CN109376903A - A kind of PM2.5 concentration value prediction technique based on game neural network - Google Patents
A kind of PM2.5 concentration value prediction technique based on game neural network Download PDFInfo
- Publication number
- CN109376903A CN109376903A CN201811050495.3A CN201811050495A CN109376903A CN 109376903 A CN109376903 A CN 109376903A CN 201811050495 A CN201811050495 A CN 201811050495A CN 109376903 A CN109376903 A CN 109376903A
- Authority
- CN
- China
- Prior art keywords
- data
- layer
- output
- network
- concentration value
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 51
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000004069 differentiation Effects 0.000 claims abstract description 14
- 230000006870 function Effects 0.000 claims description 31
- 238000012549 training Methods 0.000 claims description 25
- 230000008569 process Effects 0.000 claims description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 230000001537 neural effect Effects 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 4
- 210000005036 nerve Anatomy 0.000 claims description 3
- 230000000306 recurrent effect Effects 0.000 claims description 3
- 238000010219 correlation analysis Methods 0.000 abstract description 3
- 230000008859 change Effects 0.000 abstract description 2
- 238000009825 accumulation Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 239000002245 particle Substances 0.000 description 3
- 238000003915 air pollution Methods 0.000 description 2
- 230000002068 genetic effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000013618 particulate matter Substances 0.000 description 1
- 231100000614 poison Toxicity 0.000 description 1
- 230000007096 poisonous effect Effects 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
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
- G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
- G06N3/061—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using biological neurons, e.g. biological neurons connected to an integrated circuit
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Neurology (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Artificial Intelligence (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- General Business, Economics & Management (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of PM2.5 concentration value prediction technique based on game neural network includes the following steps: that step 1, raw data acquisition, initial data include PM2.5 concentration value historical data, PM2.5 concentration value metric history data and meteorological historical data；Step 2 generates analogue data using generation network；Step 3, the true and false that analogue data is judged using differentiation network；Step 4, using game neural network prediction PM2.5 concentration value.The present invention carries out except nonlinear correlation analysis to PM2.5 concentration value historical data, PM2.5 concentration value index of correlation historical data and meteorological historical data, it also introduces a generation network and initial data is mixed to output analogue data with noise, and analogue data is sent into and differentiates that network differentiates, it carries out handing over time iteration and adjustment according to differentiation result, for small data set, it, can be with accurate description PM2.5 concentration value Time Change without goal-selling model.
Description
Technical field
The present invention relates to the electric powder predictions of air particle PM2.5 concentration value, more particularly to one kind is based on game mind
PM2.5 concentration value prediction technique through network.
Background technique
PM2.5 refers to particulate matter of the diameter less than or equal to 2.5 microns in atmosphere, is rich in a large amount of poisonous and harmful substances
And residence time length, conveying distance in an atmosphere is remote, thus the influence to human health and atmosphere quality is bigger,
PM2.5 is exceeded also to bring another influence --- haze weather.Nowadays air pollution has become focus concerned by people,
And in air pollution regulations, PM2.5 concentration value has become the significant Testing index for measuring air quality.Nowadays, root
Grinding with stronger academic significance and application value is had become to the prediction of future time section PM2.5 concentration value according to historical data
Study carefully problem.
To solve the above-mentioned problems, Zhang Yiwen et al. passes through in paper " PM2.5 prediction model neural network based "
Neural network method is selected to carry out the concentration value prediction of PM2.5.Wang Min et al. is in the paper " city based on BP artificial neural network
The prediction of PM2.5 concentration space " in, using BP artificial nerve network model, the space of PM2.5 concentration becomes in forecasting research area air
It is different.Zheng Yi et al. proposes a kind of area based on deepness belief network in paper " the PM2.5 prediction based on deepness belief network "
Domain PM2.5 annual average prediction technique.Yang Yun et al. is proposed in paper " about PM2.5 mass concentration forecasting research in air "
The prediction of PM2.5 concentration value is realized using the prediction technique of the BP neural network of genetic algorithm optimization.Ma Tiancheng et al. exists
In paper " the fuzzy neural network PM2.5 concentration prediction based on modified PSO ", using a kind of the fuzzy of modified PSO optimization
Particle swarm algorithm is merged with fuzzy neural network the changing rule to predict PM2.5 particle concentration by neural network.Poplar
Cloud et al. proposes fuzzy based on T-S in paper " the PM2.5 mass concentration prediction based on T-S model fuzzy neural network "
The PM2.5 mass concentration prediction technique of neural network.Su Yingying et al. is in patent " based on Unscented kalman neural network
PM2.5 concentration prediction method " in, provide a kind of PM2.5 concentration prediction method based on Unscented kalman neural network.
Through document investigation and analysis, the PM2.5 concentration value prediction technique having proposed at present using neural network as core architecture,
Nonlinear regression analysis is carried out to PM2.5 concentration value and other indexs of correlation (such as AQI, PM10, NO2, CO, SO2, O3).
After neural network model includes ANN, DNN, FNN and BPNN etc., and the combination optimization algorithms such as genetic algorithm, random forest optimization
Mixed method.But through literature survey, it is a large amount of that accumulation is required in existing PM2.5 concentration value neural net prediction method
For historical data to training, the size of final precision of prediction and raw sample data also has certain relationship, and for without a large amount of
The small sample set of data accumulation, existing neural net prediction method just lose its advantage.Secondly, existing neural network is pre-
Surveying model is all to define known models to be trained data, i.e. the distribution pattern of learning objective is set, real work master
It to be the design parameter for learning and adjusting the distribution.
Summary of the invention
It is fixed in advance in order to overcome existing PM2.5 concentration value prediction mode that small data set and object module can not be trained to need
The deficiency of justice, the present invention is to PM2.5 concentration value historical data, PM2.5 concentration value index of correlation historical data and meteorological history
Data carry out except nonlinear correlation analysis, also introduce a generation network and initial data is mixed to output simulation number with noise
According to, and analogue data is sent into and differentiates that network differentiates, it carries out handing over time iteration and adjustment according to differentiation result, a kind of needle is provided
To small data set, it is not necessarily to goal-selling model, it can be with accurate description PM2.5 concentration value Time Change based on game nerve
The PM2.5 concentration value prediction technique of network.
The technical solution adopted by the present invention to solve the technical problems is:
A kind of PM2.5 concentration value prediction technique based on game neural network, described method includes following steps:
Step 1, raw data acquisition.Initial data includes that PM2.5 concentration value historical data, PM2.5 concentration value index are gone through
History data and meteorological historical data；
Step 2 generates analogue data using generation network, and process is as follows:
Step 2.1, creation one comprising input layer, hidden layer and output layer three-layer neural network, setting hidden layer and
The node number of the node number of output layer, the hidden layer provides estimated value using empirical equation, and the empirical equation is as follows:
In above formula, a and b are respectively the neuron number of input layer and output layer, and c is the constant between [0,10]；
Step 2.2, the dimension that input layer is set separately, exports layer data, the training letter of hidden layer, articulamentum and output layer
Number, contiguous function and output function, anticipation error minimum value, maximum number of iterations and the learning rate of setting network；
One group of random data is randomly generated as the input layer data for generating model, then by generating model in step 2.3
One group of new PM2.5 prediction data, which is generated, in output layer afterwards is denoted as D (m) as Fake Data；
Step 3, the true and false that analogue data is judged using differentiation network, process are as follows:
Step 3.1, creation one comprising input layer, hidden layer and output layer three-layer neural network, setting hidden layer and
The node number of output layer；The node number of the hidden layer provides estimated value using empirical equation, and the empirical equation is as follows:
In above formula, a and b are respectively the neuron number of input layer and output layer, and c is the constant between [0,10]；
Step 3.2 connects one layer of softmax function after output layer, converts polytypic output numerical value to relatively
Probability；
Wherein, ViIt is the output of classifier prime output unit.I indicates classification index, and total classification number is C, SiIt indicates
Be currentElement index and all elements index and ratio；
Step 3.3 randomly chooses one group of data as Real Data from data set, is denoted as x；
Step 3.4 inputs in differentiation network by D (m) and as x input data, and output valve is one after differentiating network
Number between a 0 to 1, for indicating that input data is the probability of Real Data, real 1, fake 0；
Step 4, using game neural network prediction PM2.5 concentration value, process is as follows:
Step 4.1 will generate the analogue data generated in network and initial data input differentiation neural network, establish game
Neural network and training；
Step 4.2, computational discrimination network losses function:
LD=-((1-y) log (1-D (G (m)))+ylogD (x)) (4)
Y is the type of input data, when input data is Real Data data, y=1, and the first half of loss function formula
Part is 0.D (x) is the output of discrimination model, indicates that input x is the probability of real data, training objective will to differentiate network
The output of output D (x) be intended to 1；
When input data is Fake Data data, y=0, the latter half of loss function formula is that 0, G (m) is to generate
The output of model, training objective at this time will make the output of D (G (m)) be intended to 0；
Step 4.3 calculates generation network losses function:
LG=(1-y) log (1-D (G (m))) is (5)
The training objective for generating network is that the data that G (m) to be made generates have same data distribution with truthful data；
Step 4.4, the loss function for calculating game neural network:
Wherein,The prediction classification for indicating discrimination model, being rounded to prediction probability is 0 or 1, for changing gradient side
To；
Step 4.5 carries out backpropagation according to the error of loss function, adjusts each layer weight of Recognition with Recurrent Neural Network, adjusts
Perfect square formula is as follows:
Adjustment rule are as follows: maximize the discrimination of D, minimize the data distribution of G and real data set；
Step 4.6 judges whether game neural network restrains, when error is less than anticipation error minimum value, algorithmic statement,
Terminate algorithm when reaching maximum number of iterations, the game neural metwork training is completed；
Testing data is input in the game neural network that the training is completed by step 4.7, exports PM2.5 concentration value
Final predicted value.
Further, the PM2.5 concentration value index includes AQI, PM10, NO2, CO, SO2 and O3 concentration.
In the present invention, in the step 3, uses and differentiated that network carries out the analogue data of initial data and generation network
Differentiate, obtains the authenticity for generating network analog data.
In the step 4, by obtaining game neural network to generating network and differentiating that network calculates separately loss function
Total loss function, and according to the weighed value adjusting rule for defining backpropagation via the penalty values that loss function is calculated.
Technical concept of the invention are as follows: to PM2.5 concentration value historical data, the history of PM2.5 concentration value index of correlation
Data (AQI, PM10, NO2、CO、SO2、O3) and meteorological historical data (temperature, relative humidity, air pressure, wind speed, precipitation etc.) number
Except progress nonlinear correlation analysis, also introduces a generation network and initial data is mixed to output analogue data with noise,
And analogue data is sent into and differentiates that network differentiates, it carries out handing over time iteration and adjustment according to differentiation result, one kind is provided and is based on
The PM2.5 concentration value prediction technique of game neural network.
Beneficial effects of the present invention are mainly manifested in: technical solution of the present invention, which can not only be handled accurately, to be had largely
The complete sample collection of historical data accumulation equally can with high-efficiency high-accuracy handle the small sample set of no mass data accumulation, and
And do not need definition known models and data are trained, i.e., without the distribution pattern for presetting learning objective, effectively improves and work as
The precision of prediction and training speed of preceding PM2.5 concentration value, have widened the limitation of neural metwork training, can be realized small sample set
Accurate prediction.
Detailed description of the invention
Fig. 1 is a kind of PM2.5 concentration value prediction technique schematic diagram based on game neural network.
Fig. 2 is the modeling process chart for generating network.
Fig. 3 is the differentiation flow chart for differentiating network.
Fig. 4 is the training flow chart of game neural network
Specific embodiment
The invention will be further described below in conjunction with the accompanying drawings.
Referring to Fig.1~Fig. 4, a kind of PM2.5 concentration value prediction technique based on game neural network, the method includes such as
Lower step:
Step 1, raw data acquisition.Initial data includes PM2.5 concentration value historical data, PM2.5 concentration value index (ratio
Such as AQI, PM10, NO2, CO, SO2, O3) historical data and meteorological historical data；
Step 2 generates analogue data using generation network, and process is as follows:
Step 2.1, creation one comprising input layer, hidden layer and output layer three-layer neural network, setting hidden layer and
The node number of output layer.The node number of the hidden layer provides estimated value using empirical equation, and the empirical equation is as follows:
In above formula, a and b are respectively the neuron number of input layer and output layer, and c is the constant between [0,10]；
Step 2.2, the dimension that input layer is set separately, exports layer data, the training letter of hidden layer, articulamentum and output layer
Number, contiguous function and output function, anticipation error minimum value, maximum number of iterations and the learning rate of setting network；
One group of random data is randomly generated as the input layer data for generating model, then by generating model in step 2.3
One group of new PM2.5 prediction data, which is generated, in output layer afterwards is denoted as D (m) as Fake Data；
Step 3, the true and false that analogue data is judged using differentiation network, process are as follows:
Step 3.1, creation one comprising input layer, hidden layer and output layer three-layer neural network, setting hidden layer and
The node number of the node number of output layer, the hidden layer provides estimated value using empirical equation, and the empirical equation is as follows:
In above formula, a and b are respectively the neuron number of input layer and output layer, and c is the constant between [0,10]；
Step 3.2 connects one layer of softmax function after output layer, converts polytypic output numerical value to relatively
Probability；
Wherein, ViIt is the output of classifier prime output unit.I indicates classification index, and total classification number is C, SiIt indicates
Be currentElement index and all elements index and ratio；
Step 3.3 randomly chooses one group of data as Real Data from data set, is denoted as x；
Step 3.4 inputs in differentiation network by D (m) and as x input data, and output valve is one after differentiating network
Number between a 0 to 1, for indicating that input data is the probability of Real Data, real 1, fake 0；
Step 4, using game neural network prediction PM2.5 concentration value, process is as follows:
Step 4.1 will generate the analogue data generated in network and initial data input differentiation neural network, establish game
Neural network and training；
Step 4.2, computational discrimination network losses function:
LD=-((1-y) log (1-D (G (m)))+ylogD (x)) (4)
Y is the type of input data, when input data is Real Data data, y=1, and the first half of loss function formula
Part is the output that 0, D (x) is discrimination model, indicates that input x is the probability of real data (y=1, representative are real data),
Training objective will make the output for the output D (x) for differentiating network be intended to 1；
When input data is Fake Data data, y=0, the latter half of loss function formula is that 0, G (m) is to generate
The output of model, training objective at this time will make the output of D (G (m)) be intended to 0；
Step 4.3 calculates generation network losses function:
LG=(1-y) log (1-D (G (m))) is (5)
The training objective for generating network is that the data that G (m) to be made generates have same data distribution with truthful data；
Step 4.4, the loss function for calculating game neural network:
Wherein,The prediction classification for indicating discrimination model, being rounded to prediction probability is 0 or 1, for changing gradient side
To threshold value is set as 0.5；
Step 4.5 carries out backpropagation according to the error of loss function, adjusts each layer weight of Recognition with Recurrent Neural Network, adjusts
Perfect square formula is as follows:
Adjustment rule are as follows: maximize the discrimination of D, minimize the data distribution of G and real data set；
Step 4.6 judges whether game neural network restrains, when error is less than anticipation error minimum value, algorithmic statement；
Terminate algorithm when reaching maximum number of iterations, the game neural metwork training is completed；
Testing data is input in the game neural network that the training is completed by step 4.7, exports PM2.5 concentration value
Final predicted value.
Claims (2)
1. a kind of PM2.5 concentration value prediction technique based on game neural network, which is characterized in that the method includes walking as follows
It is rapid:
Step 1, raw data acquisition, initial data include PM2.5 concentration value historical data, PM2.5 concentration value metric history number
According to meteorological historical data；
Step 2 generates analogue data using generation network, and process is as follows:
Step 2.1, creation one three-layer neural network comprising input layer, hidden layer and output layer, set hidden layer and output
The node number of the node number of layer, the hidden layer provides estimated value using empirical equation, and the empirical equation is as follows:
In above formula, a and b are respectively the neuron number of input layer and output layer, and c is the constant between [0,10]；
Step 2.2, the dimension that input layer is set separately, exports layer data, the training function of hidden layer, articulamentum and output layer,
Contiguous function and output function, anticipation error minimum value, maximum number of iterations and the learning rate of setting network；
Step 2.3, be randomly generated one group of random data as generate model input layer data, then after generating model
Output layer generates one group of new PM2.5 prediction data and is denoted as D (m) as FakeData；
Step 3, the true and false that analogue data is judged using differentiation network, process are as follows:
Step 3.1, creation one three-layer neural network comprising input layer, hidden layer and output layer, set hidden layer and output
The node number of layer；The node number of the hidden layer provides estimated value using empirical equation, and the empirical equation is as follows:
In above formula, a and b are respectively the neuron number of input layer and output layer, and c is the constant between [0,10]；
Step 3.2 connects one layer of softmax function after output layer, converts relative probability for polytypic output numerical value；
Wherein, ViIt is the output of classifier prime output unit.I indicates classification index, and total classification number is C, SiIndicate be
The index of currentElement and all elements index and ratio；
Step 3.3 randomly chooses one group of data as Real Data from data set, is denoted as x；
Step 3.4 inputs in differentiation network by D (m) and as x input data, and output valve arrives after differentiating network for one 0
Number between 1, for indicating that input data is the probability of Real Data, real 1, fake 0；
Step 4, using game neural network prediction PM2.5 concentration value, process is as follows:
Step 4.1 will generate the analogue data generated in network and initial data input differentiation neural network, establish game nerve
Network and training；
Step 4.2, computational discrimination network losses function:
LD=-((1-y) log (1-D (G (m)))+ylogD (x)) (4)
Y is the type of input data, when input data is Real Data data, y=1, and the first half of loss function formula
It is 0.D (x) is the output of discrimination model, indicates that input x is the probability of real data, training objective will to differentiate the defeated of network
The output of D (x) is intended to 1 out；
When input data is Fake Data data, y=0, the latter half of loss function formula is that 0, G (m) is to generate model
Output, training objective at this time will make the output of D (G (m)) be intended to 0；
Step 4.3 calculates generation network losses function:
LG=(1-y) log (1-D (G (m))) is (5)
The training objective for generating network is that the data that G (m) to be made generates have same data distribution with truthful data；
Step 4.4, the loss function for calculating game neural network:
Wherein,The prediction classification for indicating discrimination model, being rounded to prediction probability is 0 or 1, for changing gradient direction；
Step 4.5 carries out backpropagation according to the error of loss function, adjusts each layer weight of Recognition with Recurrent Neural Network, adjustment side
Formula is as follows:
Adjustment rule are as follows: maximize the discrimination of D, minimize the data distribution of G and real data set；
Step 4.6 judges whether game neural network restrains, and when error is less than anticipation error minimum value, algorithmic statement is reaching
Terminate algorithm when to maximum number of iterations, the game neural metwork training is completed；
Testing data is input in the game neural network that the training is completed by step 4.7, and output PM2.5 concentration value is most
Whole predicted value.
2. a kind of PM2.5 concentration value prediction technique based on game neural network as described in claim 1, which is characterized in that
The PM2.5 concentration value index includes AQI, PM10, NO2, CO, SO2 and O3 concentration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811050495.3A CN109376903B (en) | 2018-09-10 | 2018-09-10 | PM2.5 concentration value prediction method based on game neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811050495.3A CN109376903B (en) | 2018-09-10 | 2018-09-10 | PM2.5 concentration value prediction method based on game neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109376903A true CN109376903A (en) | 2019-02-22 |
| CN109376903B CN109376903B (en) | 2021-12-17 |
Family
ID=65405373
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811050495.3A Active CN109376903B (en) | 2018-09-10 | 2018-09-10 | PM2.5 concentration value prediction method based on game neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109376903B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110766222A (en) * | 2019-10-22 | 2020-02-07 | 太原科技大学 | Particle swarm parameter optimization and random forest based PM2.5 concentration prediction method |
| CN111737429A (en) * | 2020-06-16 | 2020-10-02 | 平安科技（深圳）有限公司 | Training method, AI interview method and related equipment |
| CN112183872A (en) * | 2020-10-10 | 2021-01-05 | 东北大学 | Blast furnace gas generation amount prediction method combining generation of countermeasure network and neural network |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060075454A1 (en) * | 2004-10-05 | 2006-04-06 | Samsung Electronics Co., Ltd. | Method and apparatus for summarizing moving picture of sports game |
| CN105956691A (en) * | 2016-04-25 | 2016-09-21 | 北京市环境保护监测中心 | Method of calculating PM2.5 background concentration in different orientations at different observation points of prediction area |
| CN106056210A (en) * | 2016-06-07 | 2016-10-26 | 浙江工业大学 | PM 2.5 concentration value prediction method based on hybrid neural network |
| CN107247888A (en) * | 2017-08-14 | 2017-10-13 | 吉林大学 | Sewage disposal water outlet total phosphorus TP flexible measurement methods based on reserve pool network |
| CN107358626A (en) * | 2017-07-17 | 2017-11-17 | 清华大学深圳研究生院 | A kind of method that confrontation network calculations parallax is generated using condition |
| CN107766815A (en) * | 2017-10-18 | 2018-03-06 | 福州大学 | A kind of vision assistant service operation system and method for running |
| CN108268935A (en) * | 2018-01-11 | 2018-07-10 | 浙江工业大学 | A kind of PM2.5 concentration values Forecasting Methodology and system based on sequential Recognition with Recurrent Neural Network |
| CN108334977A (en) * | 2017-12-28 | 2018-07-27 | 鲁东大学 | Water quality prediction method based on deep learning and system |
| CN108426812A (en) * | 2018-04-08 | 2018-08-21 | 浙江工业大学 | A kind of PM2.5 concentration value prediction techniques based on Memory Neural Networks |
| CN108491497A (en) * | 2018-03-20 | 2018-09-04 | 苏州大学 | The medical document creation method of network technology is fought based on production |
-
2018
- 2018-09-10 CN CN201811050495.3A patent/CN109376903B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060075454A1 (en) * | 2004-10-05 | 2006-04-06 | Samsung Electronics Co., Ltd. | Method and apparatus for summarizing moving picture of sports game |
| CN105956691A (en) * | 2016-04-25 | 2016-09-21 | 北京市环境保护监测中心 | Method of calculating PM2.5 background concentration in different orientations at different observation points of prediction area |
| CN106056210A (en) * | 2016-06-07 | 2016-10-26 | 浙江工业大学 | PM 2.5 concentration value prediction method based on hybrid neural network |
| CN107358626A (en) * | 2017-07-17 | 2017-11-17 | 清华大学深圳研究生院 | A kind of method that confrontation network calculations parallax is generated using condition |
| CN107247888A (en) * | 2017-08-14 | 2017-10-13 | 吉林大学 | Sewage disposal water outlet total phosphorus TP flexible measurement methods based on reserve pool network |
| CN107766815A (en) * | 2017-10-18 | 2018-03-06 | 福州大学 | A kind of vision assistant service operation system and method for running |
| CN108334977A (en) * | 2017-12-28 | 2018-07-27 | 鲁东大学 | Water quality prediction method based on deep learning and system |
| CN108268935A (en) * | 2018-01-11 | 2018-07-10 | 浙江工业大学 | A kind of PM2.5 concentration values Forecasting Methodology and system based on sequential Recognition with Recurrent Neural Network |
| CN108491497A (en) * | 2018-03-20 | 2018-09-04 | 苏州大学 | The medical document creation method of network technology is fought based on production |
| CN108426812A (en) * | 2018-04-08 | 2018-08-21 | 浙江工业大学 | A kind of PM2.5 concentration value prediction techniques based on Memory Neural Networks |
Non-Patent Citations (5)
| Title |
|---|
| IAN GOODFELLOW: "Generative Adversarial Networks", 《ADVANCES IN NEURAL》 * |
| MINGLEI FU: "Prediction of particular matter concentrations by developed feed-forward neural network with rolling mechanism and gray model", 《NEURAL COMPUTING & APPLICATIONS》 * |
| 刘婧: "基于3EDSS的京津冀雾霾形成机理及节能减排决策研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅰ辑》 * |
| 洪洋: "深度卷积对抗生成网络综述", 《第18届中国系统仿真技术及其应用学术年会论文集（18TH CCSSTA 2017）》 * |
| 莫凌飞: "基于深度学习的视频预测研究综述", 《智能系统学报》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110766222A (en) * | 2019-10-22 | 2020-02-07 | 太原科技大学 | Particle swarm parameter optimization and random forest based PM2.5 concentration prediction method |
| CN110766222B (en) * | 2019-10-22 | 2023-09-19 | 太原科技大学 | PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest |
| CN111737429A (en) * | 2020-06-16 | 2020-10-02 | 平安科技（深圳）有限公司 | Training method, AI interview method and related equipment |
| CN111737429B (en) * | 2020-06-16 | 2023-11-03 | 平安科技（深圳）有限公司 | Training method, AI interview method and related equipment |
| CN112183872A (en) * | 2020-10-10 | 2021-01-05 | 东北大学 | Blast furnace gas generation amount prediction method combining generation of countermeasure network and neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109376903B (en) | 2021-12-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108426812B (en) | PM2.5 concentration value prediction method based on memory neural network | |
| CN110782093B (en) | PM fusing SSAE deep feature learning and LSTM2.5Hourly concentration prediction method and system | |
| CN108268935B (en) | PM2.5 concentration value prediction method and system based on time sequence recurrent neural network | |
| CN111191835B (en) | IES incomplete data load prediction method and system based on C-GAN migration learning | |
| CN108009674A (en) | Air PM2.5 concentration prediction methods based on CNN and LSTM fused neural networks | |
| CN113554466B (en) | Short-term electricity consumption prediction model construction method, prediction method and device | |
| CN109376903A (en) | A kind of PM2.5 concentration value prediction technique based on game neural network | |
| CN109143408B (en) | Dynamic region combined short-time rainfall forecasting method based on MLP | |
| Ding et al. | Research of neural network algorithm based on factor analysis and cluster analysis | |
| Cacciola et al. | Aspects about air pollution prediction on urban environment | |
| Navarro et al. | Descriptive examples of the limitations of artificial neural networks applied to the analysis of independent stochastic data | |
| CN113901977A (en) | Deep learning-based power consumer electricity stealing identification method and system | |
| CN111860787A (en) | Short-term prediction method and device for coupling directed graph structure flow data containing missing data | |
| Tang et al. | A deep learning model for measuring oxygen content of boiler flue gas | |
| Song et al. | Study on turbidity prediction method of reservoirs based on long short term memory neural network | |
| Feng et al. | A dual-staged attention based conversion-gated long short term memory for multivariable time series prediction | |
| CN115096357A (en) | Indoor environment quality prediction method based on CEEMDAN-PCA-LSTM | |
| CN109920489A (en) | It is a kind of that model and method for building up are hydrocracked based on Lasso-CCF-CNN | |
| CN117370813A (en) | Atmospheric pollution deep learning prediction method based on K line pattern matching algorithm | |
| CN116720743A (en) | Carbon emission measuring and calculating method based on data clustering and machine learning | |
| CN115359197A (en) | Geological curved surface reconstruction method based on spatial autocorrelation neural network | |
| Pan et al. | Air visibility prediction based on multiple models | |
| Jiang et al. | Forecast of air temperature based on BP neural network | |
| CN109670646A (en) | A kind of PM2.5 concentration value prediction technique based on mixing threshold neural network | |
| Shayeghi et al. | A hybrid particle swarm optimization back propagation algorithm for short term load forecasting |
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 | ||
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20190222 Assignee: Hangzhou Youshu Cloud Travel Information Technology Co.,Ltd. Assignor: JIANG University OF TECHNOLOGY Contract record no.: X2023980054817 Denomination of invention: A Game Neural Network Based Method for Predicting PM2.5 Concentration Values Granted publication date: 20211217 License type: Common License Record date: 20240102 Application publication date: 20190222 Assignee: Hangzhou Tianyin Computer System Engineering Co.,Ltd. Assignor: JIANG University OF TECHNOLOGY Contract record no.: X2023980054814 Denomination of invention: A Game Neural Network Based Method for Predicting PM2.5 Concentration Values Granted publication date: 20211217 License type: Common License Record date: 20240102 |
|
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20190222 Assignee: HANGZHOU YONGGUAN NETWORK TECHNOLOGY CO.,LTD. Assignor: JIANG University OF TECHNOLOGY Contract record no.: X2024980000361 Denomination of invention: A Game Neural Network Based Method for Predicting PM2.5 Concentration Values Granted publication date: 20211217 License type: Common License Record date: 20240109 |