CN114036736B - Causal network learning method based on local Granges causal analysis - Google Patents
Causal network learning method based on local Granges causal analysis Download PDFInfo
- Publication number
- CN114036736B CN114036736B CN202111304900.1A CN202111304900A CN114036736B CN 114036736 B CN114036736 B CN 114036736B CN 202111304900 A CN202111304900 A CN 202111304900A CN 114036736 B CN114036736 B CN 114036736B
- Authority
- CN
- China
- Prior art keywords
- causal
- variable
- variables
- local
- analysis
- 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
- 230000001364 causal effect Effects 0.000 title claims abstract description 104
- 238000004458 analytical method Methods 0.000 title claims abstract description 56
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000012545 processing Methods 0.000 claims abstract description 6
- 238000013112 stability test Methods 0.000 claims abstract description 5
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 11
- 230000004044 response Effects 0.000 claims description 6
- 238000012935 Averaging Methods 0.000 claims description 4
- 238000012047 cause and effect analysis Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 238000010187 selection method Methods 0.000 claims description 4
- 210000004907 gland Anatomy 0.000 claims description 3
- 238000001744 unit root test Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 2
- 238000005070 sampling Methods 0.000 claims description 2
- 238000007418 data mining Methods 0.000 abstract description 2
- 230000001502 supplementing effect Effects 0.000 abstract 1
- 230000000694 effects Effects 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 239000000809 air pollutant Substances 0.000 description 2
- 231100001243 air pollutant Toxicity 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000006641 stabilisation Effects 0.000 description 2
- 238000011105 stabilization Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000003915 air pollution Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 238000010223 real-time analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Geometry (AREA)
Abstract
The invention belongs to the field of data mining, and provides a causal network learning method based on local Granges causal analysis. The method comprises the steps of preprocessing acquired data, and supplementing missing data by adopting an average value interpolation method. And then carrying out stability test and processing on the completed data so as to meet the assumption of building a model. And normalizing the data to eliminate the influence of different variable dimensions. And finally, establishing a causal network learning algorithm based on the local Granges causal analysis, realizing the aim of accurately exploring the causal relationship among the variables, and simultaneously displaying a dynamic causal relationship curve among different variables so as to quantitatively and definitely analyze the causal relationship among the variables among the systems.
Description
Technical Field
The invention belongs to the technical field of data mining, relates to a causal network learning method based on local Granges causal analysis, and aims to explore the relationship among variables in high-dimensional data in the fields of weather and the like.
Background
A multivariate time series is a set of discrete observations of a plurality of variables distributed over time, which are widely found in many fields of finance, industry, traffic, weather, etc. For example, in the field of research on air pollution, in recent years, with the rapid development of industries and the increase of vehicles in China, the concentration of harmful substances in the atmosphere caused by the combustion of coal, oil and the like is obviously increased, and the phenomena of air quality reduction, haze weather and the like are caused. Haze weather is an atmospheric pollution state and is a general expression for exceeding the content of various suspended particulate matters in the atmosphere, and particularly PM2.5 (particulate matters with an aerodynamic equivalent diameter of 2.5 micrometers or less) is considered as the source of the haze weather. Air pollutants such as PM2.5 not only cause poor atmospheric visibility, thereby causing social problems such as traffic jam, but also can cause harm to the physical health of people by being inhaled into the human body. Therefore, modeling is performed for PM2.5 and other air pollutants, causal relation among Air Quality Indexes (AQIs) is described, and prediction of possible phenomena in the next step is realized according to the current and historical moment states of each variable, so that theoretical support is provided for preventing and controlling the atmospheric pollution. Therefore, the establishment of an effective multivariate causal analysis model has important practical significance.
The causal relationship analysis method interprets the influence of the reason variable on the result variable by researching the driving response relationship among the variables of the system, and can effectively infer the internal structure and the operation mechanism of the system, thereby overcoming the defect that the traditional correlation analysis method is difficult to process the indirect relationship and the asymmetric relationship. The grange causal analysis method is a common method for exploring causal relationships among variables. It estimates the time dependence of variables in the model by means of a vector autoregressive model and is based on a predictable idea: for both time series X and Y, if the addition of time series X history information helps to reduce the prediction error of Y, then it can be said that there is a Granger causal relationship of X→Y. The directed graph shown in fig. 1 is a gland causal graph, and can intuitively show the causal relationship existing among the variables. Nodes represent variables and arrows represent causal relationships between two variables. For example, the cause of the glabroil causality of x 3 is x 5.
Currently, the improved method based on the grange causal analysis has been widely applied to the problems of multivariate and nonlinearity. However, the improved method only quantitatively considers the causal relation among the variables, and ignores the dynamic characteristics among the variables. STRAMAGLIA SEBASTIANO et al in paper "Stramaglia Sebastiano,Tomas Scagliarini,Yuri Antonacci,et al.Local Granger Causality[J].Physical Review E,2021,103(2)." show that a local gracile cause and effect analysis can reveal the instantaneous state of information transfer between variables, which cannot be represented by a quantitative value. This approach has a better effect, especially when exploring interactions between variables for a certain period of time. In addition, traditional methods based on graininess causal analysis are modeled based on vector autoregressive models, which results in poor performance when the data dimension is too high. RungeJakob et al in paper "RungeJakob,Sebastian Bathiany,Erik Bollt,et al.Inferring Causation from Time Series in Earth System Sciences[J].Nature Communications,2019,10(1):2553-2553." analyzed the current state of the art causal analysis method and pointed out that causal network learning methods are suitable for causal relationship exploration of high dimension.
Aiming at the problem of causal relation analysis of a complex high-dimensional system, the invention provides a causal network learning method based on local Granges causal analysis, which is used for modeling causal analysis among variables in research fields such as weather pollution and the like.
Disclosure of Invention
The invention aims at solving the technical problems that the traditional Gray causal analysis method cannot be suitable for a high-dimensional time sequence due to structural limitation, real-time dynamic exploration among variables in a process cannot be performed, an original Gray causal analysis model is expanded, a causal network learning algorithm based on local Gray causal analysis is provided, and accurate causal relation exploration of high-dimensional data and real-time display of information in the process are realized. The method aims to expand the original method to be suitable for high-dimension and can display more dynamic information.
The method is faced with complex systems such as AQI and the like, and provides a causal network learning method for exploring between variables among the systems and main pollutants PM2.5 thereof. Preprocessing the acquired data, and complementing the missing data by adopting an average value interpolation method. And then carrying out stability test and processing on the completed data so as to meet the assumption of building a model. And normalizing the data to eliminate the influence of different variable dimensions. And finally, establishing a causal network learning algorithm based on local Granges causal analysis, realizing the aim of accurately exploring causal relation among variables, and simultaneously displaying dynamic causal relation curves among different variables so as to quantitatively and definitely analyze the relation between each variable among the systems and PM2.5 and realize the analysis of PM2.5 influence factors.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: a causal network learning method based on local Granges causal analysis comprises the following specific steps:
Step 1: acquiring an air quality index AQI and meteorological observation data; preprocessing the multidimensional AQI and meteorological time sequence data, interpolating the missing samples by an average interpolation method, and analyzing outliers to obtain smooth time sequence data; adopting a unit root test method to perform stability test on the time series data, if the test result has a unit root, indicating that the data is not stable, adopting a differential method to perform stability treatment, otherwise, not obtaining the unit root, indicating that the data is stable, and not needing to perform stabilization treatment; then normalizing the time sequence data;
Step 2: determining the maximum hysteresis order tau max of the time series data obtained by the processing in the step 1 through a red pool information criterion, and generating a historical variable matrix X -;
Wherein, Data representing the current t moment, τ max is the maximum hysteresis order, d represents the variable number, and n represents the sample number;
Step 3: with the PC algorithm, with each variable Feature selection processing is carried out for the target to obtain each variable/>Related feature subset/>Wherein the variable/>For the data acquired at the time of t sampling, i=1, 2, …, d; each variable/>By its related feature subset/>The expression is represented by the following formula (2):
Wherein f i (. Cndot.) represents the mapping function, Representing the deviation;
step 4: related feature subset Sending the result as a condition set to a local Granges cause and effect analysis model to obtain quantitative cause and effect relationship values, specifically
Step 4.1: selecting a drive variableAnd target variable/>The most relevant feature subsets of the two are respectively found from a historical variable matrix X - of the time sequence through a feature selection method: /(I)And/>Variable/>The feature subset of (2) is smoothed one bit backwards to obtain the corresponding/>Is a subset of features of (a);
step 4.2: will drive the variable And target variable/>Is fed into a local glabellar causal analysis model as a condition set, as shown in formula (3):
Step 4.3: weighted summation to obtain driving variable For the target variable/>Results of causal analysis of (2);
Step 5: averaging results obtained by the local Grangel causal analysis to obtain quantitative causal relation values among variables, and drawing the results of the local Grangel causal analysis to obtain a dynamic causal relation curve among the variables;
step 6: the causal relation between any variable and other variables is obtained, and a driving variable is selected as a relevant influence factor of a target variable according to the causal relation result; and establishing a prediction model by using the echo state network for analysis to obtain a prediction analysis result of the target variable.
The local graininess causal analysis model is used for further exploring the dynamic characteristics among variables on the basis of the causal analysis result of quantitative values; the local glanger causal analysis model is represented by formula (4):
Lgc(ut,wt,yt)=GC+γ(ut,wt,yt) (4)
Wherein GC is a standard gland cause and effect analysis; gamma (u t,wt,yt) is a function about u t,wt,yt to reveal dynamic characteristics between variables; u t denotes a set of condition variables, w t denotes a driving variable, and y t denotes a response variable; the GC and γ (u t,wt,yt) are represented using the second order statistics of the unified procedure:
Where |·| represents the determinant of the matrix, Is a historical state and current state observation; /(I) And/>Is a covariance matrix; the mean value of γ (u t,wt,yt) is 0, representing < L gc (t) > = GC; </cndot > represents time series averaging.
Compared with the prior art, the invention has the following obvious advantages:
The present invention focuses on the quantitative and dynamic real-time analysis of causal analysis between high dimensional data variables. Firstly, the invention expands the original data by AIC criterion, then establishes a causal network learning algorithm based on a feature selection method, and the algorithm can be different from a general causal analysis method based on a vector autoregressive framework to realize causal analysis of high-dimensional data. And then, sending the characteristic selection result into a local Granges causal analysis model to obtain a quantitative causal relation value among variables. In addition, the local graininess causal analysis method also obtains a dynamic response curve among variables, which provides more information for exploring causal relations among variables. The invention not only obtains quantitative causal relation measurement aiming at high-dimensional data, but also can obtain dynamic causal relation among variables.
Drawings
FIG. 1 is a graph showing the cause and effect of a Grangel, nodes representing vectors, arrows representing the cause and effect relationships that exist between the vectors;
FIG. 2 is a flow chart of the present invention;
FIG. 3 is a graph showing the dynamic relationship between partial variables and PM2.5 variables.
FIG. 4 (a) is a graph of GC versus PM2.5 target and predicted values;
FIG. 4 (b) is a graph of GC prediction curve error versus PM 2.5;
FIG. 5 (a) is a graph comparing target and predicted values of PCMCI versus PM 2.5;
FIG. 5 (b) is a graph of predicted curve error for PCMCI versus PM 2.5;
FIG. 6 (a) is a graph comparing target and predicted values for PM2.5 according to the present invention;
fig. 6 (b) is a graph of prediction curve error for PM2.5 according to the present invention.
Detailed Description
The present invention will be described in further detail with reference to specific examples and simulation drawings.
The hardware equipment used by the invention comprises a PC machine.
Fig. 2 is a flow chart of a causal network learning method based on local graininess causal analysis, which specifically includes the following steps:
Step 1: acquiring 5088 groups of data from 2021, 1 and 1 to 2021, 7 and 31 in the Xuhui area of Shanghai, wherein the total data are 10-dimensional, the meaning numbers of variables are shown in table 1 and are acquired once per hour, and then interpolating and analyzing the missing values and the abnormal values of the multidimensional AQI and the meteorological dataset; adopting a unit root test method to perform stability test on the data, and performing primary differential stabilization treatment on the data according to test results; normalizing the time sequence data;
TABLE 1 Shanghai AQI and Meteorological data numbering correspondence table
Step 2: determining a maximum hysteresis coefficient tau max as 8 according to a red pool information criterion, and expanding hysteresis terms of original data according to a formula (1) to obtain a historical variable matrix X -;
Step 3: calculating the most relevant feature subset for each variable by using a PC feature selection method, taking a variable PM2.5 as an example, firstly obtaining the relevant feature subset from a history set, and obtaining PM10 (t-1), NO2 (t-1), CO (t-1), PM10 (t-2) and AQI (t-5) as results;
Step 4: and (3) sending the feature subset obtained in the step (5) into a local Grangel causal analysis model as a condition, and exploring causal relation between each variable in the historical variable matrix and the PM2.5 variable, wherein the causal relation comprises the following specific steps of:
step 4.1: taking the relation between PM10 and PM2.5 as an example. First a relevant subset of the drive variables PM10 and the response variables PM2.5 is found; according to the PC algorithm, the condition is met through n iterations, and a relevant subset containing n variables is obtained;
Step 4.2: bringing the time series of the driving variable and the response variable and the time series of the corresponding relevant subsets thereof into a local graininess causal analysis model, and calculating a quantitative causal analysis result through a formula (3);
Step 4.3: weighted summation to obtain driving variable For the target variable/>Results of causal analysis of (2);
Step 5: steps 4.1 and 4.2 are repeated until each variable in the set of historical variables is calculated as described above with PM 2.5. Through the above process, a causal value matrix is obtained 10 Represents the number of variables, 8 represents the maximum hysteresis coefficient, and then weighted summation is carried out to obtain a causal analysis result among the variables;
The results of the causal analysis are shown in table 2: the larger the values in the table, the greater the causal relationship between the two variables. Because of the interplay between variables, there are no two variables that are absolutely independent. As can be seen from table 2, the variable indexes of the Shanghai Xuhui having causal relationship with PM2.5 include SO2, O3, air temperature, etc.;
TABLE 2 causal index for PM2.5
GC | PCMCI | The method of the invention | |
PM2.5 | 0.13 | 0.37 | 0.10 |
PM10 | 0.02 | 0.39 | 0.07 |
SO2 | 0.00 | 0.13 | 0.59 |
NO2 | 0.03 | 0.33 | 0.03 |
O3 | 0.01 | 0.22 | 0.47 |
CO | 0.04 | 0.25 | 0.04 |
AQI | 0.06 | 0.20 | 0.02 |
Air temperature | 0.01 | 0.10 | 1.25 |
Dew point | 0.00 | 0.14 | 0.06 |
Wind speed | 0.00 | 0.10 | 0.01 |
Step 6: by plotting the output time sequence of the grange causal analysis, the dynamic causal relationship between partial variables and PM2.5 as shown in figure 3 is obtained, and the dynamic information of the variables on time distribution is obtained. In addition, the average value of the time sequence results is obtained, quantitative causal values among the variables are obtained, and quantitative indexes are provided for the subsequent determination of causal relations among the variables. Fig. 3 shows the dynamic profile of PM2.5 and air temperature over time, from which the dynamic relationship between the two variables over time is seen. If a specific period needs to be studied, more information needs to be obtained from the curve in detail; and obtaining the most relevant feature subset of the target variable according to the result of the causal analysis, then establishing a prediction model by using an echo state network, and checking the result of the causal relation. Wherein, the pool parameters of the echo state network are set as follows: pool dimension 100, sparsity 0.05, spectral radius 0.5, and connection input weight 0.05. A total of 10 independent experiments were performed, with the mean of 10 as the final result to eliminate the effect of error. The predictors include Root Mean Square Error (RMSE) and Symmetric Mean Absolute Percentage Error (SMAPE). The two are defined as follows:
wherein x i and Respectively representing a true value and a predicted value, and n represents the number of samples.
The prediction results are shown in fig. 4, 5 and 6. It can be seen from the figure that the variable subset selected by the method of the present invention enables the predicted PM2.5 results to even reflect a true trend of change. The errors of 10 independent replicates are shown in table 3. In continued contrast to the standard graininess causal analysis method (GC) and a two-stage causal network learning method (PCMCI), both the root mean square error and the mean absolute percentage error of symmetry obtained by the present invention gave the best results, which further illustrate the effectiveness of the present invention.
TABLE 3 prediction results of various methods
GC | PCMCI | The method of the invention | |
RMSE | 0.0189 | 0.0184 | 0.0183 |
SMAPE | 0.1132 | 0.1120 | 0.1018 |
The above examples are only for the purpose of illustrating the embodiments of the present invention and are not to be construed as limiting the scope of the invention, and it should be noted that modifications can be made by those skilled in the art without departing from the spirit of the invention, which is within the scope of the invention.
Claims (2)
1. The causal network learning method based on the local Granges causal analysis is characterized by comprising the following specific steps:
Step 1: acquiring an air quality index AQI and meteorological observation data; preprocessing the multidimensional AQI and meteorological time series data; adopting a unit root test method to carry out stability test on the time series data; then normalizing the time sequence data;
Step 2: determining the maximum hysteresis order tau max of the time series data obtained by the processing in the step 1 through a red pool information criterion, and generating a historical variable matrix X -;
Wherein, Data representing the current t moment, τ max is the maximum hysteresis order, d represents the variable number, and n represents the sample number;
Step 3: with the PC algorithm, with each variable Feature selection processing is carried out for the target to obtain each variable/>Related feature subset/>Wherein the variable/>For the data acquired at the time of t sampling, i=1, 2, …, d; each variable/>By its related feature subset/>The expression is represented by the following formula (2):
Wherein f i (. Cndot.) represents the mapping function, Representing the deviation;
step 4: related feature subset The set of conditions is sent to a local graininess causal analysis model to obtain quantitative causal relation values, specifically:
step 4.1: selecting a drive variable And target variable/>The most relevant feature subsets of the two are respectively found from a historical variable matrix X - of the time sequence through a feature selection method: /(I)And/>Variable/>The feature subset of (2) is smoothed one bit backwards to obtain the corresponding/>Is a subset of features of (a);
step 4.2: will drive the variable And target variable/>Is fed into a local glabellar causal analysis model as a condition set, as shown in formula (3):
Step 4.3: weighted summation to obtain driving variable For the target variable/>Results of causal analysis of (2);
Step 5: averaging results obtained by the local Grangel causal analysis to obtain quantitative causal relation values among variables, and drawing the results of the local Grangel causal analysis to obtain a dynamic causal relation curve among the variables;
step 6: the causal relation between any variable and other variables is obtained, and a driving variable is selected as a relevant influence factor of a target variable according to the causal relation result; and establishing a prediction model by using the echo state network for analysis to obtain a prediction analysis result of the target variable.
2. The causal network learning method based on local gracile causal analysis according to claim 1, wherein the local gracile causal analysis model further explores dynamic characteristics among variables based on causal analysis results of quantitative values; the local glanger causal analysis model is represented by formula (4):
Lgc(ut,wt,yt)=GC+γ(ut,wt,yt) (4)
Wherein GC is a standard gland cause and effect analysis; gamma (u t,wt,yt) is a function about u t,wt,yt to reveal dynamic characteristics between variables; u t denotes a set of condition variables, w t denotes a driving variable, and y t denotes a response variable; the GC and γ (u t,wt,yt) are represented using the second order statistics of the unified procedure:
Where |·| represents the determinant of the matrix, Is a historical state and current state observation; /(I) And/>Is a covariance matrix; the mean value of γ (u t,wt,yt) is 0, representing < L gc (t) > = GC; </cndot > represents time series averaging.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111304900.1A CN114036736B (en) | 2021-11-05 | 2021-11-05 | Causal network learning method based on local Granges causal analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111304900.1A CN114036736B (en) | 2021-11-05 | 2021-11-05 | Causal network learning method based on local Granges causal analysis |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114036736A CN114036736A (en) | 2022-02-11 |
CN114036736B true CN114036736B (en) | 2024-05-07 |
Family
ID=80136338
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111304900.1A Active CN114036736B (en) | 2021-11-05 | 2021-11-05 | Causal network learning method based on local Granges causal analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114036736B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114511087B (en) * | 2022-04-19 | 2022-07-01 | 四川国蓝中天环境科技集团有限公司 | Air quality space inference method and system based on double models |
CN115935283B (en) * | 2022-12-27 | 2023-07-25 | 南京信息工程大学 | Drought cause tracing method based on multi-element nonlinear causal analysis |
CN116187443B (en) * | 2023-02-10 | 2024-05-24 | 中国科学院自动化研究所 | Causal strength detection method and detection device based on multidimensional symbol dynamics |
CN116030941B (en) * | 2023-03-30 | 2023-08-25 | 北京科技大学 | Alzheimer's disease diagnosis method based on edge-centric effect connection network |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101382507B1 (en) * | 2012-10-19 | 2014-04-10 | 사단법인대기환경모델링센터 | Air quality forecast and management system |
CN109033178A (en) * | 2018-06-26 | 2018-12-18 | 北京工业大学 | A method of excavating Granger causality between visibility multidimensional space-time data |
CN111367959A (en) * | 2020-02-17 | 2020-07-03 | 大连理工大学 | Zero-time-lag nonlinear expansion Granger causal analysis method |
WO2020206466A1 (en) * | 2019-03-07 | 2020-10-08 | Wismuller Axel W E | Method and device for determining a measure of causal influence between components of complex systems |
CN112101480A (en) * | 2020-09-27 | 2020-12-18 | 西安交通大学 | Multivariate clustering and fused time sequence combined prediction method |
-
2021
- 2021-11-05 CN CN202111304900.1A patent/CN114036736B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101382507B1 (en) * | 2012-10-19 | 2014-04-10 | 사단법인대기환경모델링센터 | Air quality forecast and management system |
CN109033178A (en) * | 2018-06-26 | 2018-12-18 | 北京工业大学 | A method of excavating Granger causality between visibility multidimensional space-time data |
WO2020206466A1 (en) * | 2019-03-07 | 2020-10-08 | Wismuller Axel W E | Method and device for determining a measure of causal influence between components of complex systems |
CN111367959A (en) * | 2020-02-17 | 2020-07-03 | 大连理工大学 | Zero-time-lag nonlinear expansion Granger causal analysis method |
CN112101480A (en) * | 2020-09-27 | 2020-12-18 | 西安交通大学 | Multivariate clustering and fused time sequence combined prediction method |
Non-Patent Citations (2)
Title |
---|
基于格兰杰因果方法的注意脑电网络分析;袁勤;蒋涛;生物医学工程学杂志;20161231;第33卷(第1期);全文 * |
面向PM_(2.5)预测的递归随机森林与多层神经网络集成模型;蒋洪迅;田嘉;孙彩虹;;系统工程;20201231(第05期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN114036736A (en) | 2022-02-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN114036736B (en) | Causal network learning method based on local Granges causal analysis | |
US10830743B2 (en) | Determining the net emissions of air pollutants | |
Chen et al. | Predict the effect of meteorological factors on haze using BP neural network | |
CN111047012A (en) | Air quality prediction method based on deep bidirectional long-short term memory network | |
CN111367959B (en) | Zero-time-lag nonlinear expansion Granger causal analysis method | |
CN105574669B (en) | Power transmission and transformation equipment state abnormity detection method based on time-space joint data clustering analysis | |
CN112101480A (en) | Multivariate clustering and fused time sequence combined prediction method | |
CN112489497B (en) | Airspace operation complexity evaluation method based on deep convolutional neural network | |
CN113240170A (en) | Air quality prediction method based on seasonal cyclic neural network | |
Doran et al. | Fitting value-added models in R | |
CN111814956A (en) | Multi-task learning air quality prediction method based on multi-dimensional secondary feature extraction | |
CN116416884B (en) | Testing device and testing method for display module | |
CN110555570A (en) | Intelligent prediction method and device for gas concentration of mine limited space disaster | |
CN113361690A (en) | Water quality prediction model training method, water quality prediction device, water quality prediction equipment and medium | |
CN114217025B (en) | Analysis method for evaluating influence of meteorological data on air quality concentration prediction | |
CN114676779A (en) | Atmospheric quality monitoring data calibration method and system based on robust width network | |
Zhan et al. | A decomposition-ensemble broad learning system for AQI forecasting | |
Haas | Statistical assessment of spatio-temporal pollutant trends and meteorological transport models | |
CN117591950A (en) | Rolling bearing variable working condition fault diagnosis method, device, terminal and storage medium | |
CN116580123A (en) | Regional atmospheric environment photochemical grid model image generation system | |
CN115587650A (en) | Multi-target hybrid prediction method for normal atmospheric pollutants in medium-short term and time intervals | |
CN113225391B (en) | Atmospheric environment monitoring quality monitoring method based on sliding window anomaly detection and computing equipment | |
CN115510763A (en) | Air pollutant concentration prediction method and system based on data-driven exploration | |
Ryoo et al. | Nonlinear growth mixture models with fractional polynomials: an illustration with early childhood mathematics ability | |
CN110780604A (en) | Space-time signal recovery method based on space-time smoothness and time correlation |
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 |