CN108053071A - Regional air pollutant concentration Forecasting Methodology, terminal and readable storage medium storing program for executing - Google Patents
Regional air pollutant concentration Forecasting Methodology, terminal and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN108053071A CN108053071A CN201711393032.2A CN201711393032A CN108053071A CN 108053071 A CN108053071 A CN 108053071A CN 201711393032 A CN201711393032 A CN 201711393032A CN 108053071 A CN108053071 A CN 108053071A
- Authority
- CN
- China
- Prior art keywords
- concentration data
- pollutant concentration
- prediction
- data
- predicted
- 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
- 238000000034 method Methods 0.000 title claims abstract description 63
- 239000000809 air pollutant Substances 0.000 title claims abstract description 23
- 231100001243 air pollutant Toxicity 0.000 title claims abstract description 23
- 238000003860 storage Methods 0.000 title claims abstract description 16
- 239000003344 environmental pollutant Substances 0.000 claims abstract description 170
- 231100000719 pollutant Toxicity 0.000 claims abstract description 170
- 238000012544 monitoring process Methods 0.000 claims abstract description 106
- 238000003066 decision tree Methods 0.000 claims abstract description 85
- 238000007637 random forest analysis Methods 0.000 claims abstract description 76
- 238000013480 data collection Methods 0.000 claims abstract description 41
- 238000004422 calculation algorithm Methods 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 24
- 230000015654 memory Effects 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 16
- 238000001914 filtration Methods 0.000 claims description 15
- 238000010606 normalization Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 10
- 238000004393 prognosis Methods 0.000 claims description 6
- 230000014509 gene expression Effects 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 4
- 241001269238 Data Species 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims description 2
- 230000001537 neural effect Effects 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 abstract description 18
- 238000012549 training Methods 0.000 description 19
- 238000010586 diagram Methods 0.000 description 12
- 238000012545 processing Methods 0.000 description 9
- 238000003915 air pollution Methods 0.000 description 8
- 230000009466 transformation Effects 0.000 description 6
- 210000005036 nerve Anatomy 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000032823 cell division Effects 0.000 description 2
- 238000001311 chemical methods and process Methods 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000001932 seasonal effect Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 1
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 238000012896 Statistical algorithm Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000011109 contamination Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000012372 quality testing Methods 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
- G01N15/06—Investigating concentration of particle suspensions
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/0004—Gaseous mixtures, e.g. polluted air
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Analytical Chemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Biochemistry (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Medicinal Chemistry (AREA)
- Dispersion Chemistry (AREA)
- Food Science & Technology (AREA)
- Development Economics (AREA)
- Combustion & Propulsion (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (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 embodiment of the present invention provides a kind of regional air pollutant concentration Forecasting Methodology, terminal and computer readable storage medium.The described method includes：Using the per day Historical Pollution object concentration data collection being calculated and pretreated day history meteorological dataset as sample data set, it is trained using Random Forest model, wherein, Random Forest model includes more decision trees, and every decision tree is realized using multilayer feedforward neural network；Determine the prediction meteorological data for the following preset number of days predicted on the day of current time；Prediction meteorological data is pre-processed；According to pretreated prediction meteorological data, the pollutant concentration data of current time same day monitoring, trained Random Forest model is utilized to predict the pollutant concentration data of the following preset number of days in region to be predicted.The embodiment of the present invention improves regional air pollutant concentration precision of prediction, has stronger generalization ability.
Description
Technical field
The present invention relates to technical field of data processing more particularly to a kind of regional air pollutant concentration Forecasting Methodology, ends
End and computer readable storage medium.
Background technology
In recent years, China's most area frequently occurs the air pollution episode of large area, and air pollution problems inherent is tight
The normal life production of the people is influenced again.Common atmosphere pollution, such as PM2.5, PM10, O3 are to human health risk pole
To be serious.The generation of air pollution and variation tendency are predicted using reliable, easy air prediction methods, make portion of government
Door understands following Air Quality in time, takes counter-measure in advance in face of heavy air contamination accident, ensures the life of the people
Health realizes that maximum social benefit is particularly important with lowest economic cost.
At present, the main method of air pollution forecasting includes two kinds of statistical fluctuation and numerical forecast.Numerical Prediction Method is led to
It crosses and the physics and chemical process of atmosphere pollution is modeled, realize dirty to following air under different scale, different background
The simulation of dye situation.Need that the physical chemistry transformation mechanism of air pollution is carried out in detail to understand due to Numerical Prediction Method and
Modeling, and modeling and application process are complicated, not only need high performance computer cluster, it is also desirable to high-resolution meteorological data
With pollution sources listings data as supporting, these are resulted in the more difficult reality in the relatively low small and medium-sized cities of some detection levels or region
It is existing.
Method based on statistical fluctuation is according to statistical principle, using specific statistical algorithms from air pollutants
Concentration, meteorological factor and other relevant historical datas find out mutual transformation rule, thus this method need not be to pollution
The physical and chemical process of object is modeled.However the research of existing statistical method and application mainly using multiple linear regression as
Main, multiple linear regression nonlinear fitting ability is poor, and prediction effect precision is not high.
The content of the invention
The embodiment of the present invention provides a kind of regional air pollutant concentration Forecasting Methodology, terminal and computer-readable storage medium
Matter can predict regional air pollutant concentration, improve regional air pollutant concentration precision of prediction, have stronger
Generalization ability.
In a first aspect, an embodiment of the present invention provides a kind of regional air pollutant concentration Forecasting Methodology, this method includes：
The monitoring pollution object concentration data collection that season all monitoring points are corresponded to according to region current time to be predicted calculates
Obtain per day Historical Pollution object concentration data collection；
Determine that region current time to be predicted corresponds to the day history meteorological dataset in season；
Day history meteorological dataset is pre-processed；
Using per day Historical Pollution object concentration data collection and pretreated day history meteorological dataset as sample data
Collection, is trained using Random Forest model, wherein, Random Forest model includes more decision trees, and every decision tree uses
Multilayer feedforward neural network is realized；
Determine the prediction meteorological data for the following preset number of days predicted on the day of current time；
Prediction meteorological data is pre-processed；
According to pretreated prediction meteorological data, the pollutant concentration data of current time same day monitoring, training is utilized
Good Random Forest model predicts the pollutant concentration data of the following preset number of days in region to be predicted.
Second aspect, an embodiment of the present invention provides a kind of terminal, which includes performing above-mentioned first aspect institute
The unit for the method stated.
The third aspect, an embodiment of the present invention provides another terminal, including processor, input equipment, output equipment and
Memory, the processor, input equipment, output equipment and memory are connected with each other, wherein, the memory should for storing
With program code, the processor is arranged to call said program code, performs the method described in above-mentioned first aspect.
Fourth aspect, an embodiment of the present invention provides a kind of computer readable storage medium, the computer storage media
Computer program is stored with, the computer program includes program instruction, and described program instruction makes institute when being executed by a processor
State the method described in the above-mentioned first aspect of processor execution.
The embodiment of the present invention is by by the per day Historical Pollution object concentration data collection being calculated and pretreated day
History meteorological dataset is trained using Random Forest model as sample data set, wherein, Random Forest model includes
More decision trees, every decision tree are realized using multilayer feedforward neural network；Determine that the future of prediction on the day of current time presets
The prediction meteorological data of number of days；Prediction meteorological data is pre-processed；According to it is pretreated prediction meteorological data, it is current when
Between same day monitoring pollutant concentration data, the future for predicting region to be predicted using trained Random Forest model presets day
Several pollutant concentration data.The embodiment of the present invention carries out the pollutant concentration in region using trained Random Forest model
Prediction, improves regional air pollutant concentration precision of prediction, has stronger generalization ability.
Description of the drawings
Technical solution in order to illustrate the embodiments of the present invention more clearly, below will be to needed in embodiment description
Attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is some embodiments of the present invention, general for this field
For logical technical staff, without creative efforts, other attached drawings are can also be obtained according to these attached drawings.
Fig. 1 is a kind of flow diagram of regional air pollutant concentration Forecasting Methodology provided in an embodiment of the present invention；
Fig. 2 is a kind of sub-process schematic diagram of regional air pollutant concentration Forecasting Methodology provided in an embodiment of the present invention；
Fig. 3 is a kind of another sub-process signal of regional air pollutant concentration Forecasting Methodology provided in an embodiment of the present invention
Figure；
Fig. 4 is a kind of another sub-process signal of regional air pollutant concentration Forecasting Methodology provided in an embodiment of the present invention
Figure；
Fig. 5 is a kind of another sub-process for regional air pollutant concentration Forecasting Methodology that another embodiment of the present invention provides
Schematic diagram；
Fig. 6 is a kind of schematic block diagram for terminal that the present invention applies example offer；
Fig. 7 is the schematic block diagram of average calculation unit provided in an embodiment of the present invention；
Fig. 8 is the schematic block diagram of the first pretreatment unit provided in an embodiment of the present invention；
Fig. 9 is the schematic block diagram of predicting unit provided in an embodiment of the present invention；
Figure 10 is the schematic block diagram for the predicting unit that another embodiment of the present invention provides；
Figure 11 is a kind of schematic block diagram for terminal that another embodiment of the present invention provides.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is part of the embodiment of the present invention, instead of all the embodiments.Based on this hair
Embodiment in bright, the every other implementation that those of ordinary skill in the art are obtained without making creative work
Example, belongs to the scope of protection of the invention.
It should be appreciated that ought use in this specification and in the appended claims, term " comprising " and "comprising" instruction
Described feature, entirety, step, operation, the presence of element and/or component, but it is not precluded from one or more of the other feature, whole
Body, step, operation, element, component and/or its presence or addition gathered.
In the specific implementation, the terminal described in the embodiment of the present invention is including but not limited to such as with touch sensitive surface
The mobile phone, laptop computer or tablet computer of (for example, touch-screen display and/or touch tablet) etc it is portable
Equipment.It is to be further understood that in certain embodiments, the equipment is not portable communication device, but it is quick with touching
Feel the desktop computer of surface (for example, touch-screen display and/or touch tablet).
Pollutant concentration refers to the amount of contained pollutant in unit volume.Due to Spatial Difference, the pollutant of different zones
Concentration will be different, and the pollutant concentration of the same area can change according to the difference of seasonal climate, therefore dense to pollutant
It it is facing area, season when degree is monitored and predicts.It, can when carrying out pollutant concentration monitoring and prediction according to region
The otherness with spatial distribution is distributed in drain time to lower different zones, avoids the influence of single monitoring point exception.According to
When season carries out pollutant concentration monitoring and prediction, the influence of the pollutant emission otherness of Various Seasonal can be reduced.Wherein,
Season can be divided according to following rule：Winter is 2 months December to next years, spring is March to May, summer is June to 8
The moon, autumn are 9 to November.Different divisions can be carried out to season with specific reference to different regions.
Fig. 1 is a kind of flow diagram of regional air pollutant concentration Forecasting Methodology provided in an embodiment of the present invention.It should
Method includes S101-S107.
S101 corresponds to the monitoring pollution object concentration data collection of season all monitoring points according to region current time to be predicted
Per day Historical Pollution object concentration data collection is calculated.
Wherein, multiple miniature monitoring points can be disposed in region to be predicted, the plurality of miniature monitoring point there are multiple pollutions
Object monitor, such as PM2.5, PM10, O3 monitor.The plurality of miniature monitoring point can monitor dirt in such as 5 minutes every preset time
Object concentration data is contaminated, therefore has multiple time points in daily can all there are pollutant concentration data.It should be noted that in text
The pollutant concentration data mentioned refer to same pollutant concentration data.
The monitoring pollution object concentration data collection that season all monitoring points are corresponded to according to region current time to be predicted calculates
Obtain per day Historical Pollution object concentration data collection.Specifically, obtain region current time to be predicted and correspond to season all prisons
The monitoring pollution object concentration data collection of measuring point, wherein, generally include the pollutant concentration data monitored for many years；In daily
The data that multiple time points obtain are handled to obtain corresponding pollutant concentration data of multiple time points, by multiple time points pair
The pollutant concentration data answered averagely are obtained per day Historical Pollution object concentration data, compute repeatedly to obtain to be predicted
Region current time corresponds to the dated per day Historical Pollution object concentration data collection of season institute.Current time, which corresponds to season, to be owned
Date refers to current time and current time corresponding date in corresponding season.At regional pollution object concentration data
Reason obtains per day Historical Pollution object concentration data, can reduce the influence of the exceptional value of individual time point in one day.
In one embodiment, as shown in Fig. 2, the data obtained to a time point are handled to obtain time point correspondence
Pollutant concentration data, including S201-S204.
S201, at the exceptional value of the monitoring pollution object concentration data of a time point corresponding all monitoring points
Reason.Specifically, whether judge the monitoring pollution object concentration data of corresponding monitoring point has missing values, extremum；If there is missing values, root
Missing values are substituted according to the average value of the Proximity detection point of corresponding monitoring point；If there is extremum, using corresponding pollutant concentration data
Boundary value replace；If without monitoring point in neighborhood, then reject the corresponding data of corresponding monitoring point current point in time.Wherein, pole
End value can be understood as being more than the value for corresponding to scope, such as assume the concentration range of PM2.5 for [0,500], unit is often stood for microgram
Square rice, then extremum is exactly the value more than 500 and the value less than 0, and the boundary value of corresponding PM2.5 is respectively 0 and 500.
If monitoring, the value of PM2.5 is less than 0, then is replaced with the boundary value 0 of PM2.5；If monitoring, the value of PM2.5 is more than 500, then
It is replaced with the boundary value 500 of PM2.5.Wherein neighborhood is interpreted as all point sets of the distance less than K of the corresponding monitoring point of distance.K
3km can be taken, other can also be taken suitably to be worth.
S202 carries out gridding to region to be predicted and obtains multiple grid cells.
Wherein, grid cell size is configured according to region area size to be predicted, such as could be provided as 1km*1km.
Rule setting can be carried out, irregular setting can also be carried out, is forest as being corresponding with a big panel region in region to be predicted,
The region of forest can be so arranged to a grid cell.
S203 corresponds to the pollutant concentration data of net region using the multiple grid cells of interpolation algorithm calculating.
It is to be understood that monitoring point is not necessarily both provided in each grid cell, therefore to being not provided with monitoring point
Grid cell calculates the pollutant concentration data of corresponding net region using interpolation algorithm, and so each grid cell corresponds to grid
There are pollutant concentration data in region.Wherein, interpolation algorithm is Kriging regression algorithm, or other suitable interpolation are calculated
Method.
S204 calculates the average value that multiple grid cells correspond to the pollutant concentration data of net region, obtains described treat
The corresponding pollutant concentration data of region current point in time of prediction.
The average value that multiple grid cells correspond to the pollutant concentration data of net region is calculated, using the average value as treating
The corresponding pollutant concentration data of region current point in time of prediction.
Embodiment shown in Fig. 2 is carried out by obtaining the monitoring pollution object concentration data of all monitoring points to a time point
Outlier processing, and grid cell division is carried out to region to be predicted, and calculate each grid cell with interpolation algorithm and correspond to
The pollutant concentration data in region, the pollutant concentration data of each grid cell corresponding region to obtaining are averaged, with
The pollutant concentration data at corresponding time point are obtained, so improve the accuracy of the pollutant concentration data at corresponding time point.
S102 determines that region current time to be predicted corresponds to the day history meteorological dataset in season.
Day history meteorological dataset includes but not limited to following meteorological factor：(1) temperature：When daily 2,8 when, 14 when, 20
When temperature and it is daily 2 when, 8 when, 14 when, 20 when alternating temperature, daily mean temperature, daily minimal tcmperature, daily maximum temperature, when 7
The temperature difference of table, the temperature of 700hpa, 850hpa, 925hpa and each layer；(2) air pressure：When daily 2,8 when, 14 when, 20 when air pressure
With daily 2 when, 8 when, 14 when, 20 when transformation, per day air pressure, per day draught head；(3) humidity：When daily 2,8 when, 14
When, 20 when relative humidity, per day relative humidity；(4) wind：Cardinal wind, per day wind speed, 24 it is small when wind speed variable；(5)
It is other：Per day rainfall, per day cloud amount, per day intensity of illumination.Wherein, daily 8 when alternating temperature when referring to the previous day 8
Temperature of temperature when subtracting the same day 8, when daily mean temperature refers to daily 2,8 when, 14 when, 20 when temperature average value,
Air pressure when transformation when daily 8 refers to the previous day 8 subtracts the air pressure during same day 8, and per day draught head was referred in one day
Face highest air pressure subtracts the value of barometric minimum.
S103 pre-processes day history meteorological dataset.
In one embodiment, as shown in figure 3, being pre-processed to day history meteorological dataset, including S301-S302.
S301 is filtered the corresponding characteristic attribute of day history meteorological dataset using correlation coefficient process.
Wherein, the corresponding characteristic attribute of day history meteorological dataset, it can be understood as specific meteorological factor, such as gas when 2
Temperature, daily mean temperature, daily maximum temperature etc..By calculating each characteristic attribute and the dirt for needing to predict in day history meteorological dataset
The related coefficient of object concentration is contaminated, is filtered corresponding characteristic attribute according to related coefficient.Such as Pearson correlation coefficient, with
Related coefficient 0.2 is used as threshold values, and being less than the threshold values, then characteristic attribute filters out, and so realizes the preliminary filtering of characteristic attribute.
It is to be appreciated that some characteristic attributes in day history meteorological dataset the pollutant concentration that needs are predicted is not influenced or
Person influences little, then can will retain without influencing or influencing less corresponding characteristic attribute filtering to pollutant concentration
There is the characteristic attribute of larger impact.
S302 is normalized the day history meteorological dataset after filtering according to normalized formula.
For the day history meteorological dataset after filtering, in order to avoid scale caused by different dimensions differs, it is necessary to will not
The meteorological data obtained with dimension is converted into the meteorological data of identical dimension.It therefore can be to the day history meteorological dataset after filtering
It is normalized, min-max standardized methods may be employed in normalization processing method.Wherein, corresponding standardization
Formula is：Y*=0.1+0.8* (Yi-Ymin)/(Ymax-Ymin), wherein, YiRepresent the value of ith feature attribute, YmaxRepresent the spy
Levy the corresponding maximum of attribute, YminAfter representing that the corresponding minimum value of this feature attribute, Y* represent the normalization of this feature attribute
Value.
S104, using per day Historical Pollution object concentration data collection and pretreated day history meteorological dataset as sample
Data set is trained using Random Forest model, wherein, Random Forest model includes more decision trees, every decision tree
It is realized using multilayer feedforward neural network.
Wherein, random forests algorithm is a kind of Ensemble Learning Algorithms, which is integrated by constructing more decision trees
Learn to improve the accuracy and generalization of prediction, effectively avoid the over-fitting problem of single algorithm, improve to air pollution
The noise of quality testing measured data and the tolerance of exceptional value.Characteristic attribute is randomly selected using the algorithm, avoids feature selecting
Irrationality problem, the final prediction for realizing Regional Atmospheric Pollution object concentration.The prediction model of single tree chooses multilayer feedforward
Neural fusion.Wherein, multilayer feedforward neural network refers to BP neural network, unifies to use BP neural network below.
Specifically, random forests algorithm parameter is set, and number scale P, every decision tree such as random forest have that puts back to take
Number k of characteristic attribute etc. is extracted in the times N of sample data, each sample data；For every decision tree：Using
Bootstrap samplings have N number of sample data generation training of the random selection put back to from the sample data concentration after normalization
Collection；K feature is randomly choosed from the characteristic attribute set of sample data, is carried according to k feature of selection from training subset
Take proper subspace training set；BP neural network is initialized, input feature vector subspace training set carries out the training of single decision tree.
The training of all decision trees is completed in the operation for repeating above every decision tree.
Wherein, the number k for characteristic attribute being extracted in each sample data is generated at random in section [4, m], and m is sample
The dimension of the characteristic attribute collection of notebook data.Wherein, P can be 100 etc..Single decision tree is using BP nerves in Random Forest model
Network algorithm realizes that the input layer number of the BP neural network algorithm is using the parameter k of generation, output layer 1, hidden layer number
According toIt determines, wherein, a values in [1,10]., it is necessary to set BP neural network when initializing BP neural network
Error, maximum supervision number that end condition, each iteration allow etc..Wherein it is possible to maximum iteration is set for BP nerves
The end condition of network, maximum iteration can be 2000 etc..
Above step S101-S104 is the process of Random Forest model training.It is dense by calculating per day Historical Pollution object
Degrees of data collection and the pretreatment to day history meteorological data improve reliability and the accuracy of sample data.Further according to more
Accurate and reliable sample data, is trained using Random Forest model, improves the generalization ability of model.
Complete the training of Random Forest model, it is possible to carry out air pollution using the Random Forest model trained
Object concentration prediction.
S105 determines the prediction meteorological data for the following preset number of days predicted on the day of current time.
Wherein, weather prognosis pattern such as WRF (The Weather Research and may be employed in prediction meteorological data
Forecasting Model) etc. prediction result obtain.WRF patterns are initial and American National air may be employed in boundary condition
The global meteorological data of the Global Forecast System (GFS) of research center (NCAR) and the pre- measured center of U.S. environment (NCEP), landform
USGS 30s whole world landform and MODIS underlying surface grouped datas may be employed with underlying surface data.Predict that meteorological data can also
It is obtained using other weather prognosis patterns.Preset number of days can be 1 day, or 7 days etc..
S106 pre-processes prediction meteorological data.
Specifically, the corresponding characteristic attribute of prediction meteorological data is filtered using correlation coefficient process；According to normalization
The prediction meteorological data after filtering is normalized in processing formula.
It should be noted that it is to predicting that meteorological data is pre-processed and pretreatment is carried out to day history meteorological dataset
It is consistent, in order to obtaining the prediction meteorological data of identical data attribute.Specifically prediction meteorological data is pre-processed
Mode refer to the mode to being pre-processed to day history meteorological dataset, and details are not described herein.
S107 according to pretreated prediction meteorological data, the pollutant concentration data of current time same day monitoring, is utilized
Trained Random Forest model predicts the pollutant concentration data of the following preset number of days in region to be predicted.
In one embodiment, as shown in figure 4, being monitored according to pretreated prediction meteorological data, the current time same day
Pollutant concentration data predict that the pollutant of the following preset number of days in region to be predicted is dense using trained Random Forest model
Degrees of data, i.e. step S107 include S401-S404.
S401 obtains the prediction pollutant concentration data of nearest L days calculated using trained Random Forest model.
Specifically, the actual pollutant concentration data calculated according to the same day corresponding meteorological data and the previous day, i.e., before
The averaged historical pollutant concentration data of one day are predicted the pollutant concentration on the same day with trained Random Forest model, are obtained
To the prediction pollutant concentration data several days following later one day after and one day after.Wherein, it should be noted that carry here
To the same day corresponding meteorological data be by pretreated data.In this way, calculate the prediction pollutant concentration of nearest L days
Data, wherein, daily prediction pollutant concentration data are corresponding with multiple.Wherein, the value of L can be 7, or other
Suitable data.
S402, according to the prediction pollutant concentration data of nearest L days and the pollutant concentration data of actual monitoring, using by mistake
Poor calculation formula calculates the prediction error of every decision tree, wherein, error calculation formula is：
Wherein, MSN represents the prediction error of each tree, and the time span that n expressions are predicted every time, unit is day, such as the same day
The prediction air pollutant concentration three days following, then the value of n is 3.ytiIt is predicted for using Random Forest model within t days
The pollutant concentration i-th day following, yt' calculated for the in t days according to the pollutant concentration data of actual monitoring it is per day
Historical Pollution object concentration data.In nearest L days, daily foreseeable time span n is constant, therefore daily in L days recently
Prediction pollutant concentration data it is related with n.Prediction error daily in n days future of the prediction same day is considered in error formula,
Predict the error that following first day of prediction of the same day is not only allowed in error, it is also considered that following second day of same day prediction to the
The error of n days.So improve calculate prediction error accuracy, can so make subsequently with the relevant calculating of prediction error
As a result it is more accurate.
S403 according to pretreated prediction meteorological data, the pollutant concentration data of current time same day monitoring, is utilized
Trained Random Forest model is predicted, obtains the prediction result of more decision trees.
Worked as according to what the pollutant concentration data of pretreated prediction meteorological data, current time same day monitoring calculated
It averaged historical pollutant concentration data, are predicted, every decision tree is corresponding with using trained Random Forest model
One prediction result, then more decision trees can obtain multiple prediction results.
S404 selects to predict following preset number of days of the prediction result of the small correspondence decision tree of error as region to be predicted
Pollutant concentration data.
In another embodiment, as shown in figure 5, according to pretreated prediction meteorological data, current time same day monitoring
Pollutant concentration data, trained Random Forest model is utilized to predict the pollutant of the following preset number of days in region to be predicted
Concentration data, i.e. step S107 include S501-S504.Fig. 5 and Fig. 4 embodiments difference lies in：Add step S503 and
The difference of S505.
The prediction error of every decision tree is normalized S503, to obtain the weight of every decision tree.Specifically,
The calculation formula of weight is according to equation below：
Wherein, MSN represents the prediction error of each tree, and P represents the number of decision tree, viRepresent the transition factor, riRepresent every
The weight of decision tree.
Region to be predicted is calculated according to the weight of every decision tree, the corresponding prediction result of every decision tree in S505
Following preset number of days pollutant concentration data.Specifically, the weight of every decision tree is multiplied by every decision tree future in advance
If corresponding prediction result obtains pollutant concentration data daily in the following preset number of days in region to be predicted in number of days.
In other embodiments, after corresponding prediction result in every decision tree future preset number of days is calculated, also may be used
The corresponding prediction result weighted average of every decision tree directly is obtained daily corresponding prediction result in following preset number of days.
Step S105-S107 is the flow predicted, can be to the air pollutants of current time future preset number of days
Concentration data is predicted.The meteorological data of prediction is pre-processed, to improve the confidence level of meteorological data and and air
The degree of association between pollutant concentration, then when being predicted, according to more accurately predicting error so that the result of prediction is more accurate
Really, the generalization ability of Random Forest model is further improved.
It should be noted that trained Random Forest model also needs to periodically be trained again, so that new data can
To feed back in the model, the accuracy of model prediction is improved, improves the generalization ability of model.
Fig. 6 is a kind of schematic block diagram of terminal provided in an embodiment of the present invention.As shown in fig. 6, the terminal 60 includes meter
Calculate unit 601, the first determination unit 602, the first pretreatment unit 603, training unit 604, the second determination unit 605, second
Pretreatment unit 606, predicting unit 607.
Computing unit 601, for corresponding to the monitoring pollution of season all monitoring points according to region current time to be predicted
Per day Historical Pollution object concentration data collection is calculated in object concentration data collection.
Wherein, multiple miniature monitoring points can be disposed in region to be predicted, the plurality of miniature monitoring point there are multiple pollutions
Object monitor, such as PM2.5, PM10, O3 monitor.The plurality of miniature monitoring point can monitor dirt in such as 5 minutes every preset time
Object concentration data is contaminated, therefore has multiple time points in daily can all there are pollutant concentration data.It should be noted that in text
The pollutant concentration data mentioned refer to same pollutant concentration data.
In one embodiment, computing unit 601 includes first acquisition unit, average calculation unit.Wherein, first list is obtained
Member, for obtaining the monitoring pollution object concentration data collection that region current time to be predicted corresponds to season all monitoring points, wherein,
Generally include the pollutant concentration data monitored for many years；Average calculation unit, for being obtained to interior multiple time points daily
Data handled to obtain corresponding pollutant concentration data of multiple time points, by corresponding pollutant concentration of multiple time points
Data are averagely obtained per day Historical Pollution object concentration data, compute repeatedly to obtain region current time pair to be predicted
It should the dated per day Historical Pollution object concentration data collection of season institute.Current time corresponds to all dates in season and refers to corresponding to
Current time and current time corresponding date in season.Regional pollution object concentration data is handled to obtain per day go through
History pollutant concentration data, can reduce the influence of the exceptional value of individual time point in one day.
In one embodiment, as shown in fig. 7, average calculation unit 70 includes outlier processing unit 701, gridding unit
702nd, interpolating unit 703, time point computing unit 704.
Outlier processing unit 701, for the monitoring pollution object concentration numbers to a time point corresponding all monitoring points
According to exceptional value handled.
Specifically, whether judge the monitoring pollution object concentration data of corresponding monitoring point has missing values, whether has extremum；If
There are missing values, missing values are substituted according to the average value of the Proximity detection point of corresponding monitoring point；If there is extremum, polluted using corresponding
The boundary value of object concentration data is replaced；If without monitoring point in neighborhood, then reject the corresponding number of corresponding monitoring point current point in time
According to.Wherein, extremum can be understood as being more than the value for corresponding to scope, such as assume the concentration range of PM2.5 for [0,500], unit
For every cubic metre of microgram, then extremum is exactly the value more than 500 and the value less than 0, the boundary value point of corresponding PM2.5
It Wei 0 and 500.If monitoring, the value of PM2.5 is less than 0, then is replaced with the boundary value 0 of PM2.5；If monitor the value of PM2.5
More than 500, then replaced with the boundary value 500 of PM2.5.Wherein neighborhood is interpreted as the distance of the corresponding monitoring point of distance less than K's
All point sets.K can take 3km, other can also be taken suitably to be worth.
Gridding unit 702 obtains multiple grid cells for carrying out gridding to region to be predicted.
Wherein, grid cell size is configured according to region area size to be predicted, such as could be provided as 1km*1km.
Rule setting can be carried out, irregular setting can also be carried out, is forest as being corresponding with a big panel region in region to be predicted,
The region of forest can be so arranged to a grid cell.
Interpolating unit 703, for corresponding to the pollutant concentration of net region using the multiple grid cells of interpolation algorithm calculating
Data.It is to be understood that not necessarily it is both provided with monitoring point, therefore the grid to being not provided with monitoring point in each grid cell
Unit calculates the pollutant concentration data of corresponding net region using interpolation algorithm, and so each grid cell corresponds to net region
There are pollutant concentration data.Wherein, interpolation algorithm is Kriging regression algorithm, or other suitable interpolation algorithms.
Time point computing unit 704, for calculating the pollutant concentration data that multiple grid cells correspond to net region
Average value obtains the corresponding pollutant concentration data of the region current point in time to be predicted.Calculate multiple grid cells pair
The average value of the pollutant concentration data of net region is answered, the average value is corresponding as region current point in time to be predicted
Pollutant concentration data.
Average calculation unit 70 is carried out by obtaining the monitoring pollution object concentration data of all monitoring points to a time point
Outlier processing, and grid cell division is carried out to region to be predicted, and calculate each grid cell with interpolation algorithm and correspond to
The pollutant concentration data in region, the pollutant concentration data of each grid cell corresponding region to obtaining are averaged, with
The pollutant concentration data at corresponding time point are obtained, so improve the accuracy of the pollutant concentration data at corresponding time point.
First determination unit 602, for determining that region current time to be predicted corresponds to the day history meteorological data in season
Collection.
Day history meteorological dataset includes but not limited to following meteorological factor：(1) temperature：When daily 2,8 when, 14 when, 20
When temperature and it is daily 2 when, 8 when, 14 when, 20 when alternating temperature, daily mean temperature, daily minimal tcmperature, daily maximum temperature, when 7
The temperature difference of table, the temperature of 700hpa, 850hpa, 925hpa and each layer；(2) air pressure：When daily 2,8 when, 14 when, 20 when air pressure
With daily 2 when, 8 when, 14 when, 20 when transformation, per day air pressure, per day draught head；(3) humidity：When daily 2,8 when, 14
When, 20 when relative humidity, per day relative humidity；(4) wind：Cardinal wind, per day wind speed, 24 it is small when wind speed variable；(5)
It is other：Per day rainfall, per day cloud amount, per day intensity of illumination.Wherein, daily 8 when alternating temperature when referring to the previous day 8
Temperature of temperature when subtracting the same day 8, when daily mean temperature refers to daily 2,8 when, 14 when, 20 when temperature average value,
Air pressure when transformation when daily 8 refers to the previous day 8 subtracts the air pressure during same day 8, and per day draught head was referred in one day
Face highest air pressure subtracts the value of barometric minimum.
First pretreatment unit 603, for being pre-processed to day history meteorological dataset.
In one embodiment, as shown in figure 8, the first pretreatment unit 603 includes filter element 801, normalization unit
802。
Filter element 801, for being carried out using correlation coefficient process to the corresponding characteristic attribute of day history meteorological dataset
Filter.Wherein, the corresponding characteristic attribute of day history meteorological dataset, it can be understood as specific meteorological factor, such as temperature, day when 2
Temperature on average, daily maximum temperature etc..By calculating each characteristic attribute and the pollutant for needing to predict in day history meteorological dataset
Corresponding characteristic attribute is filtered by the related coefficient of concentration according to related coefficient.Such as Pearson correlation coefficient, with correlation
Coefficient 0.2 is used as threshold values, and being less than the threshold values, then characteristic attribute filters out, and so realizes the preliminary filtering of characteristic attribute.It can be with
With understanding, some characteristic attributes in day history meteorological dataset do not have influence or shadow to the pollutant concentration that needs are predicted
Ring little, then can will without influencing or influencing less corresponding characteristic attribute filtering, retain have to pollutant concentration compared with
The characteristic attribute influenced greatly.
Normalization unit 802, for being returned according to normalized formula to the day history meteorological dataset after filtering
One change is handled.For the day history meteorological dataset after filtering, in order to avoid scale caused by different dimensions differs, it is necessary to will not
The meteorological data obtained with dimension is converted into the meteorological data of identical dimension.It therefore can be to the day history meteorological dataset after filtering
It is normalized, min-max standardized methods may be employed in normalization processing method.Wherein, corresponding standardization
Formula is：Y*=0.1+0.8* (Yi-Ymin)/(Ymax-Ymin), wherein, YiRepresent the value of ith feature attribute, YmaxRepresent the spy
Levy the corresponding maximum of attribute, YminAfter representing that the corresponding minimum value of this feature attribute, Y* represent the normalization of this feature attribute
Value.
Training unit 604, for by per day Historical Pollution object concentration data collection and pretreated day history meteorology number
Sample data set is used as according to collection, is trained using Random Forest model, wherein, Random Forest model includes more decision-makings
Tree, every decision tree are realized using BP neural network.
Specifically, random forests algorithm parameter is set, and number scale P, every decision tree such as random forest have that puts back to take
Number k of characteristic attribute etc. is extracted in the times N of sample data, each sample data；For every decision tree：Using
Bootstrap samplings have N number of sample data generation training of the random selection put back to from the sample data concentration after normalization
Collection；K feature is randomly choosed from the characteristic attribute set of sample data, is carried according to k feature of selection from training subset
Take proper subspace training set；BP neural network is initialized, input feature vector subspace training set carries out the training of single decision tree.
The training of all decision trees is completed in the operation for repeating above every decision tree.
Wherein, the number k for characteristic attribute being extracted in each sample data is generated at random in section [4, m], and m is sample
The dimension of the characteristic attribute collection of notebook data.Wherein, P can be 100 etc..Single decision tree is using BP nerves in Random Forest model
Network algorithm realizes that the input layer number of the BP neural network algorithm is using the parameter k of generation, output layer 1, hidden layer number
According toIt determines, wherein, a values in [1,10]., it is necessary to set BP neural network when initializing BP neural network
Error, maximum supervision number that end condition, each iteration allow etc..Wherein it is possible to maximum iteration is set for BP nerves
The end condition of network, maximum iteration can be 2000 etc..
More than computing unit 601, the first determination unit 602, the first pretreatment unit 603, training unit 604 pass through calculating
Per day Historical Pollution object concentration data collection and the pretreatment to day history meteorological data improve the reliability of sample data
And accuracy.Further according to more accurate and reliable sample data, it is trained using Random Forest model, improves the general of model
Change ability.
Complete the training of Random Forest model, it is possible to carry out air pollution using the Random Forest model trained
Object concentration prediction.
Second determination unit 605, for the prediction meteorological data of the following preset number of days of prediction on the day of determining current time.
Wherein, weather prognosis pattern such as WRF (The Weather Research and may be employed in prediction meteorological data
Forecasting Model) etc. prediction result obtain.Preset number of days can be 1 day, or 7 days etc..
Second pretreatment unit 606, for being pre-processed to prediction meteorological data.Specifically, the second pretreatment unit
606, for being filtered using correlation coefficient process to the corresponding characteristic attribute of prediction meteorological data；According to normalized formula
Prediction meteorological data after filtering is normalized.It should be noted that prediction meteorological data carry out pretreatment and
It is consistent that pretreatment is carried out to day history meteorological dataset, in order to obtaining the prediction meteorological data of identical data attribute.
The mode specifically pre-processed to prediction meteorological data refer to the mode to being pre-processed to day history meteorological dataset,
Details are not described herein.
Predicting unit 607, it is dense for the pollutant according to pretreated prediction meteorological data, current time same day monitoring
Degrees of data predicts the pollutant concentration data of the following preset number of days in region to be predicted using trained Random Forest model.
In one embodiment, as shown in figure 9, predicting unit 607 includes second acquisition unit 901, the first error calculation list
First 902, first prediction result unit 903, the first result determination unit 904.
Second acquisition unit 901, it is dirty for obtaining the prediction of nearest L days calculated using trained Random Forest model
Contaminate object concentration data.Specifically, the actual pollutant concentration number calculated according to the same day corresponding meteorological data and the previous day
According to, the i.e. averaged historical pollutant concentration data of the previous day, the pollutant on the same day is predicted with trained Random Forest model
Concentration obtains the prediction pollutant concentration data several days following later one day after and one day after.Wherein, it is necessary to pay attention to
It is that same day mentioned herein, corresponding meteorological data was by pretreated data.In this way, calculate the prediction of nearest L days
Pollutant concentration data, wherein, daily prediction pollutant concentration data are corresponding with multiple.Wherein, the value of L can be 7,
Can be other suitable data.
First error calculation unit 902, for according to the prediction pollutant concentration data of nearest L days and the dirt of actual monitoring
Object concentration data is contaminated, the prediction error of every decision tree is calculated using error calculation formula, wherein, error calculation formula is：
Wherein, MSN represents the prediction error of each tree, and the time span that n expressions are predicted every time, unit is day, such as the same day
The prediction air pollutant concentration three days following, then the value of n is 3.ytiIt is predicted for using Random Forest model within t days
The pollutant concentration i-th day following, yt' calculated for the in t days according to the pollutant concentration data of actual monitoring it is per day
Historical Pollution object concentration data.In nearest L days, daily foreseeable time span n is constant, therefore daily in L days recently
Prediction pollutant concentration data it is related with n.Prediction error daily in n days future of the prediction same day is considered in error formula,
Predict the error that following first day of prediction of the same day is not only allowed in error, it is also considered that following second day of same day prediction to the
The error of n days.So improve calculate prediction error accuracy, can so make subsequently with the relevant calculating of prediction error
As a result it is more accurate.
First prediction result unit 903, for being monitored according to pretreated prediction meteorological data, the current time same day
Pollutant concentration data are predicted using trained Random Forest model, obtain the prediction result of more decision trees.According to
The averaged historical on the same day that the pollutant concentration data of monitoring calculate on the day of pretreated prediction meteorological data, current time
Pollutant concentration data are predicted that every decision tree is corresponded to there are one prediction result using trained Random Forest model,
So much decision tree can obtain multiple prediction results.
First result determination unit 904, the prediction result of the correspondence decision tree for selecting prediction error small are pre- as treating
Survey the pollutant concentration data of the following preset number of days in region.
In another embodiment, as shown in Figure 10, predicting unit 607 includes the 3rd acquiring unit 101, the second error calculation
Unit 102, weight calculation unit 103, the second prediction result unit 104, the second result determination unit 105.Wherein, shown in Figure 10
Embodiment and embodiment shown in Fig. 9 difference lies in：It adds 103 and second result of weight calculation unit and determines list
The difference of member 105.3rd acquiring unit 101, the second error calculation unit 102, the second prediction result unit 104 refer to Fig. 9
Shown embodiment describes accordingly.
Weight calculation unit 103, for the prediction error of every decision tree to be normalized, to obtain every decision tree
Weight.Specifically, the calculation formula of weight is according to equation below：
Wherein, MSN represents the prediction error of each tree, and P represents the number of decision tree, viRepresent the transition factor, riRepresent every
The weight of decision tree.
Second result determination unit 105, for weight, the corresponding prediction result of every decision tree according to every decision tree
The pollutant concentration data of the following preset number of days in region to be predicted are calculated.Specifically, the weight of every decision tree is multiplied
It is obtained with corresponding prediction result in every decision tree future preset number of days in the following preset number of days in region to be predicted daily
Pollutant concentration data.
In other embodiments, after corresponding prediction result in every decision tree future preset number of days is calculated, also may be used
The corresponding prediction result weighted average of every decision tree directly is obtained daily corresponding prediction result in following preset number of days.
Second determination unit 605, the second pretreatment unit 606, predicting unit 607 can be realized following to current time pre-
If the air pollutant concentration data of number of days are predicted.It is pre-processed by the meteorological data to prediction, to improve meteorology
The confidence level of data and the degree of association between air pollutant concentration, then when being predicted, missed according to more accurately prediction
Difference so that the result of prediction is more accurate, further improves the generalization ability of Random Forest model.
It should be noted that trained Random Forest model also needs to periodically be trained again, so that new data can
To feed back in the model, the accuracy of model prediction is improved, improves the generalization ability of model.
Figure 11 is a kind of schematic block diagram for terminal that another embodiment of the present invention provides.The terminal 110 includes one
At a or multiple input equipment 111, one or more output equipments 112, one or more memories 113 and one or more
Device 114 is managed, above-mentioned input equipment 111, output equipment 112, memory 113 and processor 114 are connected by bus 115.It deposits
Reservoir 113 is for storing computer program, and the computer program includes program instruction, and processor 114 is arranged to call
The program instruction that memory 113 stores.Wherein, processor 114 is used to perform：
The monitoring pollution object concentration data collection that season all monitoring points are corresponded to according to region current time to be predicted calculates
Obtain per day Historical Pollution object concentration data collection；Determine that region current time to be predicted corresponds to the day history meteorology number in season
According to collection；Day history meteorological dataset is pre-processed；By per day Historical Pollution object concentration data collection and pretreated day
History meteorological dataset is trained using Random Forest model as sample data set, wherein, Random Forest model includes
More decision trees, every decision tree are realized using multilayer feedforward neural network；Determine that the future of prediction on the day of current time presets
The prediction meteorological data of number of days；Prediction meteorological data is pre-processed；According to it is pretreated prediction meteorological data, it is current when
Between same day monitoring pollutant concentration data, the future for predicting region to be predicted using trained Random Forest model presets day
Several pollutant concentration data.
In one embodiment, processor 114 corresponds to season all monitorings in execution according to region current time to be predicted
It is specific to perform when per day Historical Pollution object concentration data collection is calculated in the monitoring pollution object concentration data collection of point：
Obtain the monitoring pollution object concentration data collection that region current time to be predicted corresponds to season all monitoring points；To every
The data that multiple time points obtain in it are handled to obtain corresponding pollutant concentration data of multiple time points, by multiple times
The corresponding pollutant concentration data of point are averagely obtained per day Historical Pollution object concentration data, compute repeatedly to treat with acquisition pre-
The region current time of survey corresponds to the dated per day Historical Pollution object concentration data collection of season institute.
In one embodiment, processor 114 is handled to obtain the time in the data that execution obtains a time point
It is specific to perform during the corresponding pollutant concentration data of point：
The exceptional value of the monitoring pollution object concentration data of one time point corresponding all monitoring points is handled；To institute
It states region progress gridding to be predicted and obtains multiple grid cells；Multiple grid cells, which are calculated, using interpolation algorithm corresponds to grid
The pollutant concentration data in region；The average value that multiple grid cells correspond to the pollutant concentration data of net region is calculated, is obtained
To the corresponding pollutant concentration data of the region current point in time to be predicted.
In one embodiment, processor 114 is specific to perform when execution pre-processes day history meteorological dataset：
The corresponding characteristic attribute of day history meteorological dataset is filtered using correlation coefficient process；According to normalized
The day history meteorological dataset after filtering is normalized in formula, wherein, normalized formula is：Y*=0.1+
0.8*(Yi-Ymin)/(Ymax-Ymin), wherein, YiRepresent the value of ith feature attribute, YmaxRepresent that the characteristic attribute is corresponding
Maximum, YminRepresent the corresponding minimum value of the characteristic attribute, Y* represents the value after the characteristic attribute normalization.
In one embodiment, processor 114 works as heavenly prison or jail in execution according to pretreated prediction meteorological data, current time
The pollutant concentration data of survey predict the pollution of the following preset number of days in region to be predicted using trained Random Forest model
It is specific to perform during object concentration data：
Obtain the prediction pollutant concentration data of nearest L days calculated using trained Random Forest model；According to most
The nearly prediction pollutant concentration data of L days and the pollutant concentration data of actual monitoring, every is calculated using error calculation formula
The prediction error of decision tree, wherein, error calculation formula is：
Wherein, the time span that n expressions are predicted every time, ytiThe future predicted for the t days using Random Forest model
The pollutant concentration of i-th day, yt' the per day history that is calculated for the in t days according to the pollutant concentration data of actual monitoring
Pollutant concentration data；According to pretreated prediction meteorological data, the pollutant concentration data of current time same day monitoring, profit
It is predicted with trained Random Forest model, obtains the prediction result of more decision trees；The small correspondence of selection prediction error
Pollutant concentration data of the prediction result of decision tree as the following preset number of days in region to be predicted.
In one embodiment, processor 114 works as heavenly prison or jail in execution according to pretreated prediction meteorological data, current time
The pollutant concentration data of survey predict the pollution of the following preset number of days in region to be predicted using trained Random Forest model
It is specific to perform during object concentration data：
Obtain the prediction pollutant concentration data of nearest L days calculated using trained Random Forest model；According to most
The nearly prediction pollutant concentration data of L days and the pollutant concentration data of actual monitoring calculate the prediction error of every decision tree；
The prediction error of every decision tree is normalized, to obtain the weight of every decision tree；According to pretreated prediction gas
The pollutant concentration data of monitoring, are predicted using trained Random Forest model, are obtained on the day of image data, current time
The prediction result of more decision trees；It is calculated and treats according to the weight of every decision tree, the corresponding prediction result of every decision tree
The pollutant concentration data of the following preset number of days of estimation range.
In one embodiment, processor 114 is additionally operable to perform：Determine the following preset number of days predicted on the day of current time
It predicts meteorological data, is come definite by weather prognosis pattern.
It should be appreciated that in embodiments of the present invention, alleged processor 114 can be central processing unit (Central
Processing Unit, CPU), which can also be other general processors, digital signal processor (Digital
Signal Processor, DSP), application-specific integrated circuit (Application Specific Integrated Circuit,
ASIC), ready-made programmable gate array (Field-Programmable GateArray, FPGA) or other programmable logic devices
Part, discrete gate or transistor logic, discrete hardware components etc..General processor can be microprocessor or the processing
Device can also be any conventional processor etc..
Input equipment 111 may include keyboard, mouse, acoustic input dephonoprojectoscope, touch input unit etc..Output equipment 112
It may include display, display screen, touch-screen etc..
The memory 113 can include read-only memory and random access memory, and to processor 114 provide instruction and
Data.The a part of of memory 113 can also include nonvolatile RAM.For example, memory 113 can also be deposited
Store up the information of device type.
In the specific implementation, input equipment 111, output equipment 112, the processor 114 described in the embodiment of the present invention can
The described realization method of all embodiments of optical finger print recognition methods provided in an embodiment of the present invention is performed, also can perform this
The realization method of the described terminal of inventive embodiments, details are not described herein.
A kind of computer readable storage medium, the computer-readable storage medium are provided in another embodiment of the invention
Matter is stored with computer program, and the computer program includes program instruction, and described program instruction is real when being executed by a processor
Existing following steps：
The monitoring pollution object concentration data collection that season all monitoring points are corresponded to according to region current time to be predicted calculates
Obtain per day Historical Pollution object concentration data collection；Determine that region current time to be predicted corresponds to the day history meteorology number in season
According to collection；Day history meteorological dataset is pre-processed；By per day Historical Pollution object concentration data collection and pretreated day
History meteorological dataset is trained using Random Forest model as sample data set, wherein, Random Forest model includes
More decision trees, every decision tree are realized using multilayer feedforward neural network；Determine that the future of prediction on the day of current time presets
The prediction meteorological data of number of days；Prediction meteorological data is pre-processed；According to it is pretreated prediction meteorological data, it is current when
Between same day monitoring pollutant concentration data, the future for predicting region to be predicted using trained Random Forest model presets day
Several pollutant concentration data.
In one embodiment, the processor corresponds to season all monitorings in execution according to region current time to be predicted
When per day Historical Pollution object concentration data collection is calculated in the monitoring pollution object concentration data collection of point, specific implementation：
Obtain the monitoring pollution object concentration data collection that region current time to be predicted corresponds to season all monitoring points；To every
The data that multiple time points obtain in it are handled to obtain corresponding pollutant concentration data of multiple time points, by multiple times
The corresponding pollutant concentration data of point are averagely obtained per day Historical Pollution object concentration data, compute repeatedly to treat with acquisition pre-
The region current time of survey corresponds to the dated per day Historical Pollution object concentration data collection of season institute.
In one embodiment, the processor is handled to obtain the time in the data that execution obtains a time point
During the corresponding pollutant concentration data of point, specific implementation：
The exceptional value of the monitoring pollution object concentration data of one time point corresponding all monitoring points is handled；To institute
It states region progress gridding to be predicted and obtains multiple grid cells；Multiple grid cells, which are calculated, using interpolation algorithm corresponds to grid
The pollutant concentration data in region；The average value that multiple grid cells correspond to the pollutant concentration data of net region is calculated, is obtained
To the corresponding pollutant concentration data of the region current point in time to be predicted.
In one embodiment, the processor is when execution pre-processes day history meteorological dataset, specific implementation：
The corresponding characteristic attribute of day history meteorological dataset is filtered using correlation coefficient process；According to normalized
The day history meteorological dataset after filtering is normalized in formula, wherein, normalized formula is：Y*=0.1+
0.8*(Yi-Ymin)/(Ymax-Ymin), wherein, YiRepresent the value of ith feature attribute, YmaxRepresent that the characteristic attribute is corresponding
Maximum, YminRepresent the corresponding minimum value of the characteristic attribute, Y* represents the value after the characteristic attribute normalization.
In one embodiment, the processor is on the day of performing according to pretreated prediction meteorological data, current time
The pollutant concentration data of monitoring predict the dirt of the following preset number of days in region to be predicted using trained Random Forest model
When contaminating object concentration data, specific implementation：
Obtain the prediction pollutant concentration data of nearest L days calculated using trained Random Forest model；According to most
The nearly prediction pollutant concentration data of L days and the pollutant concentration data of actual monitoring, every is calculated using error calculation formula
The prediction error of decision tree, wherein, error calculation formula is：
Wherein, the time span that n expressions are predicted every time, ytiThe future predicted for the t days using Random Forest model
The pollutant concentration of i-th day, yt' the per day history that is calculated for the in t days according to the pollutant concentration data of actual monitoring
Pollutant concentration data；According to pretreated prediction meteorological data, the pollutant concentration data of current time same day monitoring, profit
It is predicted with trained Random Forest model, obtains the prediction result of more decision trees；The small correspondence of selection prediction error
Pollutant concentration data of the prediction result of decision tree as the following preset number of days in region to be predicted.
In one embodiment, the processor is on the day of performing according to pretreated prediction meteorological data, current time
The pollutant concentration data of monitoring predict the dirt of the following preset number of days in region to be predicted using trained Random Forest model
It is specific to perform when contaminating object concentration data：
Obtain the prediction pollutant concentration data of nearest L days calculated using trained Random Forest model；According to most
The nearly prediction pollutant concentration data of L days and the pollutant concentration data of actual monitoring calculate the prediction error of every decision tree；
The prediction error of every decision tree is normalized, to obtain the weight of every decision tree；According to pretreated prediction gas
The pollutant concentration data of monitoring, are predicted using trained Random Forest model, are obtained on the day of image data, current time
The prediction result of more decision trees；It is calculated and treats according to the weight of every decision tree, the corresponding prediction result of every decision tree
The pollutant concentration data of the following preset number of days of estimation range.
In one embodiment, the processor also performs：The following preset number of days of prediction is pre- on the day of determining current time
Meteorological data is surveyed, is come definite by weather prognosis pattern.
The computer readable storage medium can be the internal storage unit of the terminal described in foregoing any embodiment, example
Such as the hard disk or memory of terminal.The computer readable storage medium can also be the External memory equipment of the terminal, such as
The plug-in type hard disk being equipped in the terminal, intelligent memory card (SmartMedia Card, SMC), secure digital (Secure
Digital, SD) card etc..Further, the computer readable storage medium can also both include the storage inside of the terminal
Unit also includes External memory equipment.In several embodiments provided herein, it should be understood that disclosed terminal and
Method can be realized by another way.For example, terminal embodiment described above is only schematical, for example, institute
The division of unit is stated, is only a kind of division of logic function, there can be other dividing mode in actual implementation.
It is apparent to those skilled in the art that for convenience of description and succinctly, the end of foregoing description
End and the specific work process of unit, may be referred to the corresponding process in preceding method embodiment, details are not described herein.More than institute
It states, is only the specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, it is any to be familiar with the art
Technical staff in the technical scope disclosed by the present invention, various equivalent modifications or substitutions can be readily occurred in, these modification or
Replacement should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with the protection model of claim
Subject to enclosing.
Claims (10)
1. a kind of regional air pollutant concentration Forecasting Methodology, which is characterized in that the described method includes：
The monitoring pollution object concentration data collection that season all monitoring points are corresponded to according to region current time to be predicted is calculated
Per day Historical Pollution object concentration data collection；
Determine that region current time to be predicted corresponds to the day history meteorological dataset in season；
Day history meteorological dataset is pre-processed；
Using per day Historical Pollution object concentration data collection and pretreated day history meteorological dataset as sample data set, profit
It is trained with Random Forest model, wherein, Random Forest model includes more decision trees, and every decision tree is used before multilayer
Present neural fusion；
Determine the prediction meteorological data for the following preset number of days predicted on the day of current time；
Prediction meteorological data is pre-processed；
According to pretreated prediction meteorological data, the pollutant concentration data of current time same day monitoring, utilization is trained
Random Forest model predicts the pollutant concentration data of the following preset number of days in region to be predicted.
2. according to the method described in claim 1, it is characterized in that, described correspond to season according to region current time to be predicted
Per day Historical Pollution object concentration data collection is calculated in the monitoring pollution object concentration data collection of all monitoring points, including：
Obtain the monitoring pollution object concentration data collection that region current time to be predicted corresponds to season all monitoring points；
The data that interior multiple time points obtain daily are handled to obtain corresponding pollutant concentration data of multiple time points, it will
Multiple time points, corresponding pollutant concentration data were averagely obtained per day Historical Pollution object concentration data, compute repeatedly with
It obtains region current time to be predicted and corresponds to the dated per day Historical Pollution object concentration data collection of season institute.
3. according to the method described in claim 2, it is characterized in that, the data obtained to a time point are handled this
Time point corresponding pollutant concentration data, including：
The exceptional value of the monitoring pollution object concentration data of one time point corresponding all monitoring points is handled；
Gridding is carried out to the region to be predicted and obtains multiple grid cells；
The pollutant concentration data of net region are corresponded to using the multiple grid cells of interpolation algorithm calculating；
The average value that multiple grid cells correspond to the pollutant concentration data of net region is calculated, obtains the region to be predicted
The corresponding pollutant concentration data of current point in time.
4. according to the method described in claim 1, it is characterized in that, described pre-process day history meteorological dataset, bag
It includes：
The corresponding characteristic attribute of day history meteorological dataset is filtered using correlation coefficient process；
The day history meteorological dataset after filtering is normalized according to normalized formula, wherein, at normalization
Managing formula is：Y*=0.1+0.8* (Yi-Ymin)/(Ymax-Ymin), wherein, YiRepresent the value of ith feature attribute, YmaxRepresent institute
State the corresponding maximum of characteristic attribute, YminRepresent the corresponding minimum value of the characteristic attribute, Y* represents the characteristic attribute normalizing
Value after change.
5. according to the method described in claim 1, it is characterized in that, it is described according to pretreated prediction meteorological data, it is current
The pollutant concentration data of monitoring on the day of time, the future for predicting region to be predicted using trained Random Forest model preset
The pollutant concentration data of number of days, including：
Obtain the prediction pollutant concentration data of nearest L days calculated using trained Random Forest model；
It is public using error calculation according to the prediction pollutant concentration data of nearest L days and the pollutant concentration data of actual monitoring
Formula calculates the prediction error of every decision tree, wherein, error calculation formula is：
<mrow>
<mi>M</mi>
<mi>S</mi>
<mi>N</mi>
<mo>=</mo>
<msqrt>
<mrow>
<mfrac>
<mn>1</mn>
<mrow>
<mi>L</mi>
<mo>*</mo>
<mi>n</mi>
</mrow>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>t</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>L</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>y</mi>
<mrow>
<mi>t</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>-</mo>
<msubsup>
<mi>y</mi>
<mi>t</mi>
<mo>&prime;</mo>
</msubsup>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
</msqrt>
</mrow>
Wherein, the time span that n expressions are predicted every time, ytiFor the t days i-th day futures predicted using Random Forest model
Pollutant concentration, yt' the per day Historical Pollution object that is calculated for the in t days according to the pollutant concentration data of actual monitoring
Concentration data；
According to pretreated prediction meteorological data, the pollutant concentration data of current time same day monitoring, utilization is trained
Random Forest model is predicted, obtains the prediction result of more decision trees；
Pollutant of the prediction result for the correspondence decision tree for selecting prediction error small as the following preset number of days in region to be predicted
Concentration data.
6. according to the method described in claim 1, it is characterized in that, it is described according to pretreated prediction meteorological data, it is current
The pollutant concentration data of monitoring on the day of time, the future for predicting region to be predicted using trained Random Forest model preset
The pollutant concentration data of number of days, including：
Obtain the prediction pollutant concentration data of nearest L days calculated using trained Random Forest model；
According to the prediction pollutant concentration data of nearest L days and the pollutant concentration data of actual monitoring, every decision tree is calculated
Prediction error；
The prediction error of every decision tree is normalized, to obtain the weight of every decision tree；
According to pretreated prediction meteorological data, the pollutant concentration data of current time same day monitoring, utilization is trained
Random Forest model is predicted, obtains the prediction result of more decision trees；
The future that region to be predicted is calculated according to the weight of every decision tree, the corresponding prediction result of every decision tree presets
The pollutant concentration data of number of days.
7. according to the method described in claim 1, it is characterized in that, the future of prediction presets day on the day of the definite current time
Several prediction meteorological datas is come definite by weather prognosis pattern.
8. a kind of terminal, which is characterized in that the terminal includes performing as described in claim 1-7 any claims
The unit of method.
9. a kind of terminal, which is characterized in that the processor, defeated including processor, input equipment, output equipment and memory
Enter equipment, output equipment and memory to be connected with each other, wherein, the memory is used to store computer program, the computer
Program includes program instruction, and the processor is arranged to call described program instruction, perform such as any one of claim 1-7
The method.
10. a kind of computer readable storage medium, which is characterized in that the computer storage media is stored with computer program,
The computer program includes program instruction, and described program instruction makes the processor perform such as right when being executed by a processor
It is required that 1-7 any one of them methods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711393032.2A CN108053071A (en) | 2017-12-21 | 2017-12-21 | Regional air pollutant concentration Forecasting Methodology, terminal and readable storage medium storing program for executing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711393032.2A CN108053071A (en) | 2017-12-21 | 2017-12-21 | Regional air pollutant concentration Forecasting Methodology, terminal and readable storage medium storing program for executing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108053071A true CN108053071A (en) | 2018-05-18 |
Family
ID=62131227
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711393032.2A Pending CN108053071A (en) | 2017-12-21 | 2017-12-21 | Regional air pollutant concentration Forecasting Methodology, terminal and readable storage medium storing program for executing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108053071A (en) |
Cited By (37)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108629980A (en) * | 2018-06-27 | 2018-10-09 | 石家庄铁道大学 | A method of improving air quality by controlling the flow of traffic |
| CN108808671A (en) * | 2018-07-03 | 2018-11-13 | 南京信息工程大学 | A kind of short-term wind speed DATA PROCESSING IN ENSEMBLE PREDICTION SYSTEM method of wind power plant |
| CN109034641A (en) * | 2018-08-10 | 2018-12-18 | 中国石油大学(北京) | Defect of pipeline prediction technique and device |
| CN109213839A (en) * | 2018-09-12 | 2019-01-15 | 北京英视睿达科技有限公司 | Hot spot grid pollutant data capture method based on the study of gridding depths of features |
| CN109298136A (en) * | 2018-10-11 | 2019-02-01 | 北京大学 | Air Quality Evaluation method, apparatus, equipment and storage medium |
| CN109543874A (en) * | 2018-10-12 | 2019-03-29 | 南京航空航天大学 | A kind of aerodrome air qualitative forecasting method that combination meteorological condition influences |
| CN109541730A (en) * | 2018-11-23 | 2019-03-29 | 长三角环境气象预报预警中心（上海市环境气象中心） | A kind of method and apparatus of pollutant prediction |
| CN109612895A (en) * | 2019-02-16 | 2019-04-12 | 新疆林科院森林生态研究所 | A kind of air quality monitoring method in urban forests environment |
| CN109934379A (en) * | 2019-01-21 | 2019-06-25 | 平安科技（深圳）有限公司 | Prediction technique, device and the electronic equipment of air pollutant concentration |
| CN109946209A (en) * | 2019-03-13 | 2019-06-28 | 北京英视睿达科技有限公司 | A kind of method and device of atmosphere pollution alarm |
| CN109977185A (en) * | 2019-03-19 | 2019-07-05 | 深圳博沃智慧科技有限公司 | Atmosphere pollution monitoring and managing method and system based on grid |
| CN110531029A (en) * | 2019-08-16 | 2019-12-03 | 北京慧辰资道资讯股份有限公司 | A kind of device based on environmentally friendly Internet of Things big data prediction air quality trend |
| CN110610258A (en) * | 2019-08-20 | 2019-12-24 | 中国地质大学（武汉） | Urban air quality refined estimation method and device fusing multi-source space-time data |
| CN111144625A (en) * | 2019-12-10 | 2020-05-12 | 北京蛙鸣信息科技发展有限公司 | Air quality prediction method and system based on adjacent space data principal component elements |
| CN111310970A (en) * | 2020-01-16 | 2020-06-19 | 集美大学 | Real-time environmental parameter prediction method, terminal equipment and storage medium |
| CN111340288A (en) * | 2020-02-25 | 2020-06-26 | 武汉墨锦创意科技有限公司 | Urban air quality time sequence prediction method considering space-time correlation |
| CN111352926A (en) * | 2018-12-20 | 2020-06-30 | 北京沃东天骏信息技术有限公司 | Data processing method, device, equipment and readable storage medium |
| CN111443163A (en) * | 2020-03-10 | 2020-07-24 | 中国科学院深圳先进技术研究院 | Interpolation method and device for ozone missing data and interpolation equipment |
| CN111507533A (en) * | 2020-04-22 | 2020-08-07 | 中科三清科技有限公司 | Air pollutant concentration prediction method and device, electronic equipment and storage medium |
| CN111754042A (en) * | 2020-06-28 | 2020-10-09 | 成都佳华物链云科技有限公司 | Atmospheric pollutant concentration prediction method and device based on Gaussian regression |
| CN111768038A (en) * | 2020-06-30 | 2020-10-13 | 平安国际智慧城市科技股份有限公司 | Pollutant monitoring method and device, terminal equipment and storage medium |
| CN111832222A (en) * | 2020-06-28 | 2020-10-27 | 成都佳华物链云科技有限公司 | Pollutant concentration prediction model training method, prediction method and device |
| CN112001520A (en) * | 2020-06-05 | 2020-11-27 | 上海眼控科技股份有限公司 | Weather forecasting method and device, computer equipment and storage medium |
| CN112800350A (en) * | 2021-02-04 | 2021-05-14 | 中科三清科技有限公司 | Live-action calendar control generation method and device, electronic equipment, storage medium and live-action calendar |
| CN112950352A (en) * | 2021-02-08 | 2021-06-11 | 北京淇瑀信息科技有限公司 | User screening strategy generation method and device and electronic equipment |
| CN113283630A (en) * | 2021-04-09 | 2021-08-20 | 中科三清科技有限公司 | Air quality prediction method, device, equipment and computer readable storage medium |
| CN113888381A (en) * | 2021-09-24 | 2022-01-04 | 中科三清科技有限公司 | Pollutant concentration forecasting method and device |
| CN114266200A (en) * | 2022-02-24 | 2022-04-01 | 山东大学 | Nitrogen dioxide concentration prediction method and system |
| CN114689478A (en) * | 2022-04-08 | 2022-07-01 | 江苏锦测环保科技有限公司 | Air quality detection device and method |
| CN115096374A (en) * | 2022-08-22 | 2022-09-23 | 中工重科智能装备有限责任公司 | Intelligent dust removal prediction compensation method and system in casting |
| CN115237896A (en) * | 2022-07-12 | 2022-10-25 | 四川大学 | Data preprocessing method and system for forecasting air quality based on deep learning |
| CN115270013A (en) * | 2022-09-22 | 2022-11-01 | 中科三清科技有限公司 | Method and device for evaluating emission reduction measures during activities and electronic equipment |
| CN115423148A (en) * | 2022-07-29 | 2022-12-02 | 江苏大学 | Agricultural machinery operation performance prediction method and device based on kriging method and decision tree |
| CN115526434A (en) * | 2022-11-07 | 2022-12-27 | 广东中思拓大数据研究院有限公司 | Object information prediction method, device, computer equipment and storage medium |
| CN115878695A (en) * | 2023-02-20 | 2023-03-31 | 中国民用航空局空中交通管理局航空气象中心 | Data visualization adjusting method and system based on meteorological database |
| CN117571976A (en) * | 2024-01-15 | 2024-02-20 | 深圳市恒义建筑技术有限公司 | Method for measuring and calculating release rate of harmful substances of interior decoration material and related equipment |
| CN117933466A (en) * | 2024-01-23 | 2024-04-26 | 中科三清科技有限公司 | Pollution prediction method and device, storage medium and electronic equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110212461A1 (en) * | 2010-02-26 | 2011-09-01 | Hans Marcus Ludwig Bitter | Prediction of cardiotoxicity |
| CN104850734A (en) * | 2015-04-21 | 2015-08-19 | 武大吉奥信息技术有限公司 | Air quality index prediction method based on spatial and temporal distribution characteristics |
| CN106019409A (en) * | 2016-05-11 | 2016-10-12 | 北京市环境保护监测中心 | Partition prediction method and system for ozone concentration |
| CN106525678A (en) * | 2016-12-03 | 2017-03-22 | 安徽新华学院 | PM2.5 concentration prediction method and device based on geographic position |
| CN106599520A (en) * | 2016-12-31 | 2017-04-26 | 中国科学技术大学 | LSTM-RNN model-based air pollutant concentration forecast method |
| CN106651100A (en) * | 2016-10-12 | 2017-05-10 | 华南理工大学 | Internet-of-Vehicles optimal vehicle-mounted monitoring point-based air quality evaluation system and method |
-
2017
- 2017-12-21 CN CN201711393032.2A patent/CN108053071A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110212461A1 (en) * | 2010-02-26 | 2011-09-01 | Hans Marcus Ludwig Bitter | Prediction of cardiotoxicity |
| CN104850734A (en) * | 2015-04-21 | 2015-08-19 | 武大吉奥信息技术有限公司 | Air quality index prediction method based on spatial and temporal distribution characteristics |
| CN106019409A (en) * | 2016-05-11 | 2016-10-12 | 北京市环境保护监测中心 | Partition prediction method and system for ozone concentration |
| CN106651100A (en) * | 2016-10-12 | 2017-05-10 | 华南理工大学 | Internet-of-Vehicles optimal vehicle-mounted monitoring point-based air quality evaluation system and method |
| CN106525678A (en) * | 2016-12-03 | 2017-03-22 | 安徽新华学院 | PM2.5 concentration prediction method and device based on geographic position |
| CN106599520A (en) * | 2016-12-31 | 2017-04-26 | 中国科学技术大学 | LSTM-RNN model-based air pollutant concentration forecast method |
Non-Patent Citations (4)
| Title |
|---|
| 侯俊雄 等: "基于随机森林的PM2. 5实时预报系统", 《测绘科学》 * |
| 周丽峰: "基于非平衡数据分类的贷款违约预测研究", 《中国优秀硕士学位论文全文数据库（经济与管理科学辑）》 * |
| 尹华: "面向高维和不平衡数据分类的集成学习研究", 《中国博士学位论文全文数据库（信息科技辑）》 * |
| 林愿仪: "基于PSO-BP神经网络预测广州市日均PM10浓度", 《2015年（第四届）全国大学生统计建模大赛论文》 * |
Cited By (55)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108629980A (en) * | 2018-06-27 | 2018-10-09 | 石家庄铁道大学 | A method of improving air quality by controlling the flow of traffic |
| CN108629980B (en) * | 2018-06-27 | 2020-04-07 | 石家庄铁道大学 | Method for improving air quality by controlling traffic flow |
| CN108808671A (en) * | 2018-07-03 | 2018-11-13 | 南京信息工程大学 | A kind of short-term wind speed DATA PROCESSING IN ENSEMBLE PREDICTION SYSTEM method of wind power plant |
| CN109034641A (en) * | 2018-08-10 | 2018-12-18 | 中国石油大学(北京) | Defect of pipeline prediction technique and device |
| CN109213839A (en) * | 2018-09-12 | 2019-01-15 | 北京英视睿达科技有限公司 | Hot spot grid pollutant data capture method based on the study of gridding depths of features |
| CN109298136B (en) * | 2018-10-11 | 2019-10-15 | 北京大学 | Air Quality Evaluation method, apparatus, equipment and storage medium |
| CN109298136A (en) * | 2018-10-11 | 2019-02-01 | 北京大学 | Air Quality Evaluation method, apparatus, equipment and storage medium |
| CN109543874A (en) * | 2018-10-12 | 2019-03-29 | 南京航空航天大学 | A kind of aerodrome air qualitative forecasting method that combination meteorological condition influences |
| CN109543874B (en) * | 2018-10-12 | 2021-06-22 | 南京航空航天大学 | Airport air quality prediction method combining meteorological condition influence |
| CN109541730A (en) * | 2018-11-23 | 2019-03-29 | 长三角环境气象预报预警中心（上海市环境气象中心） | A kind of method and apparatus of pollutant prediction |
| CN111352926A (en) * | 2018-12-20 | 2020-06-30 | 北京沃东天骏信息技术有限公司 | Data processing method, device, equipment and readable storage medium |
| CN111352926B (en) * | 2018-12-20 | 2024-03-08 | 北京沃东天骏信息技术有限公司 | Method, device, equipment and readable storage medium for data processing |
| CN109934379A (en) * | 2019-01-21 | 2019-06-25 | 平安科技（深圳）有限公司 | Prediction technique, device and the electronic equipment of air pollutant concentration |
| CN109612895A (en) * | 2019-02-16 | 2019-04-12 | 新疆林科院森林生态研究所 | A kind of air quality monitoring method in urban forests environment |
| CN109612895B (en) * | 2019-02-16 | 2021-02-23 | 新疆林科院森林生态研究所 | Air quality monitoring method in urban forest environment |
| CN109946209A (en) * | 2019-03-13 | 2019-06-28 | 北京英视睿达科技有限公司 | A kind of method and device of atmosphere pollution alarm |
| CN109977185A (en) * | 2019-03-19 | 2019-07-05 | 深圳博沃智慧科技有限公司 | Atmosphere pollution monitoring and managing method and system based on grid |
| CN110531029A (en) * | 2019-08-16 | 2019-12-03 | 北京慧辰资道资讯股份有限公司 | A kind of device based on environmentally friendly Internet of Things big data prediction air quality trend |
| CN110531029B (en) * | 2019-08-16 | 2022-02-25 | 北京慧辰资道资讯股份有限公司 | Device for predicting air quality trend based on environmental protection Internet of things big data |
| CN110610258A (en) * | 2019-08-20 | 2019-12-24 | 中国地质大学（武汉） | Urban air quality refined estimation method and device fusing multi-source space-time data |
| CN110610258B (en) * | 2019-08-20 | 2022-05-10 | 中国地质大学（武汉） | Urban air quality refined estimation method and device fusing multi-source space-time data |
| CN111144625A (en) * | 2019-12-10 | 2020-05-12 | 北京蛙鸣信息科技发展有限公司 | Air quality prediction method and system based on adjacent space data principal component elements |
| CN111310970A (en) * | 2020-01-16 | 2020-06-19 | 集美大学 | Real-time environmental parameter prediction method, terminal equipment and storage medium |
| CN111340288B (en) * | 2020-02-25 | 2024-04-05 | 武汉墨锦创意科技有限公司 | Urban air quality time sequence prediction method considering time-space correlation |
| CN111340288A (en) * | 2020-02-25 | 2020-06-26 | 武汉墨锦创意科技有限公司 | Urban air quality time sequence prediction method considering space-time correlation |
| CN111443163A (en) * | 2020-03-10 | 2020-07-24 | 中国科学院深圳先进技术研究院 | Interpolation method and device for ozone missing data and interpolation equipment |
| WO2021179742A1 (en) * | 2020-03-10 | 2021-09-16 | 中国科学院深圳先进技术研究院 | Ozone missing data interpolation method, apparatus and device |
| CN111507533B (en) * | 2020-04-22 | 2021-03-19 | 中科三清科技有限公司 | Air pollutant concentration prediction method and device, electronic equipment and storage medium |
| CN111507533A (en) * | 2020-04-22 | 2020-08-07 | 中科三清科技有限公司 | Air pollutant concentration prediction method and device, electronic equipment and storage medium |
| CN112001520A (en) * | 2020-06-05 | 2020-11-27 | 上海眼控科技股份有限公司 | Weather forecasting method and device, computer equipment and storage medium |
| CN111832222A (en) * | 2020-06-28 | 2020-10-27 | 成都佳华物链云科技有限公司 | Pollutant concentration prediction model training method, prediction method and device |
| CN111832222B (en) * | 2020-06-28 | 2023-07-25 | 成都佳华物链云科技有限公司 | Pollutant concentration prediction model training method, pollutant concentration prediction method and pollutant concentration prediction device |
| CN111754042A (en) * | 2020-06-28 | 2020-10-09 | 成都佳华物链云科技有限公司 | Atmospheric pollutant concentration prediction method and device based on Gaussian regression |
| CN111768038A (en) * | 2020-06-30 | 2020-10-13 | 平安国际智慧城市科技股份有限公司 | Pollutant monitoring method and device, terminal equipment and storage medium |
| CN111768038B (en) * | 2020-06-30 | 2023-02-03 | 平安国际智慧城市科技股份有限公司 | Pollutant monitoring method and device, terminal equipment and storage medium |
| CN112800350A (en) * | 2021-02-04 | 2021-05-14 | 中科三清科技有限公司 | Live-action calendar control generation method and device, electronic equipment, storage medium and live-action calendar |
| CN112950352A (en) * | 2021-02-08 | 2021-06-11 | 北京淇瑀信息科技有限公司 | User screening strategy generation method and device and electronic equipment |
| CN113283630A (en) * | 2021-04-09 | 2021-08-20 | 中科三清科技有限公司 | Air quality prediction method, device, equipment and computer readable storage medium |
| CN113888381A (en) * | 2021-09-24 | 2022-01-04 | 中科三清科技有限公司 | Pollutant concentration forecasting method and device |
| CN113888381B (en) * | 2021-09-24 | 2023-08-29 | 中科三清科技有限公司 | Pollutant Concentration Forecasting Method and Device |
| CN114266200B (en) * | 2022-02-24 | 2022-07-05 | 山东大学 | Nitrogen dioxide concentration prediction method and system |
| CN114266200A (en) * | 2022-02-24 | 2022-04-01 | 山东大学 | Nitrogen dioxide concentration prediction method and system |
| CN114689478A (en) * | 2022-04-08 | 2022-07-01 | 江苏锦测环保科技有限公司 | Air quality detection device and method |
| CN115237896A (en) * | 2022-07-12 | 2022-10-25 | 四川大学 | Data preprocessing method and system for forecasting air quality based on deep learning |
| CN115237896B (en) * | 2022-07-12 | 2023-07-11 | 四川大学 | Data preprocessing method and system based on deep learning forecast air quality |
| CN115423148A (en) * | 2022-07-29 | 2022-12-02 | 江苏大学 | Agricultural machinery operation performance prediction method and device based on kriging method and decision tree |
| CN115423148B (en) * | 2022-07-29 | 2023-10-31 | 江苏大学 | Agricultural machinery operation performance prediction method and device based on Ke Li jin method and decision tree |
| CN115096374B (en) * | 2022-08-22 | 2022-11-11 | 中工重科智能装备有限责任公司 | Intelligent dust removal prediction compensation method and system in casting |
| CN115096374A (en) * | 2022-08-22 | 2022-09-23 | 中工重科智能装备有限责任公司 | Intelligent dust removal prediction compensation method and system in casting |
| CN115270013A (en) * | 2022-09-22 | 2022-11-01 | 中科三清科技有限公司 | Method and device for evaluating emission reduction measures during activities and electronic equipment |
| CN115270013B (en) * | 2022-09-22 | 2022-12-20 | 中科三清科技有限公司 | Method and device for evaluating emission reduction measures during activity and electronic equipment |
| CN115526434A (en) * | 2022-11-07 | 2022-12-27 | 广东中思拓大数据研究院有限公司 | Object information prediction method, device, computer equipment and storage medium |
| CN115878695A (en) * | 2023-02-20 | 2023-03-31 | 中国民用航空局空中交通管理局航空气象中心 | Data visualization adjusting method and system based on meteorological database |
| CN117571976A (en) * | 2024-01-15 | 2024-02-20 | 深圳市恒义建筑技术有限公司 | Method for measuring and calculating release rate of harmful substances of interior decoration material and related equipment |
| CN117933466A (en) * | 2024-01-23 | 2024-04-26 | 中科三清科技有限公司 | Pollution prediction method and device, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108053071A (en) | Regional air pollutant concentration Forecasting Methodology, terminal and readable storage medium storing program for executing | |
| Yafouz et al. | Hybrid deep learning model for ozone concentration prediction: comprehensive evaluation and comparison with various machine and deep learning algorithms | |
| CN112765808B (en) | Ecological drought monitoring and evaluating method | |
| Hosseini et al. | A data fusion-based methodology for optimal redesign of groundwater monitoring networks | |
| Mateu et al. | Spatio-temporal design: Advances in efficient data acquisition | |
| CN108846503B (en) | Dynamic respiratory system disease ill person number prediction method based on neural network | |
| Yahya et al. | Designing weather forecasting model using computational intelligence tools | |
| Chen | Logistic models of fractal dimension growth of urban morphology | |
| Fan et al. | Development of PCA-based cluster quantile regression (PCA-CQR) framework for streamflow prediction: Application to the Xiangxi river watershed, China | |
| Chen et al. | An adaptive Kalman filtering approach to sensing and predicting air quality index values | |
| CN107423857B (en) | A kind of long-term water multiple target joint probability Forecasting Methodology in region | |
| Mahajan et al. | An empirical study of PM2. 5 forecasting using neural network | |
| CN106803128A (en) | The Forecasting Methodology and device of photovoltaic generation | |
| Sahraei et al. | Application of machine learning models to predict maximum event water fractions in streamflow | |
| Beuchat et al. | A robust framework for probabilistic precipitations downscaling from an ensemble of climate predictions applied to Switzerland | |
| CN115270632A (en) | System and method for predicting water bloom risk of drinking water source | |
| Richards et al. | Regression modelling of spatiotemporal extreme US wildfires via partially-interpretable neural networks | |
| Tanaka et al. | Refining coarse-grained spatial data using auxiliary spatial data sets with various granularities | |
| Başakın et al. | Developing a novel approach for missing data imputation of solar radiation: A hybrid differential evolution algorithm based eXtreme gradient boosting model | |
| Song et al. | Evaluating the performance of climate models in reproducing the hydrological characteristics of rainfall events | |
| Gao et al. | A multifactorial framework for short-term load forecasting system as well as the jinan’s case study | |
| Han et al. | Developing a multi-objective simulation-optimization model for ecological water conveyance in arid inland river basins | |
| Ali et al. | Raingauge network optimization in a tropical urban area by coupling cross-validation with the geostatistical technique | |
| Raja et al. | Drought prediction and validation for desert region using machine learning methods | |
| CN118350678B (en) | Water environment monitoring data processing method and system based on Internet of things and big data |
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: 20180518 |