CN108520311A - In conjunction with the haze prediction model method for building up and system of SOFM nets and BP neural network - Google Patents
In conjunction with the haze prediction model method for building up and system of SOFM nets and BP neural network Download PDFInfo
- Publication number
- CN108520311A CN108520311A CN201810185534.4A CN201810185534A CN108520311A CN 108520311 A CN108520311 A CN 108520311A CN 201810185534 A CN201810185534 A CN 201810185534A CN 108520311 A CN108520311 A CN 108520311A
- Authority
- CN
- China
- Prior art keywords
- sofm
- neural network
- training
- data
- pattern
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Computational Linguistics (AREA)
- Development Economics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Game Theory and Decision Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (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 the haze prediction model method for building up and system of a kind of combination SOFM nets and BP neural network, first obtain SOFM training data, then SOFM training is input to SOFM nets with data and is trained until reaching preset learning efficiency;Then the AQI values of the future time in each data group under each pattern are obtained respectively;The AQI values of data and the future time after SOFM net training are finally input to BP neural network to be trained, the value of each influence factor corresponding to training latter mode is netted when training using SOFM as inputting, the AQI values of the future time under the pattern are as corresponding output.The haze prediction model that the present invention establishes can accurately predict haze, and relative to the prediction model of BP neural network, precision of prediction higher.
Description
Technical field
The present invention relates to environmental areas, pre- more specifically to the haze of a kind of combination SOFM nets and BP neural network
Survey method for establishing model and system.
Background technology
Haze mainly by sulfur dioxide, nitrogen oxides and pellet this three form, and (air quality refers to AQI
Number) be quantitative description Air Quality zero dimension index.Now, meteorological department AQI data are real-time releases, and unfavorable
In the subsequent Forewarning Measures of arrangement, therefore haze forecast is at environmental monitoring urgent problem to be solved.However, the difficulty of forecast is not
It is only in that the factor that haze forecast is related to is numerous, many influence factors need huge all in real-time change and diffusion
Computing capability and accurate computation model.
Invention content
The technical problem to be solved in the present invention is, is difficult to complete the technological deficiency of the forecast of haze for the prior art,
Provide a kind of the haze prediction model method for building up and system of combination SOFM nets and BP neural network.
Wherein one side, the technical solution adopted by the present invention to solve the technical problems according to the present invention are:Construction one
Kind is comprised the following steps in conjunction with the haze prediction model method for building up of SOFM nets and BP neural network:
S1, SOFM training data are obtained, SOFM training data include multiple data subsamples, each data
Sample includes multiple patterns, and each pattern includes the value of multiple influence factors and AQI corresponding with the influence factor of the pattern
Value;Wherein, the pattern was divided according to one day period, and each pattern represents a period;
S2, it SOFM training is input to SOFM nets with data is trained until reach preset learning efficiency;
The AQI values of future time in each data group of S3, respectively acquisition under each pattern;
S4, the AQI values that SOFM is netted to data and the future time after training are input to BP neural network and instruct
Practice, when training, nets the value of each influence factor corresponding to training latter mode using SOFM as inputting, under the pattern under
The AQI values of one time are as corresponding output;
Wherein, the model that training is completed is for predicting haze;The model that training is completed is for each result
Prediction:Be using the value of the multiple influence factor under the pattern actually obtained and corresponding AQI values as input, it is next
The AQI values at moment are as prediction result.
Further, in the combination SOFM nets of the present invention and the haze prediction model method for building up of BP neural network, institute
It states multiple influence factors and specifically refers to humidity, temperature, wind speed, rainfall, PM10, PM2.5, SO2、NO2、CO、O3Ten classes influence because
Son.
Further, in the combination SOFM nets of the present invention and the haze prediction model method for building up of BP neural network, institute
The number of nodes for stating hidden layer in BP neural network is determined by following formula:
In formula, m is number of nodes, and n is the number of the multiple influence factor, and l is the output node number of BP neural network,
[x] indicates the maximum positive integer no more than x.
Further, in the combination SOFM nets of the present invention and the haze prediction model method for building up of BP neural network, institute
The transforming function transformation function stated in BP neural network employed in hidden layer and output layer is unipolarity Sigmoid functions.
Further, in the combination SOFM nets of the present invention and the haze prediction model method for building up of BP neural network, also
Include the following steps:
Test data are obtained, test data is handled using step S1-S4 to obtain the test result of AQI, will be surveyed
Test result is compared with actual numerical value, the prediction accuracy of judgment models.
According to another aspect of the present invention, the present invention is to solve its technical problem, additionally provide a kind of combination SOFM nets with
The haze prediction model of BP neural network establishes system, including following module:
SOFM data acquisition modules, for obtaining SOFM training data, SOFM training data include multiple numbers
According to subsample, each data subsample include multiple patterns, each pattern include multiple influence factors value and with the pattern
The corresponding AQI values of influence factor;Wherein, the pattern was divided according to one day period, and each pattern represents one
Period;
SOFM training modules are trained for SOFM training to be input to SOFM nets with data until reaching preset
Practise efficiency;
BP data acquisition modules, for obtaining the future time in each data group under each pattern respectively
AQI values;
Model building module, for the AQI values of data and the future time after SOFM net training to be input to BP
Neural network is trained, and when training nets the value of each influence factor corresponding to training latter mode using SOFM as defeated
Enter, the AQI values of the future time under the pattern are as corresponding output;
Wherein, the model that training is completed is for predicting haze;The model that training is completed is for each result
Prediction:Be using the value of the multiple influence factor under the pattern actually obtained and corresponding AQI values as input, it is next
The AQI values at moment are as prediction result.
Further, it is established in system in the combination SOFM nets of the present invention and the haze prediction model of BP neural network, institute
It states multiple influence factors and specifically refers to humidity, temperature, wind speed, rainfall, PM10, PM2.5, SO2、NO2、CO、O3Ten classes influence because
Son.
Further, it is established in system in the combination SOFM nets of the present invention and the haze prediction model of BP neural network, institute
The number of nodes for stating hidden layer in BP neural network is determined by following formula:
In formula, m is number of nodes, and n is the number of the multiple influence factor, and l is the output node number of BP neural network,
[x] indicates the maximum positive integer no more than x.
Further, it is established in system in the combination SOFM nets of the present invention and the haze prediction model of BP neural network, institute
The transforming function transformation function stated in BP neural network employed in hidden layer and output layer is unipolarity Sigmoid functions.
Further, it is established in system in the combination SOFM nets of the present invention and the haze prediction model of BP neural network, also
Including following modules:
Model authentication module is handled to obtain using step S1-S4 for obtaining test data to test data
The test result of AQI test result is compared with actual numerical value, the prediction accuracy of judgment models.
Implement the haze prediction model method for building up and system of the combination SOFM nets and BP neural network of the present invention, foundation
Haze prediction model can accurately predict haze, and relative to the prediction model of BP neural network, precision of prediction is more
It is high.
Description of the drawings
Present invention will be further explained below with reference to the attached drawings and examples, in attached drawing:
Fig. 1 is the built-up pattern topology diagram of the present invention;
Fig. 2 is the flow chart of an embodiment of the haze prediction model method for building up in conjunction with SOFM nets and BP neural network;
Fig. 3 is the comparison diagram using built-up pattern, the prediction result of BP moulds and actual value;
Fig. 4 be in conjunction with SOFM nets and BP neural network haze prediction model establish system an embodiment flow chart.
Specific implementation mode
For a clearer understanding of the technical characteristics, objects and effects of the present invention, now control attached drawing is described in detail
The specific implementation mode of the present invention.
The variation that the prediction model of haze not only will consider polynary impact factor, meet data regularity will also possess
Powerful generalization ability.The specific properties of impact factor based on haze weather, the present invention propose self-organizing feature map god
Through network and the integrated combination forecasting of BP neural network;Self-organizing feature map neural network (below with SOFM nets)
It is intended to find and conclude the inherent nature of haze data, grasps the inherent law in historical sample data;BP neural network passes through
The attribute and rule of all kinds of historical sample datas of self study form certain generalization ability, and energy is more accurately predicted to realize
Power.
An output neuron can represent a pattern class in SOFM nets, according to the origin cause of formation of Wuhan Urban gray haze phenomenon
The conclusion that analysis and research topic are obtained, the present invention was divided into four patterns by one day, respectively with daily 0:00、6:00、12:
00、18:Data carry out clustering to other sample datas centered on 00, thus SOFM nets select one dimensional linear array meaning clear and
It is simple in structure;Training function utilizes overall topological structure that can all realize with forecast function, as shown in Figure 1.
The present embodiment mainly considers humidity, temperature, wind speed, rainfall, PM10, PM2.5, SO2、NO2、CO、O3Ten class shadows
The factor is rung, in order to more accurately predict the AQI values of subsequent time, it is also necessary to use for reference the AQI values of current time, therefore data are defeated
Enter number n=11.The initial weight of SOFM nets determines the training of the cluster accuracy and sample of sample, provides a kind of simple easy
Capable method:T input sample is randomly selected from training set as initial value.Due to having had determined four kinds of patterns, i.e. t
=4, the time can be controlled near four material time points when extracting initial value from number of training (at ± 2 hours, i.e., four
Between section), initialize approached each pattern class of the input space respectively at the very start from training after weight vector in this way.
The input layer data of BP nerve nets is to net the Different categories of samples data classified, therefore n=11 through SOFM;Future time
AQI values are l=1 as tutor's data, then output layer number of nodes;A kind of determination method of node in hidden layer:If (
It is not integer, then takes and be not less thanMinimum positive integer).Transforming function transformation function employed in hidden layer and output layer is monopole
Property Sigmoid functions
The specific prediction technique of combined method and step of the embodiment are as shown in Figure 2.A kind of combination SOFM nets and BP
The haze prediction model method for building up of neural network specifically comprises the following steps:
S1, SOFM training data are obtained, SOFM training data include multiple data subsamples, each data
Sample includes multiple patterns, and each pattern includes the value of multiple influence factors and AQI corresponding with the influence factor of the pattern
Value;Wherein, the pattern was divided according to one day period, and each pattern represents a period.It is in number of modes
4, influence factor is 10, and each pattern corresponds to 11 data, and SOFM is trained using 11 data as a data cell
Obtain the submodel of the pattern.
S2, it SOFM training is input to SOFM nets with data is trained until reach preset learning efficiency;SOFM nets
In learning efficiency initial value α initial value be 0.8, termination condition be α=0.
The AQI values of future time in each data group of S3, respectively acquisition under each pattern;
S4, the AQI values that SOFM is netted to data and the future time after training are input to BP neural network and instruct
Practice, when training, nets the value of each influence factor corresponding to training latter mode using SOFM as inputting, under the pattern under
The AQI values of one time are as corresponding output.When training, BP neural network is that a data cell is instructed with 12 data
Practice, obtains the corresponding submodel of the pattern.
Wherein, the built-up pattern that training is completed is for predicting haze;The built-up pattern that training is completed is for each
The prediction of a result:Be using the value of the multiple influence factor under the pattern actually obtained and corresponding AQI values as
Input, the AQI values of subsequent time are as prediction result.
The number of nodes of hidden layer is determined by following formula in the BP neural network:
In formula, m is number of nodes, and n is the number of the multiple influence factor, and l is the output node number of BP neural network,
[x] indicates the maximum positive integer no more than x.
After training is completed, this implementation can also obtain test data, be carried out to test data using step S1-S4
Processing obtains the test result of AQI, test result is compared with actual numerical value, the prediction accuracy of judgment models.
With the morning 1 on the 12nd of September in 2015:00-7:00 carries out fine-grained prediction, prediction result such as table for test data
Shown in 1.
Table 1 predicts one hour Comparative result
The model that this patent is built, can also effective land productivity other than the AQI values for the latter hour that can be predicted
The AQI values of continuous 7 hours after being predicted with the historical data of 7 hours of past;Now by 201,5/9,/12 0:00-2015/
9/12 6:00 data predict 20,15/,912 7 as historical data:00-13:00 AQI values, specific prediction result such as table
2。
Table 2 continuously predicts 7 hours
In order to embody the difference of built-up pattern and traditional BP neural network model of the invention, the special BP god in Matlab
Identical sample data is trained and is predicted provided with identical parameters through network model, specific comparative analysis the results are shown in Table 3.
3 model of table compares
Secondly, Fig. 3 has also highlighted built-up pattern has superiority in terms of anticipation trend compared with BP models;BP model predictions
Result it is larger relative to actual value variation;And the AQI that Combined model forecast obtains and the variation of practical haze weather become
Gesture is almost the same.
With reference to figure 4, the combination SOFM nets of the present embodiment and the haze prediction model of BP neural network establish system comprising such as
Lower module:SOFM data acquisition modules 41, SOFM training modules 42, BP data acquisition modules 43 and model building module 44.
Wherein, for SOFM data acquisition modules 41 for obtaining SOFM training data, SOFM training data include multiple data
Subsample, each data subsample include multiple patterns, each pattern include multiple influence factors value and with the pattern
The corresponding AQI values of influence factor;Wherein, the pattern was divided according to one day period, when each pattern represents one
Between section;SOFM training modules 42 are used to SOFM training being input to SOFM nets with data and be trained until reaching preset study
Efficiency;BP data acquisition modules 43 for obtaining the future time in each data group under each pattern respectively
AQI values;Model building module 44 is used to the AQI values of data and the future time after SOFM net training being input to BP god
It is trained through network, using the value of each influence factor corresponding to SOFM net training latter modes as input when training,
The AQI values of future time under the pattern are as corresponding output.The model that training is completed is for predicting haze;Training
Prediction of the model of completion for each result:It is with the value of the multiple influence factor under a pattern actually obtaining
And corresponding AQI values, as input, the AQI values of subsequent time are as prediction result.
It is corresponding with above-mentioned method that combination SOFM nets and the haze prediction model of BP neural network of the present invention establishes system,
Specifically refer to the above method.
The embodiment of the present invention is described with above attached drawing, but the invention is not limited in above-mentioned specific
Embodiment, the above mentioned embodiment is only schematical, rather than restrictive, those skilled in the art
Under the inspiration of the present invention, without breaking away from the scope protected by the purposes and claims of the present invention, it can also make very much
Form, all of these belong to the protection of the present invention.
Claims (10)
1. the haze prediction model method for building up of a kind of combination SOFM nets and BP neural network, which is characterized in that include following step
Suddenly:
S1, SOFM training data are obtained, SOFM training data include multiple data subsamples, each data subsample
Including multiple patterns, each pattern includes the value of multiple influence factors and AQI values corresponding with the influence factor of the pattern;Its
In, the pattern was divided according to one day period, and each pattern represents a period;
S2, it SOFM training is input to SOFM nets with data is trained until reach preset learning efficiency;
The AQI values of future time in each data group of S3, respectively acquisition under each pattern;
S4, the AQI values that SOFM is netted to data and the future time after training are input to BP neural network and are trained, and instruct
Using the value of each influence factor corresponding to SOFM net training latter modes as input when practicing, the future time under the pattern
AQI values as corresponding output;
Wherein, the model that training is completed is for predicting haze;Prediction of the model that training is completed for each result:
It is using the value of the multiple influence factor under the pattern actually obtained and corresponding AQI values as input, subsequent time
AQI values as prediction result.
2. the haze prediction model method for building up of combination SOFM nets according to claim 1 and BP neural network, feature
It is, the multiple influence factor specifically refers to humidity, temperature, wind speed, rainfall, PM10, PM2.5, SO2、NO2、CO、O3Ten
Class impact factor.
3. the haze prediction model method for building up of combination SOFM nets according to claim 1 and BP neural network, feature
It is, the number of nodes of hidden layer is determined by following formula in the BP neural network:
In formula, m is number of nodes, and n is the number of the multiple influence factor, and l is the output node number of BP neural network, [x] table
Show the maximum positive integer no more than x.
4. the haze prediction model method for building up of combination SOFM nets according to claim 1 and BP neural network, feature
It is, the transforming function transformation function in the BP neural network employed in hidden layer and output layer is unipolarity Sigmoid functions.
5. the haze prediction model method for building up of combination SOFM nets according to claim 1 and BP neural network, feature
It is, further includes following step:
Test data are obtained, test data is handled using step S1-S4 to obtain the test result of AQI, test is tied
Fruit is compared with actual numerical value, the prediction accuracy of judgment models.
6. a kind of combination SOFM nets and the haze prediction model of BP neural network establish system, which is characterized in that comprising such as lower die
Block:
SOFM data acquisition modules, for obtaining SOFM training data, SOFM training data include multiple data
Sample, each data subsample include multiple patterns, and each pattern includes the value of multiple influence factors and the shadow with the pattern
The corresponding AQI values of the factor of sound;Wherein, the pattern was divided according to one day period, and each pattern represents a time
Section;
SOFM training modules are trained for SOFM training to be input to SOFM nets with data until reaching preset study effect
Rate;
BP data acquisition modules, the AQI for obtaining the future time in each data group under each pattern respectively
Value;
Model building module, for the AQI values of data and the future time after SOFM net training to be input to BP nerves
Network is trained, and is netted the value of each influence factor corresponding to training latter mode when training using SOFM as input, is somebody's turn to do
The AQI values of future time under pattern are as corresponding output;
Wherein, the model that training is completed is for predicting haze;Prediction of the model that training is completed for each result:
It is using the value of the multiple influence factor under the pattern actually obtained and corresponding AQI values as input, subsequent time
AQI values as prediction result.
7. combination SOFM nets according to claim 6 and the haze prediction model of BP neural network establish system, feature
It is, the multiple influence factor specifically refers to humidity, temperature, wind speed, rainfall, PM10, PM2.5, SO2、NO2、CO、O3Ten
Class impact factor.
8. combination SOFM nets according to claim 6 and the haze prediction model of BP neural network establish system, feature
It is, the number of nodes of hidden layer is determined by following formula in the BP neural network:
In formula, m is number of nodes, and n is the number of the multiple influence factor, and l is the output node number of BP neural network, [x] table
Show the maximum positive integer no more than x.
9. combination SOFM nets according to claim 6 and the haze prediction model of BP neural network establish system, feature
It is, the transforming function transformation function in the BP neural network employed in hidden layer and output layer is unipolarity Sigmoid functions.
10. combination SOFM nets according to claim 6 and the haze prediction model of BP neural network establish system, feature
It is, further includes following modules:
Model authentication module handles test data using step S1-S4 to obtain AQI's for obtaining test data
Test result test result is compared with actual numerical value, the prediction accuracy of judgment models.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810185534.4A CN108520311B (en) | 2018-03-07 | 2018-03-07 | Haze prediction model establishing method and system combining SOFM (software on a programmable) network and BP (back propagation) neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810185534.4A CN108520311B (en) | 2018-03-07 | 2018-03-07 | Haze prediction model establishing method and system combining SOFM (software on a programmable) network and BP (back propagation) neural network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108520311A true CN108520311A (en) | 2018-09-11 |
CN108520311B CN108520311B (en) | 2021-05-28 |
Family
ID=63433520
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810185534.4A Active CN108520311B (en) | 2018-03-07 | 2018-03-07 | Haze prediction model establishing method and system combining SOFM (software on a programmable) network and BP (back propagation) neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108520311B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110766219A (en) * | 2019-10-21 | 2020-02-07 | 成都理工大学工程技术学院 | Haze prediction method based on BP neural network |
CN116739394A (en) * | 2023-08-15 | 2023-09-12 | 中科三清科技有限公司 | Atmospheric pollution weather influence evaluation system and evaluation method |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102255965A (en) * | 2011-07-06 | 2011-11-23 | 无锡泛太科技有限公司 | Test platform for multi-sensor fused Internet of things (IOT) based on rough set BP (Back Propagation) neural network |
CN104112357A (en) * | 2014-07-04 | 2014-10-22 | 浙江大学 | City area traffic emergency plan method for severe haze weather |
CN104965558A (en) * | 2015-05-27 | 2015-10-07 | 华北电力大学(保定) | Photovoltaic power generation system maximum power tracking method and apparatus considering the factor of haze |
CN107590549A (en) * | 2016-07-08 | 2018-01-16 | 上海海洋大学 | A kind of Japanese squid winter life group's resource abundance Forecasting Methodology |
-
2018
- 2018-03-07 CN CN201810185534.4A patent/CN108520311B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102255965A (en) * | 2011-07-06 | 2011-11-23 | 无锡泛太科技有限公司 | Test platform for multi-sensor fused Internet of things (IOT) based on rough set BP (Back Propagation) neural network |
CN104112357A (en) * | 2014-07-04 | 2014-10-22 | 浙江大学 | City area traffic emergency plan method for severe haze weather |
CN104965558A (en) * | 2015-05-27 | 2015-10-07 | 华北电力大学(保定) | Photovoltaic power generation system maximum power tracking method and apparatus considering the factor of haze |
CN107590549A (en) * | 2016-07-08 | 2018-01-16 | 上海海洋大学 | A kind of Japanese squid winter life group's resource abundance Forecasting Methodology |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110766219A (en) * | 2019-10-21 | 2020-02-07 | 成都理工大学工程技术学院 | Haze prediction method based on BP neural network |
CN116739394A (en) * | 2023-08-15 | 2023-09-12 | 中科三清科技有限公司 | Atmospheric pollution weather influence evaluation system and evaluation method |
CN116739394B (en) * | 2023-08-15 | 2023-10-27 | 中科三清科技有限公司 | Atmospheric pollution weather influence evaluation system and evaluation method |
Also Published As
Publication number | Publication date |
---|---|
CN108520311B (en) | 2021-05-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109508360B (en) | Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton | |
CN108491970A (en) | A kind of Predict Model of Air Pollutant Density based on RBF neural | |
CN114626512A (en) | High-temperature disaster forecasting method based on directed graph neural network | |
CN112561058B (en) | Short-term photovoltaic power prediction method based on Stacking-integrated learning | |
CN106650784A (en) | Feature clustering comparison-based power prediction method and device for photovoltaic power station | |
CN108985965A (en) | A kind of photovoltaic power interval prediction method of combination neural network and parameter Estimation | |
CN107730031A (en) | A kind of ultra-short term peak load forecasting method and its system | |
CN106855957A (en) | Factory's bus load prediction based on similar day and least square method supporting vector machine | |
CN109711609A (en) | Photovoltaic plant output power predicting method based on wavelet transformation and extreme learning machine | |
CN105139264A (en) | Photovoltaic generation capacity prediction method based on particle swarm algorithm wavelet neural network | |
CN107808212A (en) | Solar energy collecting power forecasting method based on grey neural network | |
CN107657243A (en) | Neutral net Radar range profile's target identification method based on genetic algorithm optimization | |
CN108074004A (en) | A kind of GIS-Geographic Information System short-term load forecasting method based on gridding method | |
CN109376950A (en) | A kind of polynary Load Forecasting based on BP neural network | |
CN113592132B (en) | Rainfall objective forecasting method based on numerical weather forecast and artificial intelligence | |
CN110276472A (en) | A kind of offshore wind farm power ultra-short term prediction method based on LSTM deep learning network | |
CN113011630B (en) | Short-term prediction method for space-time load of big data distribution network area | |
CN107895202A (en) | Short-term wind speed forecasting method based on various visual angles wind speed mode excavation | |
CN108520311A (en) | In conjunction with the haze prediction model method for building up and system of SOFM nets and BP neural network | |
CN112801411A (en) | Network flow prediction method based on generation countermeasure network | |
CN114492941A (en) | Whole-county photovoltaic prediction method based on cluster division and data enhancement | |
CN115830473A (en) | Water quality inversion method and system based on satellite remote sensing and automatic monitoring | |
CN109800517A (en) | Improved reverse modeling method for magnetorheological damper | |
Ao et al. | Hybrid model of air quality prediction using k-means clustering and deep neural network | |
Ashrafi et al. | Prediction of climate change induced temperature rise in regional scale using neural network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |