CN110766222B - PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest - Google Patents
PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest Download PDFInfo
- Publication number
- CN110766222B CN110766222B CN201911004338.3A CN201911004338A CN110766222B CN 110766222 B CN110766222 B CN 110766222B CN 201911004338 A CN201911004338 A CN 201911004338A CN 110766222 B CN110766222 B CN 110766222B
- Authority
- CN
- China
- Prior art keywords
- data
- random forest
- particle swarm
- training
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000002245 particle Substances 0.000 title claims abstract description 57
- 238000007637 random forest analysis Methods 0.000 title claims abstract description 39
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000005457 optimization Methods 0.000 title claims abstract description 15
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 22
- 238000012549 training Methods 0.000 claims abstract description 22
- 238000012795 verification Methods 0.000 claims abstract description 12
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 230000008569 process Effects 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 8
- 238000005070 sampling Methods 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 3
- 239000003344 environmental pollutant Substances 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 5
- 231100000719 pollutant Toxicity 0.000 description 5
- 238000011160 research Methods 0.000 description 4
- 239000000809 air pollutant Substances 0.000 description 2
- 231100001243 air pollutant Toxicity 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 238000003915 air pollution Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000000356 contaminant Substances 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000036039 immunity Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012546 transfer 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Marketing (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Educational Administration (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Primary Health Care (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest, which comprises the following steps: step 1, acquiring air quality data needing predictive analysis, and dividing the air quality data into a training data set and a verification data set; step 2, preprocessing the acquired data; step 3, initializing various parameters of a Particle Swarm Optimization (PSO), wherein the set parameters are as follows: group size m, the range of values of the position and the speed of particles, fitness function and the like; step 4, inputting training set data and particle positions, and training a prediction model by using a random forest algorithm; and 5, taking the root mean square error of the verification set data in the training model as an adaptability value, and continuously optimizing the number t and the feature number f of subtrees of the random forest through the particle swarm to obtain a final optimal prediction model. By adopting the technical scheme of the invention, the data quality is improved, the self-adaptive selection of the prediction model parameters is realized, and the PM2.5 concentration in the air in a future period of time can be effectively predicted.
Description
Technical Field
The invention relates to the technical field of air quality prediction, in particular to a PM2.5 pollutant concentration prediction method participating in air quality evaluation.
Background
The problem of air pollution has become a global hot spot in recent years, and along with the frequent occurrence of severe polluted weather such as haze and the like, and the improvement of environmental awareness and health awareness of people, the concentration of PM2.5 has become an important information of frequent attention in daily life of people. In order to timely and accurately perform PM2.5 concentration early warning work, a plurality of scholars and research institutions at home and abroad start to transfer research emphasis to air quality prediction.
Currently, various algorithmic models have been successfully applied to air quality prediction with good results. These prediction methods can be roughly classified into three types: a prediction method based on a statistical model, a prediction method based on a physicochemical mechanism model and a prediction method based on machine learning.
The prediction method based on machine learning utilizes a specific learning algorithm to search the conversion rule among the air pollutant concentration, meteorological parameters and other related historical data, and simplifies the prediction flow of air quality. Meanwhile, mass data are reserved in various industries due to the arrival of big data age, sufficient learning samples are provided for machine learning, and development of a machine learning method and application are greatly promoted. In recent years, many students began using machine learning algorithms to make air pollutant concentration predictions. The random forest algorithm has the advantages of high prediction accuracy, capability of preventing overfitting, strong noise immunity and the like, and research experiments of Du Xu et al (PM 2.5 concentration prediction model based on random forest regression analysis) and Yang Saiqi et al (application of the random forest algorithm in urban air quality prediction) respectively prove that the random forest algorithm is superior to algorithms such as a neural network and a support vector machine, so that the PM2.5 concentration prediction method for optimizing random forest parameters by the particle swarm algorithm is designed.
Disclosure of Invention
The invention aims to solve the technical problem of providing a PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest, which is used for obtaining the evolution rule of PM2.5 in the atmosphere by analyzing a large amount of historical data so as to predict the mass concentration of PM2.5 in the air in time and accurately in a future period of time. According to the invention, historical data of an air quality monitoring station, including meteorological data, pollutant concentration, time, related stations and the like, are collected, then the collected data are preprocessed, a prediction model is built by a random forest algorithm according to the processed data, and parameters of the random forest are optimized by a particle swarm algorithm, so that a final prediction result is obtained. The whole process is carried out on a big data platform, so that the PM2.5 concentration in the air in a future period of time is predicted with high precision, rapidness and intellectualization.
The invention achieves the aim through the following technical scheme: the PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest, as shown in FIG. 1, comprises the following steps:
step 1, acquiring air quality data needing predictive analysis, and dividing the air quality data into a training data set and a verification data set;
step 2, preprocessing the acquired data, wherein the preprocessing comprises filling of a missing value, balanced sampling and the like;
step 3, initializing various parameters of a Particle Swarm Optimization (PSO), wherein the set parameters are as follows: the group scale m, the value range of the position and the speed of each particle, the iteration number k, the fitness function setting and the like;
step 4, inputting training set data and positions of particles (the number t of subtrees and the characteristic number f) to train a random forest prediction model;
and 5, continuously optimizing the number t and the feature number f of subtrees of the random forest by using the root mean square error of the verification set data in the training model as an fitness value, so that the root mean square error of the model is minimum as the fitness value, and obtaining the final optimal prediction model.
The method has the beneficial effects that:
(1) Relevant characteristics in the fields of meteorology, local pollutants and the like are introduced, and time characteristics including the characteristics of working days, seasons and the like and the spatial characteristics of the influence of surrounding air monitoring stations on the PM2.5 concentration of a prediction area are added in consideration of the retention characteristics of suspended particles, so that the PM2.5 concentration in the air can be predicted more accurately.
(2) The invention adopts a large amount of historical data in the process of training the prediction model, so that the problems of unbalanced large amount of missing values and concentration levels of each PM2.5 in the data are solved by adopting an effective missing value processing and balanced sampling method in the stage of data preprocessing, and the data quality is improved, thereby improving the prediction precision.
(3) Compared with the traditional random forest, the parameters of the random forest need to be manually set, and in the process of building the prediction model, the particle swarm algorithm is adopted to optimize the model built by the random forest, so that the self-adaptive optimization parameters are realized in the whole modeling process, and the whole prediction model is biased to be intelligent. The whole modeling process is realized in the spark environment of a large data platform, so that the invention has higher operation efficiency and faster modeling speed than the traditional single machine.
(4) The prior research results and the implementation results prove that the random forest has obvious advantages compared with other traditional prediction methods, and has higher prediction accuracy. The invention is not only suitable for air quality prediction in theory, but also suitable for prediction of other time series data.
Drawings
Fig. 1 shows a flowchart of the PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest according to the present invention.
Detailed Description
Specific embodiments of the present invention are described in detail below.
A PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest, as shown in FIG. 1, is a method flow chart, comprising the following steps:
step 1, acquiring air quality data needing predictive analysis, and dividing the air quality data into a training data set and a verification data set; the method comprises the following specific steps:
1.1A data record is formed by taking the average value of meteorological data and the concentration of primary pollutants in the air per hour as attribute values, wherein the meteorological data comprise meteorological characteristics such as air temperature, humidity, wind direction, air pressure, visibility and the like, and the meteorological data are respectively attribute values, and the pollutants comprise PM2.5 and SO 2 、PM10、NO 2 、CO、O 3 O and O 3 8 hour and 24 hour running averages of concentrations, etc., each meteorological feature and each contaminant concentration is a property.
And 1.2, the training set is used for training a prediction model, the verification set is used for calculating the fitness value of the particle swarm algorithm, and the optimal parameters are adjusted.
1.3, the collected data is historical data of five years, the first three years are used as training sets, the next year is used as verification sets, and the last year is used as test set.
Step 2, preprocessing the acquired data, wherein the preprocessing comprises filling of a missing value, balanced sampling and the like; the method comprises the following specific steps:
2.1, (1) deleting a record when two or more attribute values are missing in the record; (2) During the experimental process, the correlation between PM2.5 and PM10 is found to be the strongest, and the correlation has obvious linear correlation trend; when one record lacks only the PM2.5 concentration value, the invention uses the PM10 concentration value to replace the PM2.5 value; (3) If a certain attribute value is continuously missing, the attribute missing value is replaced by taking an average value according to the data of the former and latter hours of the attribute.
2.2, clustering meteorological data according to contour coefficients after missing value processing in order to reduce errors and improve prediction accuracy;
2.3, improving the influence of PM2.5 concentration level unbalance on random forest prediction performance by using a data undersampling method;
2.4, clustering and undersampling the data in the steps 2.2 and 2.3 to form data with different PM2.5 concentration levels, and then establishing different prediction models by using a random forest algorithm according to different kinds of data in the step 4.
Step 3, initializing various parameters of a Particle Swarm Optimization (PSO), wherein the set parameters are as follows: the group scale m, the value range of the position and the speed of each particle, the iteration number k, the fitness function setting and the like; the step is the parameter initialization of the step 4, and the specific steps are as follows:
3.1, when the random forest algorithm is used for establishing a model, the parameters for determining the performance of the model are mainly the number t of decision trees and the characteristic number f used for constructing the decision trees, and the optimal parameter combination can improve the model prediction precision and approach to a true value, but in the traditional random forest algorithm, the two parameters are manually set, and in the particle swarm optimization process, the position of the particles is set as a two-dimensional vector X i ＝(x 1 ,x 2 ) Let x 1 ＝t、x 2 In the process of continuously updating the particle position, the process of parameter optimization is also referred to as =f.
3.2 initializing a particle group, setting the position X of the particle, wherein the group size m=100 i ＝(x 1 ,x 2 ) And velocity V i ＝(v 1 ,v 2 ) Is a two-dimensional vector, wherein the position of each particle is made to represent two parameters (the number t of subtrees and the characteristic number f) of a random forest, x 1 The value interval of (2) is [1,20 ]]，x 2 The value interval of (5) is [100,1500 ]]，v 1 The value interval of (1) is [ -1,1]，v 2 The value interval of (2) is [ -20,20]The number of iterations k=200, and two learning factors C 1 And C 2 ；
3.3, setting a fitness function: the root mean square error of the prediction model trained by the verification set at the optimal position of each particle (namely the parameter of the random forest) is used as the fitness function of the particle swarm, and the step is the presetting and the setting of the iterative judgment condition in the step 5.
Step 4, inputting training set data and the positions of particles (the number t of subtrees and the characteristic number f), and training a prediction model by using a random forest algorithm; the method comprises the following specific steps:
and 4.1, sampling the training set data processed in the step 2 through a self-help sampling method, inputting the training set data and the initialized particle positions in the step 3 into a random forest algorithm, constructing t regression sub-trees with the characteristic number f in the random forest algorithm, deciding output results of the sub-trees to obtain a final predicted value, and training a current optimal random forest prediction model of each particle according to the stored current optimal position of the particle swarm, thereby obtaining the prediction models of different concentration levels after clustering in the step 2.
Step 5, taking root mean square error of verification set data in a training model as an adaptability value, and continuously optimizing the number t and the feature number f of subtrees of a random forest through a particle swarm to obtain a final optimal prediction model; the method comprises the following specific steps:
5.1, inputting the verification set processed in the step 2 into a trained prediction model to obtain root mean square error (namely fitness value in the step 3) of each particle building model;
and 5.2, in the conventional particle swarm optimization process, calculating the fitness value of each particle, comparing the fitness value with the optimal fitness value and the optimal position of the swarm and the particle, then updating the position and the speed of each particle, and searching for an optimal final result, wherein a specific updating formula is as follows:
wherein inertial weight ω = 0.75; learning factor c 1 ＝c 2 ＝1.5；r 1 And r 2 Is interval [0,1 ]]Random number on, dimension d=1, 2; particle count i=1, 2, …, m; k is the current iteration number; gbest represents the global optimal position of the particle swarm; pbest represents the locally optimal position of the particle swarm (i.e., the optimal position currently found by the particle itself);
from step 3, the two parameter subtrees of the random forest are respectivelyThe number and the feature number are set as two coordinates x of the particle position 1 And x 2 According to step 5.1, the current position (t and f) and fitness value of the particles can be calculated, compared with the optimal position and fitness value found so far, the position X (update parameter combination) and the speed V of the particles are updated, and the optimal position and fitness value of the group are updated and stored; thus, the updated formulas for the modified position and velocity are as follows:
the number of the two parameter subtrees of the random forest and the modeling process of the used feature numbers are integers, so that the position and the speed of the particles are made to be integers by using a rounding function, and the optimal solution of the model is made to be integers;
and 5.3, judging whether an algorithm ending condition is reached, namely, the maximum iteration times are reached or the optimal parameter combination of the algorithm enables the fitness function value to be minimum, outputting a result if the condition is met, and otherwise, turning to the iterative learning of the step 4. And (3) until the ending condition is reached, establishing different prediction models according to different data types after clustering in the step (2).
The single machine environment of the experiment of the embodiment of the invention is a PC with 1 PC with 8 cores and 16G memories. And taking python as a main programming language, wherein Spark clusters are 1 Master node and 2 Slave nodes, wherein an HDFS file system of Hadoop is responsible for storing data, and Spark is responsible for experimental calculation. Spark version is Spark-1.6.0, hadoop version is Hadoop-2.6.2, and the platform is built on a Linux operating system CentOS6.5.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the invention in any way, and any simple modification, equivalent variation, of the above embodiments according to the technical principles of the present invention are within the scope of the technical solutions of the present invention.
Claims (1)
1. The PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest is characterized by comprising the following steps:
step 1, acquiring air quality data needing predictive analysis, and dividing the air quality data into a training data set and a verification data set;
step 2, preprocessing the acquired data, wherein the preprocessing comprises filling of a missing value and balanced sampling; when filling the obtained data with the missing value, when one record only lacks the PM2.5 concentration value, replacing the PM2.5 value with the PM10 concentration value;
step 3, initializing each parameter of a particle swarm algorithm, wherein the set parameters are as follows: the group scale m, the value range of the position and the speed of each particle, the iteration number k, and the fitness function are set;
step 4, inputting training set data and the positions of the particles, and setting the positions of the particles as two-dimensional vectors X i ＝(x 1 ,x 2 ) Let x1=t, x2=f, t and f be the number of subtrees and the feature number respectively, training a prediction model by using a random forest algorithm;
step 5, taking the root mean square error of the verification set data in the training model as an fitness value, continuously optimizing the number t and the characteristic number f of subtrees of the random forest through a particle swarm, and enabling the root mean square error of the model to be the minimum fitness value to obtain a final optimal prediction model; the updated formulas for the modified position and velocity are as follows:
x id (k+1)＝[x id (k)+V id (k+1)-0.5]
wherein inertial weight ω = 0.75; learning factor c 1 ＝c 2 ＝1.5；r 1 And r 2 Is interval [0,1 ]]Random number on, dimension d=1, 2; particle count i=1, 2, …, m; k is the current iteration number; gbest represents the global optimal position of the particle swarm; pbest represents the locally optimal position of the particle swarm;
since the number of the two parameter subtrees of the random forest and the modeling process of the used feature numbers are necessarily integers, the position and the speed of the particles are made to be integers by using a rounding function, and thus the optimal solution of the model is also made to be integers.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911004338.3A CN110766222B (en) | 2019-10-22 | 2019-10-22 | PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911004338.3A CN110766222B (en) | 2019-10-22 | 2019-10-22 | PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110766222A CN110766222A (en) | 2020-02-07 |
| CN110766222B true CN110766222B (en) | 2023-09-19 |
Family
ID=69332520
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911004338.3A Active CN110766222B (en) | 2019-10-22 | 2019-10-22 | PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110766222B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112070103B (en) * | 2020-04-26 | 2021-04-30 | 河海大学 | Method for inverting atmospheric visibility through microwave link network gridding self-adaptive variable scale |
| CN111461263A (en) * | 2020-05-19 | 2020-07-28 | 昆明理工大学 | Method for predicting concentration of dissolved gas in transformer oil based on EMD-RF |
| CN112116058B (en) * | 2020-09-16 | 2022-05-31 | 昆明理工大学 | Transformer fault diagnosis method for optimizing multi-granularity cascade forest model based on particle swarm algorithm |
| CN112381290A (en) * | 2020-11-13 | 2021-02-19 | 辽宁工程技术大学 | Method for predicting coal gas content by random forest and gray wolf optimization |
| CN112686296B (en) * | 2020-12-29 | 2022-07-01 | 昆明理工大学 | Octane loss value prediction method based on particle swarm optimization random forest parameters |
| CN114169106A (en) * | 2021-12-13 | 2022-03-11 | 汕头大学 | Method for evaluating lateral displacement caused by jet grouting pile based on improved random forest model |
| CN114662611B (en) * | 2022-04-07 | 2022-12-09 | 中科三清科技有限公司 | Method and device for restoring particulate component data, electronic equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108038306A (en) * | 2017-12-11 | 2018-05-15 | 太原理工大学 | A kind of method of power boiler burning modeling and optimization towards magnanimity high dimensional data |
| CN108426812A (en) * | 2018-04-08 | 2018-08-21 | 浙江工业大学 | A kind of PM2.5 concentration value prediction techniques based on Memory Neural Networks |
| CN109376903A (en) * | 2018-09-10 | 2019-02-22 | 浙江工业大学 | A kind of PM2.5 concentration value prediction technique based on game neural network |
| CN110059852A (en) * | 2019-03-11 | 2019-07-26 | 杭州电子科技大学 | A kind of stock yield prediction technique based on improvement random forests algorithm |
| CN110334452A (en) * | 2019-07-09 | 2019-10-15 | 中南大学 | A kind of wisdom agricultural air pollution concentration method for early warning by different level |
| CN113128134A (en) * | 2021-06-17 | 2021-07-16 | 中国矿业大学(北京) | Mining area ecological environment evolution driving factor weight quantitative analysis method |
-
2019
- 2019-10-22 CN CN201911004338.3A patent/CN110766222B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108038306A (en) * | 2017-12-11 | 2018-05-15 | 太原理工大学 | A kind of method of power boiler burning modeling and optimization towards magnanimity high dimensional data |
| CN108426812A (en) * | 2018-04-08 | 2018-08-21 | 浙江工业大学 | A kind of PM2.5 concentration value prediction techniques based on Memory Neural Networks |
| CN109376903A (en) * | 2018-09-10 | 2019-02-22 | 浙江工业大学 | A kind of PM2.5 concentration value prediction technique based on game neural network |
| CN110059852A (en) * | 2019-03-11 | 2019-07-26 | 杭州电子科技大学 | A kind of stock yield prediction technique based on improvement random forests algorithm |
| CN110334452A (en) * | 2019-07-09 | 2019-10-15 | 中南大学 | A kind of wisdom agricultural air pollution concentration method for early warning by different level |
| CN113128134A (en) * | 2021-06-17 | 2021-07-16 | 中国矿业大学(北京) | Mining area ecological environment evolution driving factor weight quantitative analysis method |
Non-Patent Citations (6)
| Title |
|---|
| Particulate matter concentration from open-cut coal mines: A hybrid machine learning estimation;Qi, CC 等;《environmental pollution》;第263卷;第1-10页 * |
| 城市空气质量的PM2.5浓度预测及监测网络优化研究;李松洲;《中国优秀硕士学位论文全文数据库（电子期刊）工程科技Ⅰ辑》(第7期);第B027-707页 * |
| 基于随机森林和气象参数的PM2.5浓度等级预测;任才溶等;《计算机工程与应用》;第214页右栏第2段、第218页图7 * |
| 基于随机森林模型的城市空气质量评价;杨瑞君等;《计算机工程与设计》(第11期);第3151-3156页 * |
| 寒区露天煤矿粉尘污染特征及预测;罗怀廷;《中国知网》;第1-146页 * |
| 采空区煤自燃预测的随机森林方法;邓军等;《煤炭学报》;第2800页"摘要"、第2804页左栏第5-6段 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110766222A (en) | 2020-02-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110766222B (en) | PM2.5 concentration prediction method based on particle swarm parameter optimization and random forest | |
| CN112465243B (en) | Air quality forecasting method and system | |
| CN108491970B (en) | Atmospheric pollutant concentration prediction method based on RBF neural network | |
| CN111798051A (en) | Air quality space-time prediction method based on long-short term memory neural network | |
| CN106650767B (en) | Flood forecasting method based on cluster analysis and real-time correction | |
| CN116186566B (en) | Diffusion prediction method and system based on deep learning | |
| CN108009674A (en) | Air PM2.5 concentration prediction methods based on CNN and LSTM fused neural networks | |
| CN110619360A (en) | Ultra-short-term wind power prediction method considering historical sample similarity | |
| CN112085163A (en) | Air quality prediction method based on attention enhancement graph convolutional neural network AGC and gated cyclic unit GRU | |
| CN111814956B (en) | Multi-task learning air quality prediction method based on multi-dimensional secondary feature extraction | |
| CN112016175B (en) | Water supply pipe network pressure measuring point optimal arrangement method based on tree hierarchical clustering | |
| CN107346459B (en) | Multi-mode pollutant integrated forecasting method based on genetic algorithm improvement | |
| CN108399470B (en) | Indoor PM2.5 prediction method based on multi-example genetic neural network | |
| CN115759488A (en) | Carbon emission monitoring and early warning analysis system and method based on edge calculation | |
| CN113987912A (en) | Pollutant on-line monitoring system based on geographic information | |
| CN110610209A (en) | Air quality prediction method and system based on data mining | |
| CN115526298A (en) | High-robustness comprehensive prediction method for concentration of atmospheric pollutants | |
| CN113516304A (en) | Space-time joint prediction method and device for regional pollutants based on space-time graph network | |
| CN116757057A (en) | Air quality prediction method based on PSO-GA-LSTM model | |
| CN108564110A (en) | A kind of Air Quality Forecast method based on clustering algorithm | |
| CN113762591A (en) | Short-term electric quantity prediction method and system based on GRU and multi-core SVM counterstudy | |
| CN113377750B (en) | Hydrological data cleaning method and system | |
| CN110852493A (en) | Atmospheric PM2.5 concentration prediction method based on multiple model comparisons | |
| CN112598050A (en) | Ecological environment data quality control method | |
| CN117350146A (en) | GA-BP neural network-based drainage pipe network health evaluation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |