea level height duration statistics
Prediction model, and precision of prediction can satisfy Service assurance demand.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811593352.7A CN109614742A (en) | 2018-12-25 | 2018-12-25 | A kind of sea level height duration prediction algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811593352.7A CN109614742A (en) | 2018-12-25 | 2018-12-25 | A kind of sea level height duration prediction algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109614742A true CN109614742A (en) | 2019-04-12 |
Family
ID=66012468
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811593352.7A Pending CN109614742A (en) | 2018-12-25 | 2018-12-25 | A kind of sea level height duration prediction algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109614742A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112465203A (en) * | 2020-11-19 | 2021-03-09 | 中国石油大学(华东) | Sea level height intelligent prediction and forecast system based on gate control circulation unit neural network, computer equipment and storage medium |
| CN118155080A (en) * | 2024-05-10 | 2024-06-07 | 国家***北海预报中心((国家***青岛海洋预报台)(国家***青岛海洋环境监测中心站)) | Enteromorpha coverage area prediction method based on exponential regression model |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103903071A (en) * | 2014-04-17 | 2014-07-02 | 上海电机学院 | Wind power forecast combination method and system |
| CN105139079A (en) * | 2015-07-30 | 2015-12-09 | 广州时韵信息科技有限公司 | Tax revenue prediction method and device based on hybrid model |
| CN105158598A (en) * | 2015-08-15 | 2015-12-16 | 国家电网公司 | Fault prediction method suitable for power equipment |
| CN107085750A (en) * | 2017-03-10 | 2017-08-22 | 广东工业大学 | A kind of mixing dynamic fault Forecasting Methodology based on ARMA and ANN |
-
2018
- 2018-12-25 CN CN201811593352.7A patent/CN109614742A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103903071A (en) * | 2014-04-17 | 2014-07-02 | 上海电机学院 | Wind power forecast combination method and system |
| CN105139079A (en) * | 2015-07-30 | 2015-12-09 | 广州时韵信息科技有限公司 | Tax revenue prediction method and device based on hybrid model |
| CN105158598A (en) * | 2015-08-15 | 2015-12-16 | 国家电网公司 | Fault prediction method suitable for power equipment |
| CN107085750A (en) * | 2017-03-10 | 2017-08-22 | 广东工业大学 | A kind of mixing dynamic fault Forecasting Methodology based on ARMA and ANN |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112465203A (en) * | 2020-11-19 | 2021-03-09 | 中国石油大学(华东) | Sea level height intelligent prediction and forecast system based on gate control circulation unit neural network, computer equipment and storage medium |
| CN118155080A (en) * | 2024-05-10 | 2024-06-07 | 国家***北海预报中心((国家***青岛海洋预报台)(国家***青岛海洋环境监测中心站)) | Enteromorpha coverage area prediction method based on exponential regression model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | An edge intelligence empowered flooding process prediction using Internet of things in smart city | |
| Li et al. | Prediction for tourism flow based on LSTM neural network | |
| Kisi et al. | Precipitation forecasting using wavelet-genetic programming and wavelet-neuro-fuzzy conjunction models | |
| CN107153874B (en) | Water quality prediction method and system | |
| Li et al. | Short-term traffic flow prediction using a methodology based on ARIMA and RBF-ANN | |
| Thong et al. | A new approach to multi-variable fuzzy forecasting using picture fuzzy clustering and picture fuzzy rule interpolation method | |
| Üneş et al. | Prediction of millers ferry dam reservoir level in USA using artificial neural network | |
| Taşar et al. | Forecasting of suspended sediment in rivers using artificial neural networks approach | |
| Kaveh et al. | A comparative study of three different learning algorithms applied to ANFIS for predicting daily suspended sediment concentration | |
| CN103730006A (en) | Short-time traffic flow combined forecasting method | |
| CN115688579B (en) | Drainage basin multipoint water level prediction and early warning method based on generation countermeasure network | |
| Behboudian et al. | A long-term prediction of domestic water demand using preprocessing in artificial neural network | |
| CN105760952A (en) | Load prediction method based on Kalman filtering and self-adaptive fuzzy neural network | |
| Remesan et al. | Effect of data time interval on real-time flood forecasting | |
| Awad et al. | Prediction of water demand using artificial neural networks models and statistical model | |
| Wu et al. | The prediction of monthly average solar radiation with TDNN and ARIMA | |
| CN107704426A (en) | Water level prediction method based on extension wavelet-neural network model | |
| Sharifi et al. | Optimum combination of variables for runoff simulation in Amameh watershed using gamma test | |
| CN109614742A (en) | A kind of sea level height duration prediction algorithm | |
| Okuda et al. | Non-parametric prediction interval estimate for uncertainty quantification of the prediction of road pavement deterioration | |
| Shiri et al. | Forecasting daily stream flows using artificial intelligence approaches | |
| Wu et al. | Local and global Bayesian network based model for flood prediction | |
| Kosana et al. | Hybrid wind speed prediction framework using data pre-processing strategy based autoencoder network | |
| Zhou et al. | Prediction and early warning method of inundation process at waterlogging points based on Bayesian model average and data-driven | |
| CN105071961A (en) | Method and device for predicting service quality of Web service |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190412 |