KR102307898B1 - Groundwater potential mapping using integrated ensemble of three bivariate statistical models with random froest and logistic tree models - Google Patents

Abstract

A groundwater potential mapping technique using an integrated ensemble of three bivariate statistical models with random forest and logistic tree models is disclosed. The groundwater possibility mapping method includes the steps of selecting effective factors including topographic parameters and hydrological parameters of a target area; and generating a groundwater potential map through a bivariate statistical model using the effective factor.

Description

GROUNDWATER POTENTIAL MAPPING USING INTEGRATED ENSEMBLE OF THREE BIVARIATE STATISTICAL MODELS WITH RANDOM FROEST AND LOGISTIC TREE MODELS

The description below relates to groundwater potential mapping (GPM) technology.

Groundwater has been identified as one of the largest natural resources in many countries in recent decades due to population growth and industrialization. Groundwater provides about 50% of the water needed for drinking water, 40% of the water needed for industry and 20% of the water needed for agriculture. One of the advantages of groundwater over surface water is that it is stored naturally, does not occupy a large area, is safe from evaporation, is less susceptible to pollution and sudden drought, and can be used in all seasons. Given the growing use of groundwater for a variety of purposes and the over-exploitation of most aquifers and groundwater reservoirs, it is important to identify areas of varying potential for groundwater. Groundwater storage potential is related to the maximum permanent storage of aquifers. Groundwater potential information can play an important role in local decision-making. Given the fact that developing countries such as Iran have limited access to hydrological information, it is essential to understand the current state of the groundwater system.

Conventional groundwater exploration methods such as excavation, geophysical and geological methods require high cost, time, and manpower. For example, Korean Patent Registration No. 10-1718522 discloses a system and method for selecting a suitable site for a clean underground reservoir.

GIS and RS (remote sensing) methods are very effective in preparing groundwater potential maps (GPMs) and can improve the accuracy and speed of groundwater studies. Various methods such as frequency ratio (FR), certainty factor (CF), evidence confidence function (EBF), logistic regression (LR), weight of evidence (WOE), and entropy are used in GIS-based GPM. Random forest (RF), logistic model tree (LMT), decision tree (DT), classification and regression tree (CART), artificial neural network (ANN), support vector machine (SVM) and ensemble of metaheuristic algorithms. With the advancement of IT and big data, data mining algorithms are being widely used in GPM together with ANFIS (Adaptive Neuro-fuzzy Reasoning System).

Groundwater data analysis requires powerful and flexible analytical methods that can control non-linear relationships, interactions, and loss information. Also, understanding and presenting the results in this way should be simple and easily interpretable. Decision tree (DT) is easy to classify and interpret, but its biggest weakness is that the variance of the fitted model is large, making it difficult to interpret classification.

The present invention includes FR (Frequency Ratio), CF (Certainty Factor), EBF (Evidence Confidence Function), RF (Random Forest) and LMT (Logistic Model Tree) to create a Groundwater Probability Map (GPM) for a plain. We apply a novel ensemble data mining approach using a bivariate statistical model.

Random Forest (RF) and Logistic Model Tree (LMT) are among the most powerful methods in the field of GPM. One of the important issues in data mining algorithms is to preprocess and prepare the data to increase accuracy and select the best input data for the algorithm. There are no studies that have used a combination of bivariate statistical models and data mining algorithms to provide a GPM. The present invention aims to prepare GPM using three bivariate statistical models (FR, EBF, and CF) including random forest (RF) and logistic model tree (LMT) and select the most suitable hybrid model for plains. do.

A method of groundwater potential mapping executed on a computer system, the computer system comprising at least one processor configured to execute computer readable instructions contained in a memory, the method comprising: , selecting effective factors including topographic parameters and hydrological parameters of the target area; and generating, by the at least one processor, a groundwater potential map through a bivariate statistical model using the effective factor.

According to one aspect, the selecting step includes at least one of an altitude, a slope angle, a slope aspect, a slope length, a plan curvature, and a profile curvature. selecting a terrain parameter comprising one; and selecting a hydrological parameter including at least one of a topographic wetness index (TWI), a distance from a river, and a drainage density.

According to another aspect, the selecting comprises: Geological parameters including at least one of lithology, distance from fault, and fault density, rainfall. The method may further include selecting at least one of climate parameters, and ecological parameters including at least one of land use and soil.

According to another aspect, in the generating, the groundwater likelihood map may be generated through a bivariate statistical model including a random forest (RF) model and a logistic model tree (LMT) model.

According to another aspect, the generating includes two of a frequency ratio (FR) model, a certainty factor (CF) model, an evident belief function (EBF) model, a random forest (RF) model, and a logistic model tree (LMT). The groundwater potential map may be generated through a bivariate statistical model including a dog model.

According to another aspect, the groundwater likelihood mapping method may further comprise, by the at least one processor, verifying, by the at least one processor, the groundwater likelihood map using a receiver operating characteristic (ROC) curve and an area under curve (AUC). can

A computer system comprising: at least one processor configured to execute computer readable instructions contained in a memory, wherein the at least one processor is configured to select an effective factor comprising a topographical parameter and a hydrological parameter of a target area ; and using the effective factor as input values of a bivariate statistical model including a random forest (RF) model and a logistic model tree (LMT) model to provide a computer system that processes a process of generating a groundwater potential map.

1 is a flowchart illustrating an example of a groundwater potential mapping (GPM) method according to an embodiment of the present invention.
2 shows an example of a target area according to an embodiment of the present invention.
3 shows the characteristics of rock quality of a target area according to an embodiment of the present invention.
4 to 18 are exemplary views for explaining the groundwater effective factors in an embodiment of the present invention.
19 is a block diagram for explaining an example of an internal configuration of a computer system according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Embodiments of the present invention relate to groundwater potential mapping (GPM) technology.

Embodiments, including those specifically disclosed herein, provide FR (Frequency Ratio), CF (Correctiveness Factor), EBF (Evidence Confidence Function), RF (Random Forest) to create a groundwater likelihood map (GPM) for a plain. and LMT (Logistic Model Tree), a novel ensemble data mining approach using a bivariate statistical model.

Random Forest (RF) and Logistic Model Tree (LMT) are among the most powerful methods in the field of GPM. One of the important issues in data mining algorithms is to preprocess and prepare the data to increase accuracy and select the best input data for the algorithm. There are no studies that have used a combination of bivariate statistical models and data mining algorithms to provide a GPM. The present invention aims to prepare GPM using three bivariate statistical models (FR, EBF, and CF) including random forest (RF) and logistic model tree (LMT) and select the most suitable hybrid model for plains. do.

The present invention may include five steps to provide a groundwater potential map (GPM). The first step is to characterize the area being studied and identify existing wells. In the second stage, the necessary data is collected and a spatial database of effective criteria is established. At this time, well-distributed elevation, slope angle, slope, plane curvature, longitudinal curvature, slope length, terrain humidity index (TWI), rainfall, distance from river, fault distance, drainage density, fault density, rock quality, land use and soil 15 Prepare for modeling by selecting individual factors. In the third step, spatial relationships are calculated using the FR, CF, and EBF models between the existing well and the effective reference. In the fourth step, the weights of the three statistical models are considered as input values of the RF and LMT models to generate the GPM. In the fifth step, the GPM is verified using ROC and AUC and finally the best model is selected. 1 shows an example of a groundwater potential mapping (GPM) method according to the present invention.

target area

For example, the Booshehr Plain lies on an area of 2696 km2 in southeastern Iran between 51°20' and 52°10' east longitude and 27°50' and 28°30' north latitude. Much of the Bushehr Plain is at low elevations. The elevation of the Bushehr Plain is between 3 and 1490 m above sea level. The average annual temperature in this area is 24°C and the highest temperature is 50°C in summer. Rainfall is low, with an average of less than 255 mm. Most of the rain falls between November and May. Rainfall in spring and autumn is strong and short, and sparse and irregular in winter. The average annual humidity in this area is over 71%. FIG. 2 shows an example of a target area for groundwater potential mapping (GPM), and FIG. 3 shows parameters for a local rock quality unit.

well list

11m³/h)과 평균 pH, 6.9와 495 μmhos/cm의 전기전도도(EC) 데이터가 사용된다. 우물들을 랜덤으로 두 개의 훈련과 시험 데이터 세트로 나뉜다. 총 장소(238개 우물)의 70%를 훈련으로, 나머지 우물(30%, 101개 우물)을 검증으로 간주한다. 우물 위치는 도 2와 같다.Information on groundwater wells is provided by the Department of Water Resources Management, Bushehr Province. According to previous studies and water quality management reports, each of the high-probability groundwater wells (

11 m ³ /h) and average pH, electrical conductivity (EC) data of 6.9 and 495 μmhos/cm are used. The wells are randomly divided into two training and test data sets. Consider 70% of the total sites (238 wells) as training and the remaining wells (30%, 101 wells) as validation. The well location is shown in FIG. 2 .

Conditioning Factors

(1) Topographic Parameters

The digital elevation model (DEM) was downloaded with a spatial resolution of 30 m × 30 m. The topographical parameters of the target area consisting of elevation, longitudinal curvature, slope length, slope angle, planar curvature and slope can be obtained from DEM. The classification of all maps is based on the characteristics of the area and on previous studies as well as on natural cleavage techniques. Different elevations create different climatic conditions, resulting in different types of plants and soils.

The elevation map is divided into five grades (less than 108 m, 108-287 m, 287-535 m, 553-851 m, greater than 851 m) (FIG. 4).

The angle of inclination mainly controls the groundwater supply, infiltration and runoff processes and the rate of groundwater movement. The angle of inclination is classified into five classes: 0°-6°, 6°-14°, 14°-24°, 24°-39°, and greater than 39° (Fig. 5).

Slopes are affected by rainfall and topographical processes and affect precipitation and vegetation types. The slopes are classified into nine categories (FIG. 6).

The slope length is measured by Equation 1 as a function of the reservoir area and the slope slope.

[Equation 1]

는 경사각이다. 경사 길이는 5개 등급(0-10m, 10-20m, 20-30m, 30-40m, 40m 초과)으로 분류된다(도 7).where LS is the slope length, B _S is the reservoir area (m ² ),

is the angle of inclination. The slope length is classified into 5 grades (0-10 m, 10-20 m, 20-30 m, 30-40 m, greater than 40 m) (FIG. 7).

Planar curvature affects flow convergence and dispersion, and longitudinal curvature coincides with the maximum slope direction and mainly affects surface flow velocity. Planar curvature was classified into 5 classes (less than -2.22 100/m, -2.22-0.8 100/m, 0.8-0.4 100/m, 0.4-2.2 100/m, greater than 2.2 100/m) (Fig. 8), The longitudinal curvature layer is classified into five grades (less than 3.4 100/m, -3.4-1 100/m, -1-0.4 100/m, -0.4-2.8 100/m, greater than 2.8 100/m) (Fig. 9) ).

(2) Hydrological Parameters

In hydrological systems, hydrological parameters such as distance to river, TWI and drainage density play an important role. In soil moisture, slope stability, groundwater flow and TWI play important roles. TWI represents the topographical control over the hydrological process, and can be defined as Equation (2).

[Equation 2]

는 라디언 경사각이다. TWI는 2.92 미만, 2.92-3.84, 3.84-4.69, 4.69-6.57, 6.57 초과의 5가지 등급으로 분류된다(도 10).where A _S is the area of the cumulative ascending slope,

is the angle of inclination in radians. TWI is classified into five grades: less than 2.92, 2.92-3.84, 3.84-4.69, 4.69-6.57, and greater than 6.57 (FIG. 10).

Local drainage systems depend on geological formation, soil absorption capacity, permeability and slope. High drainage density increases surface runoff and reduces penetration. Areas of high drainage density are not suitable for the production of groundwater resources.

The distance to the river is divided into seven grades of less than 100 m, 100-200 m, 200-500 m, 500-1000 m, 1000-1500 m, 1500-2000 m, and more than 2000 m (FIG. 11).

Drainage density is 0.13 km/km ² Less than 0.13-0.27 km/km ² , 0.27-0.4 km/km ² , 0.4-0.58 km/km ² , and more than 0.58 km/km ² are classified into five grades ( FIG. 12 ).

(3) Geological Parameters

Rock quality affects groundwater potential through hydraulic conductivity.

For example, in the Booshehr Plain, there are 11 types of rock quality maps including Qft2, MuPlaj, Plbk, Mmn, Mgs, Eoas-ja, KEpd-gu, Kbgp, JKkgp, OMr, and Pc-ch (Fig. 13).

Faults primarily control the distribution of spatial networks and the accumulation of groundwater. The faults of the Booshehr Plain are determined on a scale of 1:100,000 on the geological maps of the Booshehr province, and the distances from the faults are created to be less than 100 m, 100-200 m, 200-500 m, 500-1000 m, 1000-2000 m, 2000-5000 m. , are classified into 7 grades, such as over 5000 m (FIG. 14).

Fault densities are classified into five classes ( ^{less than 0.03 km/km 2} , 0.03-0.09 km/km ² , 0.09-0.13 km/km ² , 0.13-0.19 km/km ² , 0.19 km/km ^{2 ).} excess) (Fig. 15).

(4) Climate Parameters

Rain is the climatic variable that has the greatest impact on groundwater recharge. Rainfall is very important for evaluating the flow of water into the basin and understanding the trophic state of the basin.

The rainfall is divided into five grades of less than 247 mm, 247-264 mm, 264-281 mm, 281-297 mm, and more than 297 mm (FIG. 16).

(5) Ecological Parameters

The most important ecological parameters are land use and soil. Land use directly or indirectly affects permeability, runoff and evaporation.

The land use map is divided into 11 classes: mangrove forest, forest, urban, saltland, very low forest, low-range land, sand dune, mid-range land, reforestation, rock and water itself (FIG. 17).

One of the key parameters for the generation and accumulation of surface runoff and subsurface runoff is soil type. The soil map is classified into three groups (Entisols/Aridisoils, rock outcrops/Entisols, badlands) (FIG. 18).

Model

(1) FR (frequency ratio) model

FR shows the relationship between the effective factors of wells and groundwater. FR is the ratio of the area where the well is located. Calculate the ratio of well occurrences to nonoccurrences to calculate FR values for each class or factor. The FR model is calculated through Equation 3.

[Equation 3]

는 우물 위치와 함께 각 기준의 등급 별로 있는 픽셀의 총 값이고,

는 각 기준 j의 등급 별로 있는 픽셀 수이며, m과 n은 각각 기준당 등급 수와 기준의 총 수이다.here,

is the total number of pixels per class of each criterion, along with the well location,

is the number of pixels per grade of each criterion j, and m and n are the number of grades per criterion and the total number of criteria, respectively.

In the present invention, 15 parameters affecting groundwater are used, each of these parameters is referred to as a standard, and each standard is classified into different categories and referred to as a grade. Although the FR model uses simple and understandable concepts, it can also analyze bivariate statistical analysis and classification of each factor.

(2) CF (certainty factor) model

The CF model integrates the results into a spatial database using GIS. The CF model is calculated using Equation 4 as the well event frequency of each layer.

[Equation 4]

where pp _a is the conditional probability of a well in a class and pp _s is the likelihood of a well transfer in the area. pp _a is the ratio of the number of well pixels in a class to the total pixels in that class, and pp _s is the ratio of pixels with wells in the target area to the total pixels in the map.

The CF model ranges from -1 to +1, with positive values indicating an increase in certainty and negative values indicating a decrease in certainty.

(3) EBF (evidential belief function) model

The EBF model is a statistical bivariate technique based on the Dempster Shafer theory. In the EBF model, the Bel, Unc, Dis, and Pls parameters are confidence rank, uncertainty rank, distrust, and reliability rank, respectively. In the EBF model, the Bel parameter considers the pessimistic mode and low probability, and the Pls parameter considers the optimistic mode and the high probability state, so the value of the Bel parameter is less than or equal to the Pls parameter, and the difference between these two parameters is called Unc. do. The data extracted from the EBF model not only estimate the spatial correlation between effective factors and well occurrence, but also estimate the spatial correlation of each class factor.

The EBF model is calculated through Equations 5 and 6.

[Equation 5]

[Equation 6]

A conditional probability representing the probability of an existing well (ie, groundwater generation) in the absence of C _{ij (each grade for each factor) is shown in Equation 6.} WC _ij D is the weight of _{C ij} representing the confidence that wells exist instead of scarce. In this relationship, m denotes the number of criteria considered for modeling, i denotes each grade for each criterion, and j denotes each criterion. In this relationship, N(T) and N(D) represent the total number of pixels in the target region and the total number of well pixels in the target region, respectively.

(4) RF (random forest) model

The RF model is one of the data mining algorithms used by various trees in classification. The RF model creates a massive decision tree by replacing and changing the variables that affect the target. Then, in the prediction, the algorithm merges all the trees. In the training process, the original data of each tree is randomly selected. RF contains three user-defined parameters, which include the number of arguments used to construct each tree, the number of trees, and the minimum number of tree nodes. By increasing the strength of the autonomous tree and reducing the correlation between them, the predictive power of the RF model is improved. The RF system does not use all the accessible information to grow the tree, but utilizes 66% of the bootstrap information. Then, in the growth phase, predictor variables are implemented randomly and these variables are used to generate nodes in the tree. Thus, the decision tree is produced at its maximum size. 33% of the remaining information is also used to evaluate the tree fit. This process is repeated several times, and the final prediction of the algorithm is used as the average of all expected values.

(5) LMT (logistic model tree) model

The LMT model is a nonparametric method for predicting a quantitative variable or a classification variable based on a set of quantitative and qualitative predictor variables. In fact, the decision tree of the hierarchical model is composed of decision tools that are decomposed into homogeneous regions of independent variables. A decision tree is a way of defining a set of laws leading to a category or value. One of the differences between the construction techniques of decision trees is the method of measuring this distance. Decision trees used to predict discrete variables are called classification trees because they classify samples. Decision trees are called regression trees and are used to predict continuous variables. The purpose of a decision tree is to discover an approach in the form of a set of rules and present the results of predictions extracted from a set of input factors. A classification tree model is a combination of logistic regression and decision tree learning. To isolate the logistic-type data increments, we use the LogitBoost algorithm to create an LR model from the leaves of each tree and use the CART algorithm to truncate the tree. For each class C (well or no wells), the LogitBoost algorithm uses a logistic regression analysis of the addition factor with least squares as Equation (7).

[Equation 7]

[Equation 8]

where c is the class number

Verification

The accuracy of the groundwater potential map can be determined using a receiver operating characteristic (ROC) curve and an area under the curve (AUC). The accuracy of the groundwater potential map can be verified by determining the AUC using the CF-RF, EBF-RF, FR-RF, CF-LMT, EBF-LMT, and FR-LMT methods.

One of the appropriate methods for assessing classification results and for assessing the ability to identify assigned classes is to use ROC curves and AUC to validate the sensitivity of the method. Sensitivity refers to the relationship between classified and unclassified values. The greater the deviation from the baseline for a particular class in the ROC curve, the more efficient the classifier is at identifying the class. In addition to taking into account the trend chart of a particular class, the AUC is also calculated. The target area represents the probability that a randomly selected value will be correctly classified. A higher value indicates the reliability of the method. The rating indices are: values properly assigned to target ratings (True Positive), values assigned to incorrect ratings (False Positive), values not assigned to defined ratings (True Negative), and values not assigned to incorrect ratings (False Negative). evaluate The evaluation curve is composed of a horizontal axis (X axis) and a vertical axis (Y axis) calculated by Equations (9) and (10).

[Equation 9]

[Equation 10]

When the actual output is positive and the predicted value is positive, this state is called TP (True Positive), and FN (False negative) indicates the state in which the actual output is negative and the predicted value is also negative. TN (True Negative) indicates a state in which the actual output is positive and the predicted value is negative, and FP (False Positive) indicates a state in which the actual output is negative and the predicted value is positive. These evaluation indices are derived from the confusion matrix and the ROC curve is calculated according to this criterion.

19 is a block diagram illustrating an example of a computer system according to an embodiment of the present invention. For example, the groundwater potential mapping system according to embodiments of the present invention may be implemented by the computer system 1900 illustrated in FIG. 19 .

As shown in FIG. 19 , the computer system 1900 is a component for executing the groundwater potential mapping method according to embodiments of the present invention, and includes a memory 1910 , a processor 1920 , a communication interface 1930 , and It may include an input/output interface 1940 .

The memory 1910 is a computer-readable recording medium and may include a random access memory (RAM), a read only memory (ROM), and a permanent mass storage device such as a disk drive. Here, a non-volatile mass storage device such as a ROM and a disk drive may be included in the computer system 1900 as a separate permanent storage device distinct from the memory 1910 . Also, an operating system and at least one program code may be stored in the memory 1910 . These software components may be loaded into the memory 1910 from a computer-readable recording medium separate from the memory 1910 . The separate computer-readable recording medium may include a computer-readable recording medium such as a floppy drive, a disk, a tape, a DVD/CD-ROM drive, and a memory card. In another embodiment, the software components may be loaded into the memory 1910 through the communication interface 1930 instead of a computer-readable recording medium. For example, software components may be loaded into memory 1910 of computer system 1900 based on computer programs installed by files received over network 1960 .

The processor 1920 may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input/output operations. Instructions may be provided to processor 1920 by memory 1910 or communication interface 1930 . For example, the processor 1920 may be configured to execute a received instruction according to program code stored in a recording device such as the memory 1910 .

The communication interface 1930 may provide functions for the computer system 1900 to communicate with each other with other devices over the network 1960 . For example, a request, command, data, file, etc. generated by the processor 1920 of the computer system 1900 according to the program code stored in a recording device such as the memory 1910 is transmitted to the network ( 1960) to other devices. Conversely, signals, commands, data, files, etc. from other devices may be received by the computer system 1900 via the communication interface 1930 of the computer system 1900 via the network 1960 . A signal, command, or data received through the communication interface 1930 may be transferred to the processor 1920 or the memory 1910 , and the file, etc. may be a storage medium (described above) that the computer system 1900 may further include. persistent storage).

The communication method is not limited, and short-distance wired/wireless communication between devices as well as a communication method using a communication network (eg, mobile communication network, wired Internet, wireless Internet, broadcasting network) that the network 1960 may include may also be included. have. For example, the network 1960 may include a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), and a broadband network (BBN). , the Internet, and the like. In addition, the network 1960 may include any one or more of a network topology including a bus network, a star network, a ring network, a mesh network, a star-bus network, a tree or a hierarchical network, etc. not limited

The input/output interface 1940 may be a means for an interface with the input/output device 1950 . For example, the input device may include a device such as a microphone, keyboard, camera, or mouse, and the output device may include a device such as a display or speaker. As another example, the input/output interface 1940 may be a means for an interface with a device in which functions for input and output are integrated into one, such as a touch screen. The input/output device 1950 may be configured as one device with the computer system 1900 .

This embodiment of FIG. 19 is only an example of the computer system 1900, and the computer system 1900 further includes additional components not shown in FIG. 19, or may have a configuration or arrangement combining two or more components. can The components that may be included in the computer system 1900 may be implemented in hardware, software, or a combination of both hardware and software including one or more signal processing or application-specific integrated circuits.

The device described above may be implemented as a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, the devices and components described in the embodiments may include a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), and a programmable logic unit (PLU). It may be implemented using one or more general purpose or special purpose computers, such as a logic unit, microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that can include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

The software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be embodied in any type of machine, component, physical device, computer storage medium or device for interpretation by or providing instructions or data to the processing device. have. The software may be distributed over networked computer systems, and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. In this case, the medium may be to continuously store the program executable by the computer, or to temporarily store the program for execution or download. In addition, the medium may be a variety of recording means or storage means in the form of a single or several hardware combined, it is not limited to a medium directly connected to any computer system, and may exist distributed on a network. Examples of the medium include a hard disk, a magnetic medium such as a floppy disk and a magnetic tape, an optical recording medium such as CD-ROM and DVD, a magneto-optical medium such as a floppy disk, and those configured to store program instructions, including ROM, RAM, flash memory, and the like. In addition, examples of other media include recording media or storage media managed by an app store that distributes applications, sites that supply or distribute various other software, and servers.

As described above, although the embodiments have been described with reference to the limited embodiments and drawings, various modifications and variations are possible from the above description by those skilled in the art. For example, the described techniques are performed in a different order than the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (7)

A method of groundwater potential mapping executed on a computer system, comprising:
the computer system comprising at least one processor configured to execute computer readable instructions contained in a memory;
The groundwater potential mapping method is,
selecting, by the at least one processor, effective factors including topographic parameters and hydrological parameters of a target area; and
generating, by the at least one processor, a groundwater potential map through a bivariate statistical model using the effective factor;
including,
The selecting step is
selecting a terrain parameter including an altitude, a slope angle, a slope aspect, a slope length, a plan curvature, and a profile curvature;
selecting hydrological parameters including topographic wetness index (TWI), distance from river, and drainage density; and
Geological parameters including lithology, distance from fault, fault density, climate parameters including rainfall, land use and Step of selecting ecological parameters including soil (soil)
including,
The generating step is
Three bivariate statistical models, including a random forest (RF) model and a logistic model tree (LMT) model, the FR (frequency ratio) model, the CF (certainty factor) model, and the EBF (evidential belief function) model By creating a map,
calculating a spatial relationship for the effective factor through the FR model, the CF model, and the EBF model; and
generating the groundwater possibility map by using the weights of the three bivariate statistical models as input values of the RF model and the LMT model
A groundwater potential mapping method comprising a. delete delete delete delete According to claim 1,
The groundwater potential mapping method is,
verifying, by the at least one processor, the groundwater potential map using a receiver operating characteristic (ROC) curve and an area under curve (AUC);
A groundwater potential mapping method further comprising a. In a computer system,
at least one processor configured to execute computer readable instructions contained in memory
including,
the at least one processor,
a process of selecting an effective factor including a topographical parameter and a hydrological parameter of the target area; and
The process of generating a groundwater potential map by using the effective factor as an input value of a bivariate statistical model including an RF (random forest) model and an LMT (logistic model tree) model
process the
The selection process is
a process of selecting a terrain parameter including an altitude, a slope angle, a slope aspect, a slope length, a plan curvature, and a profile curvature;
a process of selecting hydrological parameters including topographic wetness index (TWI), distance from river, and drainage density; and
Geological parameters including lithology, distance from fault, fault density, climate parameters including rainfall, land use and The process of selecting ecological parameters including soil
including,
The creation process is
Three bivariate statistical models, including a random forest (RF) model and a logistic model tree (LMT) model, the FR (frequency ratio) model, the CF (certainty factor) model, and the EBF (evidential belief function) model By creating a map,
calculating a spatial relationship for the effective factor through the FR model, the CF model, and the EBF model; and
The process of generating the groundwater possibility map by using the weights of the three bivariate statistical models as input values of the RF model and the LMT model
A computer system comprising a.

Images