CN113887635B - Basin similarity classification method and classification device - Google Patents
Basin similarity classification method and classification device Download PDFInfo
- Publication number
- CN113887635B CN113887635B CN202111170738.9A CN202111170738A CN113887635B CN 113887635 B CN113887635 B CN 113887635B CN 202111170738 A CN202111170738 A CN 202111170738A CN 113887635 B CN113887635 B CN 113887635B
- Authority
- CN
- China
- Prior art keywords
- basin
- similarity
- average
- preset
- area
- 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
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 27
- 230000004927 fusion Effects 0.000 claims abstract description 19
- 238000005259 measurement Methods 0.000 claims abstract description 8
- 238000013316 zoning Methods 0.000 claims abstract description 5
- 210000002569 neuron Anatomy 0.000 claims description 36
- 238000012549 training Methods 0.000 claims description 22
- 238000009736 wetting Methods 0.000 claims description 22
- 238000013528 artificial neural network Methods 0.000 claims description 17
- 239000002689 soil Substances 0.000 claims description 16
- 238000005192 partition Methods 0.000 claims description 12
- 238000001556 precipitation Methods 0.000 claims description 12
- 150000001875 compounds Chemical class 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 10
- 239000004576 sand Substances 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 7
- 239000004927 clay Substances 0.000 claims description 7
- 238000001704 evaporation Methods 0.000 claims description 6
- 230000008020 evaporation Effects 0.000 claims description 6
- 230000000694 effects Effects 0.000 claims description 5
- 238000011156 evaluation Methods 0.000 claims description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 4
- 230000010354 integration Effects 0.000 claims description 3
- 239000000463 material Substances 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 238000000638 solvent extraction Methods 0.000 claims description 3
- 239000002245 particle Substances 0.000 claims description 2
- 238000011161 development Methods 0.000 abstract description 4
- 230000018109 developmental process Effects 0.000 abstract description 4
- 230000007261 regionalization Effects 0.000 abstract description 3
- 238000002054 transplantation Methods 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 2
- 238000013139 quantization Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000006424 Flood reaction Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/043—Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Operations Research (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Automation & Control Theory (AREA)
- Fuzzy Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Cultivation Of Plants (AREA)
Abstract
The invention discloses a basin similarity classification method and a classification device, wherein the classification method comprises the following steps: collecting hydrological, meteorological and underlying surface data of a preset basin, extracting meteorological and underlying surface factors, and constructing characteristic indexes; setting a correlation threshold value to determine a weather and underlying surface similarity index; performing grid scale hydrological meteorological zoning in a preset basin based on a SOM and FCM fusion algorithm; constructing a sub-basin scale clustering set in a meteorological homogeneous region, and classifying the sub-basins by using a fusion model; according to the weather subareas where the drainage basins are located and the sub-drainage basin types contained in the drainage basins, fuzzy membership degrees of the sub-drainage basins are utilized to construct drainage basin similarity comprehensive measurement, the similarity degree between the drainage basins is evaluated by using a maximum and minimum closeness method, and the similarity between the drainage basins is identified. The invention provides theoretical basis for parameter transplantation and parameter selection of the data-free basin, and promotes the deep development of the work of regionalization of model parameters, hydrological early warning and forecasting and the like of the data-free area.
Description
Technical Field
The invention belongs to the technical field of hydrology, and particularly relates to a basin similarity classification method and a basin similarity classification device, which are mainly used for similar basin discrimination and hydrology early warning and forecasting and the like in data-free areas.
Background
In recent years, under the influence of climate change and human activities, extreme weather events show an increasing and strengthening trend, and serious water disaster disasters such as mountain floods, debris flows and the like are frequent. Particularly for the regions with deficient hydrologic data, due to the lack of effective hydrologic monitoring means, the detection of disasters is insufficient, the disasters happen rapidly, the damage is serious, and the sustainable development of the economy and the society is severely restricted. Aiming at the hydrologic forecast problem of the data-free area, the international hydrologic scientific association (IAHS) started the PUB international hydrologic program in 2002, and the program aims to explore a new method for converting the hydrologic work of the data-free area from a traditional method for calibrating observed data into a mechanism analysis. The method for determining model parameters of a data-free area is a regionalization method, namely, a data basin similar to a target basin is selected, and the model parameters of the data-free basin are calculated by utilizing the calibrated model parameters of the data basin, so that hydrologic prediction of the target basin is realized.
With the development of technologies such as remote sensing and GIS, climate and underlying surface data are easy to obtain, the hydrologic similarity is calculated in an approximate mode by mining effective information of the structure characteristics of the non-data basin and utilizing the similarity of the climate and the basin characteristics, and the possibility is provided for runoff prediction of the non-data areas. In recent years, computer technology is rapidly developed, pattern recognition, clustering algorithm and fuzzy mathematics are greatly developed from research theory to practical application, and technical support is provided for hydrologic similarity research. Therefore, the comprehensive criterion for establishing the similarity between the basin weather and the underlying surface and the method for establishing the high-dimensional heterogeneous characteristic space hydrologic similarity area identification have certain theoretical basis and practical significance.
The selection of the traditional similar drainage basin usually depends on the experience of a decision maker, has no uniform quantization standard and has stronger uncertainty in practical application. Meanwhile, a general similarity index combination is not formed at present, a specific similarity index may not capture comprehensive hydrologic rules, and how to select the similarity index and construct a similar basin evaluation system becomes a key step in hydrologic similarity and parameter regional research. Therefore, the present invention provides a general watershed similarity classification method and a classification apparatus.
Disclosure of Invention
The invention aims to provide a watershed similarity classification method and a watershed similarity classification device, which provide a theoretical basis for selection and parameter transplantation of similar watersheds in a data-free area.
In order to achieve the purpose, the invention adopts the following technical scheme: a basin similarity classification method is characterized by comprising the following steps:
step 1, collecting hydrological, meteorological and underlying surface data of a preset watershed, extracting meteorological and underlying surface factors, and constructing characteristic indexes;
step 3, constructing a grid scale cluster based on the SOM and FCM fusion algorithm, and carrying out hydrological meteorological zoning in a preset basin;
and 5, constructing a basin similarity comprehensive measurement by using the fuzzy membership degree of the sub-basins according to the meteorological subareas where the basins are located and the sub-basin types contained in the meteorological subareas. And (4) evaluating the similarity degree between the flow domains by using a maximum and minimum closeness method, and identifying the similarity between the flow domains.
The step 1 comprises the following steps:
step 11, collecting hydrological, meteorological and underlying surface data of a preset watershed, wherein the hydrological data comprise runoff coefficients and multi-year average daily runoff of the preset watershed, the meteorological data comprise multi-year average monthly precipitation, multi-year average monthly potential evaporation and multi-year average monthly temperature data of the preset watershed, and the underlying surface data comprise topographic characteristics, vegetation types, soil types and land utilization types of the preset watershed;
Im,r=max(MI(1,2,...12))-min(MI(1,2,...12))
Tm,r=max(T(1,2,...12))-min(T(1,2,...12))
wherein MI (t) is the wetting index at month t; i ism,Im,r,fs,Tm,Tm,r,DsRespectively representing the annual average wetting index, the annual monthly wetting index maximum difference, the snow falling ratio, the annual average air temperature, the annual monthly air temperature maximum difference and the snow falling time ratio; p (t), EP(t), t (t) is the annual average precipitation in month tth, the annual average potential evaporation in month tth and the annual average temperature value in month tth, respectively; d (t) days of month t; t is0For a temperature threshold, the form of precipitation below this temperature is snow, which is set by this patent to be 0 ℃.
And step 13, extracting the following underlying surface factors as characteristic indexes including topographic characteristic factors, soil vegetation characteristic factors and shape characteristic factors according to the underlying surface information of the preset watershed. The terrain characterization factors include an average elevation, a maximum elevation difference, an area-elevation integral, an area elevation curve slope, an average terrain index, and an average grade of the predetermined watershed. The method comprises the following steps that (1) area-elevation integration represents the quality of earth surface materials in a preset drainage basin, the slope of an area-elevation curve reflects the topographic relief degree of the preset drainage basin, and the area-elevation integral and the topographic relief degree are calculated through the area-elevation curve; the average terrain index is an arithmetic average of unit grid terrain indexes in a preset basin, and the calculation formula is as follows:
in the formula, HI、ASAnd TI is an area-elevation integral value, an area-elevation curve slope and an average topographic index respectively; the area-elevation curve f (x) is a fitting curve consisting of x ═ a/A and y ═ H/H, and a is the area above a certain contour line in a predetermined flow domain; h is the height difference between the contour line and the lowest point in the preset drainage basin; a is the total area of the preset drainage basin; h is the maximum relative height difference in the preset basin; f (0.2) and f (0.8) represent the corresponding relative height differences on the area-elevation curve at area-to-elevation ratios of 0.2 and 0.8, respectively; a isiThe water collecting area of the ith unit grid in the preset drainage basin is obtained; beta is aiIs the slope of the ith unit grid; n is the total number of grids in the predetermined flow domain;
the soil vegetation characteristic factors comprise the sand grain content, the particle content, the clay grain content and the normalized vegetation index of the watershed soil; the shape characteristic factors comprise the river basin area, the river basin length, the river basin form factor, the river basin elongation ratio and the river basin and river network density.
The step 2 comprises the following steps:
the method for selecting the distance correlation coefficient to select the characteristic indexes of the weather and underlying surface factors and setting the correlation threshold value to determine the similarity indexes of the weather and underlying surface comprises the following steps:
calculating distance correlation coefficients between different factors:
in the formula, X is a certain underlying surface or meteorological characteristic factor sequence, and Y is a hydrological characteristic factor sequence; dCor (X, Y) is the distance correlation coefficient between the X and Y sequences; dCov (X, Y) is the distance covariance of X and Y sequences; dVar (X) and dVar (Y) are respectively the standard deviation of the distance between the X sequence and the Y sequence;
and setting a threshold value omega, and clustering by using the underlying surface or meteorological characteristic factor with the distance correlation coefficient dCor (X, Y) > omega as a similarity index. The threshold omega is selected to be 0.5, and similarity indexes determined according to the distance correlation coefficient are respectively as follows:
the annual average precipitation, the annual average wetting index and the annual monthly wetting index maximum difference and other 3 meteorological indexes; and 7 underlay surface indexes such as the sand grain content of the soil in the drainage basin, the clay grain content of the drainage basin, the normalized vegetation index of the drainage basin, the area of the drainage basin, the length of the drainage basin, the form factor of the drainage basin, the elongation ratio of the drainage basin and the like.
The step 3 comprises the following steps:
a hydrological meteorological clustering set is constructed on a grid scale, a SOM and FCM fusion model is used for training and partitioning, and the clustering process is divided into two stages: the first stage, carrying out primary clustering by using an SOM algorithm, and obtaining a competition output layer after training is finished; and in the second stage, the corresponding weight vector of the node of the SOM output layer is used as a clustering sample of the FCM algorithm, and iterative computation is carried out until a convergence condition is reached.
In step 3, the clustering process based on the SOM-FCM fusion algorithm is as follows:
step 31, initializing the SOM neural network: setting the competition layer structure of the SOM neural network, the initial domain radius delta (0) and the initial learning rateNumber of iterations k and total number of iterations ks(ii) a Carrying out normalization processing on the N-dimensional similarity index to obtain a training sample G; setting weight vector W corresponding to each neuron in competition layerj(k)=(Wj,1(k),Wj,2(k),...,Wj,N(k) (j ═ 1, 2.. times, M), initializing a weight vector, where M is the number of contention layer neurons, and k has an initial value of 0;
step 32, inputting training samples: randomly selecting the ith sample Gi=(G1,G2,...,GN)TTo the input layer;
step 33, search for the winning neuron and calculate GiAnd Wj(k) Selecting the neuron with the smallest distance as a winning neuron r;
step 34, adjusting the neuron connection weight: connected weight vector W for neurons in neighborhood of winning neuron rjAdjusting;
in the formula (I), the compound is shown in the specification,is the learning rate;is the radius of the field of the winning neuron r;
step 36, repeating steps 32-35 until the training iteration count k reaches the preset total number k of iterationssObtaining a neural network of an SOM output layer;
step 37, using SOM to output the weight W of each neuron of the neural network of the layerjSetting the clustering number c, membership factor m, limited error epsilon and maximum iteration number k of the FCM algorithm as the input vector of the FCM algorithmfInitializing a membership matrix U, and setting the iteration number k as 0;
step 38, calculating a clustering center according to the membership matrix U, wherein the ith clustering center CiComprises the following steps:
in the formula (I), the compound is shown in the specification,is an input vector WjFor CiDegree of membership of;
and 39, updating the membership matrix U according to the clustering center:
in the formula (d)ijIs an input vector WjFor CiThe Euclidean distance of; dqjIs an input vector WjFor C q1,2, c;
step 310, training an iteration counter k ═ k + 1;
step 311, repeat steps 39-310 until | U(k)-U(k+1)| < epsilon or the iteration number k reaches the maximum iteration number kf。
In step 3, the SOM network structure is selected according to the principle of minimum error, quantitative error QE and topological error TE are selected for evaluation, and two indexes can express the quality of SOM neural network clustering:
TE=∑v(Gi)
where QE is the average relative distance between the input sample and the corresponding winning neuron, Wr(Gi) Is GiThe weight vector of the corresponding winning neuron; TE is the degree of proximity of samples adjacent to the input space in the competition layer network, if the samples are GiThe neighboring samples remain adjacent in the output space, v (G)i) Is 1.
In step 3, the clustering number c of the SOM-FCM algorithm is determined according to the Davies-Bouldin index (DBI), and the clustering effect under different clustering numbers is evaluated to optimize the c:
wherein c is the number of clusters; miThe number of samples belonging to the category i; ciAs the cluster center of class i, CqA cluster center of class q, wherein i 1,2,.., c, q 1,2,.., c;class i and class q sample points to C, respectivelyiAnd CqThe average distance of (c).
And step 5, constructing a basin similarity comprehensive measurement by using the fuzzy membership degree of the sub-basins according to the weather subareas where the basins are located and the included sub-basin categories. The similarity degree between the flow domains is evaluated by using a maximum and minimum closeness method, and the flow domain B1And basin B2The similarity of (a) is expressed as:
in the formula (I), the compound is shown in the specification,is a drainage basin B1And basin B2The similarity of (2);represents basin B1And B2Weighting membership degrees for the areas of the basin class i;andrespectively is a drainage basin B1And B2The ratio of the area of the jth sub-basin to the total area of the basin; n is a radical ofB1And NB2Is a drainage basin B1And B2The number of sub-watersheds in (a);andrepresents basin B1And B2Fuzzy membership degree of the jth sub-basin to the basin class i; the A and V respectively represent the calculation of the minimum value and the maximum value.
The device for classifying the basin similarity is characterized by comprising a processor and a memory; stored in the memory are programs or instructions that are loaded and executed by the processor to implement the steps of the classification method according to any one of claims 1 to 9.
The invention has the following beneficial results: according to the basin similarity classification method and the classification device, the basin similarity index system is constructed by calculating the distance correlation coefficient among basin factors; the SOM and FCM fusion model is adopted to divide the hydrometeorology homogeneous region, and on the basis, the fusion model is utilized to realize sub-basin classification; and constructing a basin similarity comprehensive measurement by combining the maximum and minimum closeness method with the fuzzy membership of the sub-basins, and realizing comprehensive evaluation of the similarity degree between the basins. The invention adopts a model combining SOM and FCM to perform hydrological meteorological partition and sub-basin clustering, the fusion model has the characteristics of SOM algorithm self-organization and strong nonlinear mapping, and simultaneously, the concept of fuzzy set is fused, and nonlinear and highly heterogeneous hydrological data with difficult-to-define boundary can be explained. Meanwhile, the optimal clustering number is selected based on the internal indexes of the clusters, so that the high efficiency and the objective stability of the clustering result are ensured. The similarity between the drainage basins is expressed through the fuzzy membership degree, an objective quantitative criterion is provided for the identification of the similar drainage basins, a theoretical basis can be provided for the selection and parameter transplantation of the non-data drainage basin parametrization drainage basins, and the deep development of the work of model parameter regionalization, hydrological early warning and forecasting and the like of the non-data areas is promoted.
Drawings
Fig. 1 is a schematic diagram of a computing flow of a basin similarity classification method provided by the present invention;
FIG. 2 is a graph of the correlation coefficient of the distance between the hydrological features and the features of the meteorological and underlying surfaces in an exemplary embodiment;
FIG. 3 is a test chart of the structure of the SOM neural network in the hydrometeorology partition in the embodiment;
FIG. 4 is a diagram of the hydrometeorology partition number-Davies-Bouldin index change relationship in the embodiment;
FIG. 5 is a diagram of the relationship between the cluster number of the sub-domains-Davies-Bouldin index change in the embodiment;
fig. 6 is a diagram illustrating exemplary inter-basin similarity in an exemplary embodiment.
Detailed Description
The invention is further described with reference to the accompanying drawings and specific examples.
It should be understood that the detailed description and specific examples, while indicating the invention, are intended for purposes of illustration only and are not intended to limit the scope of the invention.
As shown in fig. 1, the present invention provides a watershed similarity classification method, which includes the following steps:
step 1, collecting hydrological, meteorological and underlying surface data of a preset watershed, extracting meteorological and underlying surface factors, and constructing characteristic indexes;
step 3, constructing a grid scale clustering set based on a SOM and FCM fusion algorithm, and carrying out hydrological meteorological zoning in a preset basin;
and 5, constructing a basin similarity comprehensive measurement by utilizing the fuzzy membership degree of the sub-basins according to the weather subareas where the basins are located and the included sub-basin categories. And (4) evaluating the similarity degree between the drainage domains by using a maximum and minimum closeness method, and identifying the similarity between the drainage domains.
Taking 10 drainage basins of Tunxi, Shuishiang, Ma Shuwang, banqiao, big river dam, old river, camp, big pavilion, Shidan and Suide as typical drainage basins, and carrying out small drainage basin classification and drainage basin similarity research. The method specifically comprises the following steps:
step 1, collecting hydrological, meteorological and underlying surface data of a preset basin, extracting meteorological and underlying surface factors, and constructing characteristic indexes;
the method comprises the following steps:
step 11, collecting hydrological, meteorological and underlying surface data of a preset watershed, wherein the hydrological data comprise runoff coefficients and multi-year average daily runoff of the preset watershed, the meteorological data comprise multi-year average monthly precipitation, multi-year average monthly potential evaporation and multi-year average monthly temperature data of the preset watershed, and the underlying surface data comprise topographic characteristics, vegetation types, soil types and land utilization types of the preset watershed;
Im,r=max(MI(1,2,...12))-min(MI(1,2,...12)) (3)
Tm,r=max(T(1,2,...12))-min(T(1,2,...12)) (6)
where MI (t) is the wet index at month t; i ism,Im,r,fs,Tm,Tm,r,DsRespectively representing the annual average wetting index, the annual monthly wetting index maximum difference, the snow falling ratio, the annual average air temperature, the annual monthly air temperature maximum difference and the snow falling time ratio; p (t), EP(t), t (t) is the annual average precipitation in month tth, the annual average potential evaporation in month tth and the annual average temperature value in month tth, respectively; d (t) days of month t; t is0For a temperature threshold, the form of precipitation below this temperature is snow, which is set by this patent to be 0 ℃.
And step 13, extracting the following underlying surface factors as characteristic indexes including topographic characteristic factors, soil vegetation characteristic factors and shape characteristic factors according to the underlying surface information of the preset watershed. The topographic characteristic factor includes an average elevation (H) of the predetermined drainage basinm) Maximum height difference (H)r) Area-elevation integral (HI), area-elevation slope (A)s) Average terrain index (Ti) and average slope (β). The method comprises the following steps that (1) area-elevation integration represents the quality of earth surface materials in a preset drainage basin, the slope of an area-elevation curve reflects the topographic relief degree of the preset drainage basin, and the area-elevation integral and the topographic relief degree are calculated through the area-elevation curve; the average terrain index is an arithmetic average of unit grid terrain indexes in a preset basin, and the calculation formula is as follows:
in the formula (I), the compound is shown in the specification,HI、ASand TI is an area-elevation integral value, an area-elevation curve slope and an average topographic index respectively; the area-elevation curve f (x) is a fitting curve consisting of x ═ a/A and y ═ H/H, and a is the area above a certain contour line in a predetermined flow domain; h is the height difference between the contour line and the lowest point in the preset drainage basin; a is the total area of the preset drainage basin; h is the maximum relative height difference in the preset basin; f (0.2) and f (0.8) represent the corresponding relative height differences on the area-elevation curve at area-to-elevation ratios of 0.2 and 0.8, respectively; a isiThe water collecting area of the ith unit grid in the preset drainage basin is obtained; beta is aiIs the slope of the ith unit grid; n is the total number of grids in the predetermined flow domain;
the soil vegetation characteristic factors comprise drainage basin soil Sand content (Sand), drainage basin powder content (Clay), drainage basin Clay content (Silt) and drainage basin normalized vegetation index (NDVI); the shape feature factors include a watershed area (a), a watershed length (L), a watershed form factor (Rf), a watershed elongation ratio (Re), and a watershed network density (Rd).
the method for selecting the distance correlation coefficient to select the characteristic indexes of the weather and underlying surface factors and setting the correlation threshold value to determine the similarity indexes of the weather and underlying surface comprises the following steps:
calculating distance correlation coefficients between different factors:
in the formula, X is a certain underlying surface or meteorological characteristic factor sequence, and Y is a hydrological characteristic factor sequence; dCor (X, Y) is the distance correlation coefficient between the X and Y sequences; dCov (X, Y) is the distance covariance of X and Y sequences; dVar (X) and dVar (Y) are the standard deviation of the distance between the X sequence and the Y sequence, respectively.
And setting a threshold value omega, and clustering by using the underlying surface or meteorological characteristic factor with the distance correlation coefficient dCor (X, Y) > omega as a similarity index. The threshold omega is selected to be 0.5, and similarity indexes determined according to the distance correlation coefficient are respectively as follows:
the annual average precipitation, the annual average wetting index and the annual monthly wetting index maximum difference and other 3 meteorological indexes; and 7 underlying surface indexes such as the sand grain content of the watershed soil, the clay grain content of the watershed, the watershed normalized vegetation index, the watershed area, the watershed length, the watershed form factor, the watershed elongation ratio and the like.
Step 3, constructing a grid scale clustering set based on a SOM and FCM fusion algorithm, and carrying out hydrological meteorological zoning in a preset basin;
a hydrological meteorological clustering set is constructed on a grid scale, a SOM and FCM fusion model is used for training and partitioning, and the clustering process is divided into two stages: the first stage, carrying out primary clustering by using an SOM algorithm, and obtaining a competition output layer after training is finished; and in the second stage, the corresponding weight vector of the node of the SOM output layer is used as a clustering sample of the FCM algorithm, and iterative computation is carried out until a convergence condition is reached.
The clustering process based on the SOM-FCM fusion algorithm comprises the following steps:
step 31, initializing the SOM neural network: setting the competition layer structure of the SOM neural network, the initial domain radius delta (0) and the initial learning rateNumber of iterations k and total number of iterations ks(ii) a Carrying out normalization processing on the N-dimensional similarity index to obtain a training sample G; setting weight vector W corresponding to each neuron in competition layerj(k)=(Wj,1(k),Wj,2(k),...,Wj,N(k) (j ═ 1, 2.. times, M), initializing a weight vector, where M is the number of contention layer neurons, and k has an initial value of 0;
step 32, inputting training samples: randomly selecting the ith sample Gi=(G1,G2,...,GN)TTo the input layer;
step 33, search for the winning neuron and calculate GiAnd Wj(k) Distance between, distance selectedThe smallest neuron is taken as a winning neuron r;
step 34, adjusting the neuron connection weight, and carrying out connection weight vector W on neurons in the neighborhood of the winning neuron rjAdjusting;
in the formula (I), the compound is shown in the specification,is the learning rate;is the radius of the field of the winning neuron r;
step 36, repeating steps 32-35 until the training iteration count k reaches the preset total number k of iterationssObtaining a neural network of an SOM output layer;
step 37, using SOM to output the weight W of each neuron of the neural network of the layerjSetting the clustering number c, membership factor m, limited error epsilon and maximum iteration number k of the FCM algorithm as the input vector of the FCM algorithmfInitializing a membership matrix U, and setting the iteration number k as 0;
step 38, calculating a clustering center according to the membership matrix U, wherein the ith clustering center CiComprises the following steps:
in the formula (I), the compound is shown in the specification,is an input vector WjFor CiDegree of membership of;
and 39, updating the membership matrix U according to the clustering center:
in the formula, dijIs an input vector WjFor CiThe Euclidean distance of;
step 310, training an iteration counter k ═ k + 1;
step 311, repeat steps 39-310 until | U(k)-U(k+1)| < epsilon or the iteration number k reaches the maximum iteration number kf。
The SOM network structure in the SOM-FCM algorithm is selected according to the principle of minimum error, quantitative error QE and topological error TE are selected for evaluation, and two indexes can express the quality of SOM neural network clustering:
TE=∑v(Gi) (17)
where QE is the average relative distance between the input sample and the corresponding winning neuron, Wr(Gi) Is GiA weight vector of the corresponding winning neuron; TE is the degree of proximity of samples adjacent to the input space in the competition layer network, if the samples are GiThe neighboring samples remain adjacent in the output space, v (G)i) Is 1.
Fig. 3 shows the test result of the hydrometeorology partition SOM neural network structure, and the number of the finally selected model output nodes is 14 × 22.
The clustering number c of the SOM-FCM algorithm is determined according to the Davies-Bouldin index (DBI), and c is optimized by evaluating clustering effects under different clustering numbers:
wherein c is the number of clusters; m is a group ofiThe number of samples belonging to the category i; ciClustering centers for class i, CqA cluster center of class q, wherein i 1,2,.., c, q 1,2,.., c;class i and class q sample points to C, respectivelyiAnd CqThe average distance of (c).
And (4) optimizing k, wherein when k is 7, DBI reaches the minimum value, and the optimization process of the hydrometeorology partition number k is shown in FIG. 4.
and classifying the sub-basin by using an SOM-FCM algorithm, selecting a SOM network structure by selecting a quantization error QE and a topology error TE according to a minimum error principle, evaluating clustering effects under different clustering numbers according to the DBI, and optimizing k. When k is 9, DBI reaches the minimum value, and the preferred process of the sub-basin classification number k is shown in fig. 5.
And 5, constructing a basin similarity comprehensive measurement by utilizing the fuzzy membership degree of the sub-basins according to the weather subareas where the basins are located and the sub-basins contained in the weather subareas. The similarity between the drainage domains is evaluated by using a maximum and minimum closeness method, the similarity between the drainage domains is identified, the drainage domain with a large pavilion is taken as a target drainage domain, and the typical similarity between the drainage domains is shown in fig. 6.
The similarity degree between the flow domains is evaluated by using a maximum and minimum closeness method, and the flow domain B1And basin B2The similarity of (a) is expressed as:
in the formula (I), the compound is shown in the specification,is a drainage basin B1And basin B2The similarity of (2);represents basin B1And B2Weighting the membership degree of the area of the basin class i;andrespectively is a drainage basin B1And B2The ratio of the area of the jth sub-basin to the total area of the basin; n is a radical ofB1And NB2Is a drainage basin B1And B2The number of sub-watersheds in (a);andrepresents basin B1And B2The fuzzy membership degree of the jth sub-basin to the basin class i; the A and V respectively represent the calculation of the minimum value and the maximum value.
The embodiment of the invention also provides a basin similarity classification device, which comprises a processor and a memory; the memory has stored therein a program or instructions that is loaded and executed by the processor to implement the steps of the classification method in the above embodiments.
Claims (9)
1. A watershed similarity classification method is characterized by comprising the following steps:
step 1, collecting hydrological, meteorological and underlying surface data of a preset basin, extracting meteorological and underlying surface factors, and constructing characteristic indexes;
step 2, selecting characteristic indexes by using distance correlation coefficients of the hydrological characteristics of the preset basin and weather and underlying surface factors, and setting a correlation threshold value to determine weather and underlying surface similarity indexes;
step 3, constructing a grid scale clustering set based on a SOM and FCM fusion algorithm, and carrying out hydrological meteorological zoning in a preset basin;
step 4, constructing a sub-basin scale clustering set in a meteorological homogeneous area based on the hydrometeorology partition result, and classifying the sub-basins by using a fusion model;
step 5, constructing a basin similarity comprehensive measurement by utilizing the fuzzy membership degree of the sub-basins according to the meteorological partition where the basin is located and the sub-basin categories contained in the meteorological partition; evaluating the similarity degree between the river domains by using a maximum and minimum closeness method, and identifying the similarity between the river domains;
in the step 1, the constructed characteristic indexes comprise an annual average wetting index, a annual monthly wetting index maximum difference, a snowfall proportion, an annual average air temperature, an annual monthly air temperature maximum difference and a snowfall time proportion, a terrain characteristic factor, a soil vegetation characteristic factor and a shape characteristic factor; the terrain characteristic factors comprise the average elevation, the maximum elevation difference, the area-elevation integral, the area-elevation curve slope, the average terrain index and the average gradient of a preset drainage basin;
in step 2, selecting distance correlation coefficients to select characteristic indexes of weather and underlying surface factors, and setting a correlation threshold value to determine weather and underlying surface similarity indexes, wherein the method comprises the following steps:
calculating distance correlation coefficients between different factors:
in the formula, X is a certain underlying surface or meteorological characteristic factor sequence, and Y is a hydrological characteristic factor sequence; dCor (X, Y) is the distance correlation coefficient between the X and Y sequences; dCov (X, Y) is the distance covariance of X and Y sequences; dVar (X) and dVar (Y) are respectively the standard deviation of the distance between the X sequence and the Y sequence;
setting a threshold value omega, and clustering by using an underlying surface or meteorological characteristic factor with a distance correlation coefficient dCor (X, Y) > omega as a similarity index.
2. The method of claim 1, wherein step 1 comprises:
step 11, collecting hydrological, meteorological and underlying surface data of a preset watershed, wherein the hydrological data comprise runoff coefficients and multi-year average daily runoff of the preset watershed, the meteorological data comprise multi-year average monthly precipitation, multi-year average monthly potential evaporation and multi-year average monthly temperature data of the preset watershed, and the underlying surface data comprise topographic characteristics, vegetation types, soil types and land utilization types of the preset watershed;
step 12, according to the meteorological data of the preset basin, calculating the following meteorological factors as characteristic indexes: the average annual wetting index, the maximum difference of annual monthly wetting indexes, the snow falling proportion, the average annual temperature, the maximum difference of annual monthly temperature and the snow falling time proportion;
Im,r=max(MI(1,2,...12))-min(MI(1,2,...12))
Tm,r=max(T(1,2,...12))-min(T(1,2,...12))
wherein MI (t) is the wetting index at month t; I.C. Am,Im,r,fs,Tm,Tm,r,DsRespectively representing the annual average wetting index, the annual monthly wetting index maximum difference, the snow falling ratio, the annual average air temperature, the annual monthly air temperature maximum difference and the snow falling time ratio; p (t), EP(t), t (t) is the annual average precipitation in month tth, the annual average potential evaporation in month tth and the annual average temperature value in month tth, respectively; d (t) days of month t; t is a unit of0At a temperature threshold below which precipitation is in the form of snow, T0Is 0 ℃;
step 13, extracting the following underlying surface factors as characteristic indexes including topographic characteristic factors, soil vegetation characteristic factors and shape characteristic factors according to the underlying surface information of the preset watershed; the terrain characteristic factors comprise the average elevation, the maximum elevation difference, the area-elevation integral, the area-elevation curve slope, the average terrain index and the average gradient of a preset drainage basin; the method comprises the following steps that (1) area-elevation integration represents the quality of earth surface materials in a preset drainage basin, the slope of an area-elevation curve reflects the topographic relief degree of the preset drainage basin, and the area-elevation integral and the topographic relief degree are calculated through the area-elevation curve; the average terrain index is an arithmetic average of unit grid terrain indexes in a preset basin, and the calculation formula is as follows:
in the formula, HI and ASAnd TI is an area-elevation integral, an area-elevation curve slope and an average topographic index respectively; the area-elevation curve f (x) is a fitting curve consisting of x ═ a/A and y ═ H/H, and a is the area above a certain contour line in a predetermined flow domain; h is the height difference between the contour line and the lowest point in the preset drainage basin; a is the total area of the preset drainage basin; h is the maximum relative height difference in the preset basin; f (0.2) and f (0.8) represent the corresponding relative height differences on the area-elevation curve at area-to-elevation ratios of 0.2 and 0.8, respectively; a isiThe water collecting area of the ith unit grid in the preset drainage basin is obtained; beta is a betaiIs the slope of the ith unit grid; n is the total number of grids in the predetermined flow domain;
the soil vegetation characteristic factors comprise the sand grain content, the particle content, the clay grain content and the normalized vegetation index of the watershed soil; the shape characteristic factors include a watershed area, a watershed length, a watershed form factor, a watershed elongation ratio and a watershed river network density.
3. The method of claim 1,
the threshold value omega is selected to be 0.5, and the similarity index determined according to the distance correlation coefficient is as follows:
the maximum difference between the annual average precipitation, the annual average wetting index and the annual monthly wetting index is 3 meteorological indexes; the drainage basin soil sand grain content, the drainage basin clay grain content, the drainage basin normalized vegetation index, the drainage basin area, the drainage basin length, the drainage basin form factor and the drainage basin elongation ratio are 7 underlying surface indexes.
4. The method according to claim 1, wherein in the step 3, a hydrological meteorological clustering set is constructed on a grid scale, a SOM and FCM fusion model is used for training partitioning, and the clustering process is divided into two stages: the first stage, carrying out primary clustering by using an SOM algorithm, and obtaining a competition output layer after training is finished; and in the second stage, the SOM output layer node corresponding weight vector is used as a clustering sample of the FCM algorithm, and iterative computation is carried out until a convergence condition is reached.
5. The method according to claim 1, wherein in the step 3, the clustering process based on the SOM-FCM fusion algorithm is as follows:
step 31, initializing the SOM neural network: setting the competition layer structure of the SOM neural network, the initial domain radius delta (0) and the initial learning rateNumber of iterations k and total number of iterations ks(ii) a Carrying out normalization processing on the N-dimensional similarity index to obtain a training sample G; setting weight vector W corresponding to each neuron in competition layerj(k)=(Wj,1(k),Wj,2(k),...,Wj,N(k) (j ═ 1, 2.. times, M), initializing a weight vector, where M is the number of contention layer neurons, and k has an initial value of 0;
step 32, inputting a training sample: randomly selecting the ith sample Gi=(G1,G2,...,GN)TTo the input layer;
step 33, finding winning neurons: calculation of GiAnd Wj(k) The neuron with the smallest distance is selected as a winning neuron r;
step 34, adjusting the neuron connection weight: connected weight vector W for neurons in neighborhood of winning neuron rjAdjusting;
in the formula (I), the compound is shown in the specification,for learningRate;is the radius of the field of the winning neuron r;
step 36, repeating steps 32-35 until the training iteration number k reaches the preset total iteration number ksObtaining an SOM output layer neural network;
step 37, using SOM to output the weight W of each neuron of the neural network of the layerjSetting the clustering number c, membership factor m, limited error epsilon and maximum iteration number k of the FCM algorithm as the input vector of the FCM algorithmfInitializing a membership matrix U, and setting the iteration number k as 0;
step 38, calculating a clustering center according to the membership matrix U, wherein the ith clustering center CiComprises the following steps:
in the formula (I), the compound is shown in the specification,is an input vector WjFor CiDegree of membership of;
and 39, updating the membership matrix U according to the clustering center:
in the formula, dijIs an input vector WjFor CiThe Euclidean distance of; dqjIs an input vector WjFor CqEuclidean distance of, q ═ 1, 2.., c;
step 310, training an iteration counter k ═ k + 1;
step 311, repeat steps 39-310 until | U(k)-U(k+1)| < epsilon or the iteration number k reaches the maximum iteration number kf。
6. The method according to claim 5, wherein in step 3, the SOM network structure is selected according to the principle of minimum error, and the quantitative error QE and the topological error TE are selected for evaluation, and two indexes can express the quality of SOM neural network clustering:
TE=∑v(Gi)
where QE is the average relative distance between the input sample and the corresponding winning neuron, Wr(Gi) Is GiA weight vector of the corresponding winning neuron; TE is the degree of proximity of samples adjacent to the input space in the competition layer network, if the samples are GiThe neighboring samples remain adjacent in the output space, v (G)i) Is 1.
7. The method according to claim 6, wherein in the step 3, the clustering number c of the SOM-FCM algorithm is determined according to Davies-Bouldin index, and c is optimized by evaluating clustering effect under different clustering numbers:
wherein c is the number of clusters; m is a group ofiThe number of samples belonging to the category i; ciAs the cluster center of class i, CqIs the cluster center of class q, whichWherein, i is 1,2,., c, q is 1,2,. and c;class i and class q sample points to C, respectivelyiAnd CqThe average distance of (c).
8. The method according to claim 1, wherein in the step 5, according to the weather partition where the drainage basin is located and the included sub-drainage basin category, the fuzzy membership of the sub-drainage basins is used for constructing the drainage basin similarity comprehensive measurement; the degree of similarity between the domains is evaluated by using the maximum and minimum closeness method, and the domain B1And basin B2The similarity of (a) is expressed as:
in the formula (I), the compound is shown in the specification,is a drainage basin B1And basin B2The similarity of (2);represents basin B1And B2Weighting membership degrees for the areas of the basin class i;andrespectively is a drainage basin B1And B2The ratio of the area of the jth sub-basin to the total area of the basin; n is a radical ofB1And NB2Is a drainage basin B1And B2The number of sub-watersheds in (a);andrepresents basin B1And B2The fuzzy membership degree of the jth sub-basin to the basin class i; the A and V respectively represent the calculation of the minimum value and the maximum value.
9. The device for classifying the basin similarity is characterized by comprising a processor and a memory; stored in the memory are programs or instructions which are loaded and executed by the processor to implement the steps of the classification method according to any one of claims 1 to 8.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111170738.9A CN113887635B (en) | 2021-10-08 | 2021-10-08 | Basin similarity classification method and classification device |
GB2213018.1A GB2612682A (en) | 2021-10-08 | 2022-09-06 | Basin similarity classification method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111170738.9A CN113887635B (en) | 2021-10-08 | 2021-10-08 | Basin similarity classification method and classification device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113887635A CN113887635A (en) | 2022-01-04 |
CN113887635B true CN113887635B (en) | 2022-07-01 |
Family
ID=79005316
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111170738.9A Active CN113887635B (en) | 2021-10-08 | 2021-10-08 | Basin similarity classification method and classification device |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN113887635B (en) |
GB (1) | GB2612682A (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115293577B (en) * | 2022-08-05 | 2023-07-21 | 水利部珠江水利委员会技术咨询(广州)有限公司 | Machine learning-based high-cold-flow-domain groundwater chemical control factor analysis method |
CN117291349B (en) * | 2023-11-24 | 2024-03-19 | 济宁御龙源水务有限公司 | Groundwater level restoration prediction method and system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106815467A (en) * | 2016-12-12 | 2017-06-09 | 河海大学 | A kind of River Basin Hydrology zoning methods that similarity analysis are counted based on big data |
CN107092653A (en) * | 2017-03-15 | 2017-08-25 | 西安工程大学 | A kind of landslide Critical Rainfall Threshold based on method of fuzzy cluster analysis |
CN111985389A (en) * | 2020-08-18 | 2020-11-24 | 中国电建集团成都勘测设计研究院有限公司 | Basin similarity discrimination method based on basin attribute distance |
CN112347155A (en) * | 2020-10-29 | 2021-02-09 | 南京大学 | Site pollution characteristic factor identification and monitoring index optimization method based on data mining |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
SG11201706602RA (en) * | 2015-02-13 | 2017-09-28 | Beijing Didi Infinity Tech And Dev Co Ltd | Methods and systems for transport capacity scheduling |
CN109388664A (en) * | 2018-09-29 | 2019-02-26 | 河海大学 | A kind of middle and small river basin similitude method of discrimination |
CN109840873A (en) * | 2019-02-02 | 2019-06-04 | 中国水利水电科学研究院 | A kind of Cross Some Region Without Data Hydro-Model Parameter Calibration Technology fields method based on machine learning |
CN111242472B (en) * | 2020-01-09 | 2022-08-19 | 河海大学 | Basin similarity degree identification method based on Copula empirical density function |
CN113377750B (en) * | 2021-05-19 | 2022-03-29 | 山东大学 | Hydrological data cleaning method and system |
-
2021
- 2021-10-08 CN CN202111170738.9A patent/CN113887635B/en active Active
-
2022
- 2022-09-06 GB GB2213018.1A patent/GB2612682A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106815467A (en) * | 2016-12-12 | 2017-06-09 | 河海大学 | A kind of River Basin Hydrology zoning methods that similarity analysis are counted based on big data |
CN107092653A (en) * | 2017-03-15 | 2017-08-25 | 西安工程大学 | A kind of landslide Critical Rainfall Threshold based on method of fuzzy cluster analysis |
CN111985389A (en) * | 2020-08-18 | 2020-11-24 | 中国电建集团成都勘测设计研究院有限公司 | Basin similarity discrimination method based on basin attribute distance |
CN112347155A (en) * | 2020-10-29 | 2021-02-09 | 南京大学 | Site pollution characteristic factor identification and monitoring index optimization method based on data mining |
Non-Patent Citations (1)
Title |
---|
"Regional flood frequency analysis by combining;V.V. Srinivas etal.;《ELSEVIER》;20081231;第148-166页 * |
Also Published As
Publication number | Publication date |
---|---|
CN113887635A (en) | 2022-01-04 |
GB2612682A (en) | 2023-05-10 |
GB202213018D0 (en) | 2022-10-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108304668B (en) | Flood prediction method combining hydrologic process data and historical prior data | |
Shahabi et al. | Flash flood susceptibility mapping using a novel deep learning model based on deep belief network, back propagation and genetic algorithm | |
Termeh et al. | Flood susceptibility mapping using novel ensembles of adaptive neuro fuzzy inference system and metaheuristic algorithms | |
Pham et al. | Can deep learning algorithms outperform benchmark machine learning algorithms in flood susceptibility modeling? | |
Huang et al. | Uncertainty study of landslide susceptibility prediction considering the different attribute interval numbers of environmental factors and different data-based models | |
CN110471950B (en) | Intelligent model forecasting method for real-time flood forecasting of medium and small rivers | |
Soma et al. | Optimization of causative factors using logistic regression and artificial neural network models for landslide susceptibility assessment in Ujung Loe Watershed, South Sulawesi Indonesia | |
CN113887635B (en) | Basin similarity classification method and classification device | |
Rodrigues et al. | Estimation of renewable energy and built environment-related variables using neural networks–A review | |
CN113705877B (en) | Real-time moon runoff forecasting method based on deep learning model | |
CN115271255B (en) | Rainfall flood similarity analysis method and system based on knowledge graph and machine learning | |
CN114254561A (en) | Waterlogging prediction method, waterlogging prediction system and storage medium | |
CN113139329B (en) | Xinanjiang model parameter calibration method based on hydrological similarity and artificial neural network | |
CN111027764A (en) | Flood forecasting method suitable for runoff data lack basin based on machine learning | |
CN116090625A (en) | Coastal city flood rapid prediction method based on LightGBM and hydrological hydrodynamic model | |
Dehghanian et al. | Evaluation of coupled ANN-GA model to prioritize flood source areas in ungauged watersheds | |
Saab et al. | Review on generating optimal operation for dam and reservoir water system: simulation models and optimization algorithms | |
Liu et al. | Gully erosion susceptibility assessment based on machine learning-A case study of watersheds in Tuquan County in the black soil region of Northeast China | |
CN117808214A (en) | Hydraulic engineering data analysis system | |
Huang et al. | Modelling landslide susceptibility prediction: a review and construction of semi-supervised imbalanced theory | |
CN114372631A (en) | Data-lacking area runoff prediction method based on small sample learning and LSTM | |
CN117494419A (en) | Multi-model coupling drainage basin soil erosion remote sensing monitoring method | |
Chen et al. | Development of design hyetographs for ungauged sites using an approach combining PCA, SOM and kriging methods | |
Refonaa et al. | Cognitive computing techniques based rainfall prediction—a study | |
Feyzolahpour et al. | Estimating suspended sediment concentration using neural differential evolution (NDE), multi layer perceptron (MLP) and radial basis function (RBF) models |
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 |