Geological disaster risk prediction method and system

Abstract

The invention discloses a geological disaster risk prediction method. The method comprises the following steps: acquiring existing geological disaster data information; a plurality of geological factors are selected as geological disaster risk candidate indexes, correlation analysis is carried out, final geological disaster risk indexes are obtained, and a training data set is constructed; constructing a geological disaster risk prediction initial model based on SPM chaotic mapping, a sparrow algorithm and a limit gradient lifting algorithm, and training to obtain a geological disaster risk prediction model; and adopting the geological disaster risk prediction model to carry out actual geological disaster risk prediction of the target region. The invention also discloses a system for realizing the geological disaster risk prediction method. According to the method, the risk indexes are selected based on the correlation analysis scheme, and the geological disaster risk prediction model is constructed and trained based on the SPM chaotic mapping, the sparrow algorithm and the limit gradient lifting algorithm, so that the geological disaster risk prediction can be realized, the reliability is higher, and the accuracy is better.

Description

Technical Field

[0001] The present invention belongs to the field of safety engineering, and particularly relates to a geological disaster risk prediction method and system. Background Art

[0002] Geological disasters are disasters caused by geological processes that lead to the deterioration of the geological natural environment. At present, the losses caused by geological disasters show an upward trend; therefore, it is particularly important to predict geological disasters.

[0003] At present, traditional geological disaster prediction schemes include qualitative evaluation schemes and quantitative evaluation schemes. Qualitative evaluation schemes mainly include the analytic hierarchy process scheme, fuzzy comprehensive evaluation scheme, and determination coefficient scheme, etc. However, since the results of qualitative schemes are only qualitative results without corresponding data support, the practicality and accuracy of qualitative schemes are not high. Quantitative evaluation schemes include multivariate statistical analysis schemes and machine learning schemes. However, due to the complex mechanism of geological disasters, multivariate statistical analysis schemes have limitations. Existing machine learning schemes include support vector machines, random forests, logistic regression, artificial neural networks, and BP neural networks, etc. However, such schemes still have the defect of low model prediction accuracy in the process of geological disaster prediction. Summary of the Invention

[0004] One object of the present invention is to provide a geological disaster risk prediction method with high reliability and good accuracy.

[0005] Another object of the present invention is to provide a system for implementing the geological disaster risk prediction method.

[0006] The geological disaster risk prediction method provided by the present invention includes the following steps:

[0007] S1. Obtain existing geological disaster data information;

[0008] S2. According to the data information obtained in step S1, select several geological factors as candidate indicators for geological disaster risk;

[0009] S3. Conduct a correlation analysis on the candidate indicators for geological disaster risk selected in step S2 to obtain the final geological disaster risk indicators, and construct a training data set;

[0010] S4. Based on the SPM chaotic mapping, sparrow algorithm, and extreme gradient boosting algorithm, construct an initial geological disaster risk prediction model;

[0011] S5. Use the training data set constructed in step S3 to train the initial geological disaster risk prediction model constructed in step S4 to obtain a trained geological disaster risk prediction model;

[0012] S6. Use the geological disaster risk prediction model obtained in step S5 to predict the geological disaster risk of the actual target area.

[0013] According to the data information obtained in step S1, select several geological factors as candidate indicators for geological disaster risk, which specifically include the following steps:

[0014] According to the data information obtained in step S1, select several geological factors as candidate indicators for geological disaster risk;

[0015] The selected geological disaster levels include extremely low incidence, low incidence, medium incidence, and high incidence;

[0016] The selected candidate indicators for geological disaster risk include slope, aspect, elevation, curvature, topographic relief, distance from river, annual average precipitation, vegetation index, engineering rock group, distance from fault, land cover type, distance from road, soil permeability coefficient, surface humidity index, distance from mining area, and peak ground acceleration of earthquake.

[0017] Perform a correlation analysis on the candidate indicators for geological disaster risk selected in step S2 to obtain the final geological disaster risk indicators, which specifically include the following steps:

[0018] Perform a correlation analysis on the candidate indicators for geological disaster risk selected in step S2 using the Pearson correlation coefficient method;

[0019] Finally, according to the results of the correlation analysis, select the final geological disaster risk indicators including slope, elevation, curvature, topographic relief, distance from river, annual average precipitation, vegetation index, engineering rock group, distance from fault, land cover type, and distance from road.

[0020] Based on the SPM chaos mapping, sparrow algorithm, and extreme gradient boosting algorithm, construct an initial model for predicting geological disaster risk, which specifically includes the following steps:

[0021] A. Use the SPM chaos mapping method to optimize the initial population of the sparrow algorithm;

[0022] B. Calculate the fitness value of the current population;

[0023] C. Divide the sparrows into discoverers and followers according to the fitness value obtained in step B;

[0024] D. The discoverers update their own positions; the followers update their own positions according to the positions of the discoverers and their own states;

[0025] E. Determine the iteration termination: If the iteration termination condition is reached, the parameter values corresponding to the individual sparrow with the highest current fitness are used as the output parameter values; if the iteration termination condition is not reached, return to step B for the next round of iteration;

[0026] F. Use the output parameter values obtained in step E as the parameter values of the extreme gradient boosting model, thereby constructing an initial geological disaster risk prediction model.

[0027] The use of the SPM chaotic mapping scheme in step A to optimize the initial population of the sparrow algorithm specifically includes the following steps:

[0028] Use the following formula to perform SPM chaotic mapping to optimize the initial population of the sparrow algorithm:

[0029]

[0030] where x i is the position of the i-th sparrow; mod() is the modulo operator; η is the set first parameter, and η ∈ (0, 1); μ is the set second parameter, and μ ∈ (0, 1); tur is a random number, and tur ∈ (0, 1).

[0031] The calculation of the fitness value of the current population in step B specifically includes the following steps:

[0032] Use the accuracy rate of the extreme gradient boosting model as the fitness function and calculate the fitness value of the current population.

[0033] According to the fitness value obtained in step B in step C, divide the sparrows into discoverers and followers, specifically including the following steps:

[0034] Sort the fitness values of each individual in the current population obtained in step B in descending order, select the top n% of the individuals as discoverers, and use the remaining individuals as followers; n takes a positive value.

[0035] The update of the position of the discoverer in step D specifically includes the following steps:

[0036] The discoverer updates its position using the following formula:

[0037]

[0038] where: is the position of the i-th discoverer in the j-th dimension at the (t + 1)-th iteration; is the position of the $i$-th discoverer in the $j$-th dimension at the $t$-th iteration; rand is a random number uniformly distributed in the interval $[0, 1]$; $T$ is the maximum number of iterations; $ST$ is the set safety value; $R2$ is a random number with a value range of $[0, 1]$; $Q$ is a random number from the standard normal distribution.

[0039] The follower in step D updates its own position according to the position of the discoverer and its own state, which specifically includes the following steps:

[0040] The follower updates its own position using the following formula:

[0041]

[0042] where: is the position of the $i1$-th follower in the $j$-th dimension at the $(t + 1)$-th iteration; is the position of the $i1$-th follower in the $j$-th dimension at the $t$-th iteration; $Q$ is a random number from the standard normal distribution; is the position of the individual corresponding to the global worst fitness; $n$ is the number of joiners in the sparrow algorithm; is the position of the individual corresponding to the global best fitness; $d$ is the number of spatial dimensions; $A$ + is a random matrix of size $1\times d$, and each element in the matrix is randomly assigned a value of 1 or -1.

[0043] The construction of the initial geological disaster risk prediction model in step F specifically includes the following steps:

[0044] The constructed initial geological disaster risk prediction model is expressed as

[0045]

[0046] where is the log-odds of the initial disaster probability; $\gamma$ is the log-odds of the initial set proportion of disaster samples; $L(y$ i , $\gamma)$ is the set loss function; $N$ is the number of samples; $i$ represents the sample index.

[0047] The model update process is expressed as

[0048]

[0049] where is the predicted value after the $t$-th round of iteration; $\eta$ is the learning rate; $f$ t (x i ) is the predicted output value of the $t$-th tree;

[0050] The final prediction formula is expressed as

[0051]

[0052] where \(T\) is the total number of trees; \(\eta_f\) t (\(x\) i ) is the weighted prediction value of the \(t\)-th tree.

[0053] The present invention also provides a system for implementing the geological disaster risk prediction method, including a data acquisition module, an index selection module, a data determination module, a model construction module, a model training module, and a risk prediction module; the data acquisition module, the index selection module, the data determination module, the model construction module, the model training module, and the risk prediction module are connected in series in sequence; the data acquisition module is used to acquire existing geological disaster data information and upload the data information to the index selection module; the index selection module is used to select several geological factors as candidate indicators for geological disaster risk according to the received data information and upload the data information to the data determination module; the data determination module is used to perform a correlation analysis on the selected candidate indicators for geological disaster risk according to the received data information, obtain the final geological disaster risk indicators, construct a training data set, and upload the data information to the model construction module; the model construction module is used to construct an initial model for geological disaster risk prediction based on the SPM chaotic mapping, the sparrow algorithm, and the extreme gradient boosting algorithm according to the received data information and upload the data information to the model training module; the model training module is used to train the initial model for geological disaster risk prediction constructed in step S4 by using the constructed training data set according to the received data information, obtain a trained geological disaster risk prediction model, and upload the data information to the risk prediction module; the risk prediction module is used to perform the geological disaster risk prediction of the actual target area by using the obtained geological disaster risk prediction model according to the received data information.

[0054] This geological disaster risk prediction method and system provided by the present invention select risk indicators based on a correlation analysis scheme, and construct and train a geological disaster risk prediction model based on the SPM chaotic mapping, the sparrow algorithm, and the extreme gradient boosting algorithm. Therefore, the present invention can not only realize the risk prediction of geological disasters, but also has higher reliability and better accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a schematic flowchart of the method of the present invention.

[0056] Figure 2 It is a schematic diagram of the performance comparison between the method of the present invention and the existing scheme.

[0057] Figure 3 It is a schematic diagram of the performance comparison between the method of the present invention and the existing SSA-extreme gradient boosting algorithm on unimodal and multimodal test functions.

[0058] Figure 4Schematic diagram of the search space for the method of the present invention.

[0059] Figure 5 Schematic diagram of the functional modules of the system of the present invention. Detailed implementation manners

[0060] As Figure 1 Shown is the schematic diagram of the method flow of the method of the present invention: The geological disaster risk prediction method disclosed by the present invention includes the following steps:

[0061] S1. Obtain existing geological disaster data information;

[0062] S2. According to the data information obtained in step S1, select several geological factors as candidate indicators for geological disaster risks; specifically including the following steps:

[0063] According to the data information obtained in step S1, select several geological factors as candidate indicators for geological disaster risks;

[0064] The selected geological disaster levels include extremely low incidence rate, low incidence rate, medium incidence rate and high incidence rate;

[0065] The selected candidate indicators for geological disaster risks include slope, aspect, elevation, curvature, topographic relief, distance from river, annual average precipitation, vegetation index, engineering rock group, distance from fault, land cover type, distance from road, soil permeability coefficient, surface humidity index, distance from mining area and peak ground acceleration of earthquake, etc.;

[0066] S3. Conduct a correlation analysis on the candidate indicators for geological disaster risks selected in step S2 to obtain the final geological disaster risk indicators and construct a training data set; specifically including the following steps:

[0067] For the candidate indicators for geological disaster risks selected in step S2, adopt the Pearson correlation coefficient scheme to conduct a correlation analysis;

[0068] Finally, according to the correlation analysis results, the selected final geological disaster risk indicators include slope, elevation, curvature, topographic relief, distance from river, annual average precipitation, vegetation index, engineering rock group, distance from fault, land cover type and distance from road;

[0069] S4. Based on the SPM chaotic mapping, sparrow algorithm and extreme gradient boosting algorithm, construct an initial model for geological disaster risk prediction; specifically including the following steps:

[0070] A. Adopt the SPM chaotic mapping scheme to optimize the initial population of the sparrow algorithm; specifically including the following steps:

[0071] The following formula is used to perform the SPM chaotic mapping to optimize the initial population of the sparrow algorithm:

[0072]

[0073] where x i is the position of the i-th sparrow; mod() is the modulo operator; η is the set first parameter and η ∈ (0, 1); μ is the set second parameter and μ ∈ (0, 1); tur is a random number and tur ∈ (0, 1);

[0074] B. Calculate the fitness values of the current population; specifically, it includes the following steps:

[0075] Use the accuracy rate of the extreme gradient boosting model as the fitness function and calculate the fitness values of the current population;

[0076] C. According to the fitness values obtained in step B, divide the sparrows into discoverers and followers; specifically, it includes the following steps:

[0077] Sort the fitness values of each individual in the current population obtained in step B in descending order, select the top n% of the individuals as discoverers, and regard the remaining individuals as followers; the value of n is a positive number, preferably 70;

[0078] D. The discoverers update their own positions; the followers update their own positions according to the positions of the discoverers and their own states;

[0079] Among them, the discoverers update their own positions, specifically including the following steps:

[0080] The discoverers update their own positions using the following formula:

[0081]

[0082] where: is the position of the i-th discoverer in the j-th dimension at the (t + 1)-th iteration; is the position of the i-th discoverer in the j-th dimension at the t-th iteration; rand is a random number uniformly distributed in the interval [0, 1]; T is the maximum number of iterations; ST is the set safety value; R2 is a random number with a value range of [0, 1]; Q is a random number from the standard normal distribution;

[0083] The followers update their own positions according to the positions of the discoverers and their own states, specifically including the following steps:

[0084] The followers update their own positions using the following formula:

[0085]

[0086] In the formula: is the position of the i1-th follower at the j-th dimension in the (t + 1)-th iteration; is the position of the i1-th follower at the j-th dimension in the t-th iteration; Q is a random number from the standard normal distribution; is the position of the individual corresponding to the global worst fitness; n is the number of joiners in the sparrow algorithm; is the position of the individual corresponding to the global best fitness; d is the number of spatial dimensions; A + is a random matrix of size 1×d, and each element in the matrix is randomly assigned a value of 1 or -1;

[0087] E. Determine whether to terminate the iteration: If the iteration termination condition is reached, use the parameter values corresponding to the individual sparrow with the highest current fitness as the output parameter values; if the iteration termination condition is not reached, return to step B for the next round of iteration;

[0088] F. Use the output parameter values obtained in step E as the parameter values of the extreme gradient boosting model to construct an initial geological disaster risk prediction model; specifically, it includes the following steps:

[0089] The constructed initial geological disaster risk prediction model is expressed as

[0090]

[0091] In the formula is the log-odds of the initial disaster probability; γ is the log-odds of the initial proportion of disaster samples; L(y i , γ) is the set loss function; N is the number of samples; i represents the sample index.

[0092] The model update process is expressed as

[0093]

[0094] In the formula is the predicted value after the t-th round of iteration; η is the learning rate; f t (x i ) is the predicted output value of the t-th tree;

[0095] The final prediction formula is expressed as

[0096]

[0097] In the formula, T is the total number of trees; ηf t (x i ) is the weighted predicted value of the t-th tree;

[0098] S5. Use the training data set constructed in step S3 to train the initial geological disaster risk prediction model constructed in step S4 to obtain a trained geological disaster risk prediction model;

[0099] S6. Use the geological disaster risk prediction model obtained in step S5 to conduct geological disaster risk prediction for the actual target area.

[0100] The method of the present invention will be further described below in conjunction with an embodiment:

[0101] Compare the method of the present invention with traditional SGBoost algorithm (extreme gradient boosting algorithm), SSA-XGBoost (sparrow algorithm coupled with extreme gradient boosting algorithm), RF (random forest algorithm) and SVM algorithm (support vector machine) algorithms. Use accuracy, precision, recall rate and F1 value as indicators for comparison. The comparison data is shown in Table 1, and the specific comparison curves are as Figure 2 shown:

[0102] Table 1 Schematic table of experimental data comparison

[0103] Accuracy Precision Recall F1 Score The present invention 95.56 95.68 95.16 95.51 SSA-XGBoost 93.99 94.09 93.99 94.76 XGBoost 82.22 82.12 82.22 80.65 RF 81.11 79.64 81.11 75.54 SVM 67.78 66.47 67.78 63.83

[0104] According to Table 1 and Figure 2 it can be seen that the solution of the present invention is significantly higher than the existing solutions in all four indicators, which shows that the solution of the present invention has good reliability and accuracy.

[0105] Figure 3 This is a schematic diagram of the performance comparison between the method of the present invention and the existing SSA-extreme gradient boosting algorithm on unimodal and multimodal test functions. It can be seen from the figure that the method of the present invention has a better initial population than SSA-XGBoost, and thus converges faster during the iteration process. Figure 4 This is a schematic diagram of the search space of the present invention.

[0106] As Figure 5The following is a schematic diagram of the functional modules of the system of the present invention: The system for implementing the geological disaster risk prediction method disclosed in the present invention includes a data acquisition module, an index selection module, a data determination module, a model construction module, a model training module, and a risk prediction module; the data acquisition module, the index selection module, the data determination module, the model construction module, the model training module, and the risk prediction module are connected in series in sequence; the data acquisition module is used to acquire existing geological disaster data information and upload the data information to the index selection module; the index selection module is used to select a number of geological factors as candidate indicators for geological disaster risks according to the received data information and upload the data information to the data determination module; the data determination module is used to perform a correlation analysis on the selected candidate indicators for geological disaster risks according to the received data information to obtain the final geological disaster risk indicators, construct a training data set, and upload the data information to the model construction module; the model construction module is used to construct an initial geological disaster risk prediction model based on the SPM chaotic mapping, the sparrow algorithm, and the extreme gradient boosting algorithm according to the received data information and upload the data information to the model training module; the model training module is used to train the initial geological disaster risk prediction model constructed in step S4 using the constructed training data set according to the received data information to obtain a trained geological disaster risk prediction model and upload the data information to the risk prediction module; the risk prediction module is used to perform the geological disaster risk prediction of the actual target area using the obtained geological disaster risk prediction model according to the received data information.

Claims

1. A geological disaster risk prediction method, comprising the following steps: S1. Obtain existing geological disaster data information; S2. According to the data information obtained in step S1, select several geological factors as candidate indicators for geological disaster risk; S3. Conduct a correlation analysis on the candidate indicators for geological disaster risk selected in step S2 to obtain the final geological disaster risk indicators, and construct a training data set; S4. Based on the SPM chaotic mapping, sparrow algorithm, and extreme gradient boosting algorithm, construct an initial geological disaster risk prediction model; S5. Use the training data set constructed in step S3 to train the initial geological disaster risk prediction model constructed in step S4 to obtain a trained geological disaster risk prediction model; S6. Use the geological disaster risk prediction model obtained in step S5 to predict the geological disaster risk in the actual target area.

2. The geological disaster risk prediction method according to claim 1, wherein The step S2 of selecting several geological factors as candidate indicators for geological disaster risk according to the data information obtained in step S1 specifically includes the following steps: According to the data information obtained in step S1, select several geological factors as candidate indicators for geological disaster risk; The selected geological disaster levels include extremely low incidence, low incidence, medium incidence, and high incidence; The selected candidate indicators for geological disaster risk include slope, aspect, elevation, curvature, topographic relief, distance from river, annual average precipitation, vegetation index, engineering rock group, distance from fault, land cover type, distance from road, soil permeability coefficient, surface humidity index, distance from mining area, and peak ground acceleration of earthquake.

3. The geological disaster risk prediction method according to claim 2, wherein The step S3 of conducting a correlation analysis on the candidate indicators for geological disaster risk selected in step S2 to obtain the final geological disaster risk indicators specifically includes the following steps: Conduct a correlation analysis on the candidate indicators for geological disaster risk selected in step S2 using the Pearson correlation coefficient method; Finally, according to the correlation analysis results, select the final geological disaster risk indicators including slope, elevation, curvature, topographic relief, distance from river, annual average precipitation, vegetation index, engineering rock group, distance from fault, land cover type, and distance from road.

4. The geological hazard risk prediction method according to claim 3, wherein The step S4 of constructing an initial geological disaster risk prediction model based on the SPM chaotic mapping, sparrow algorithm, and extreme gradient boosting algorithm specifically includes the following steps: A. Use the SPM chaotic mapping method to optimize the initial population of the sparrow algorithm; B. Calculate the fitness value of the current population; C. According to the fitness value obtained in step B, divide the sparrows into discoverers and followers; D. The discoverers update their own positions; the followers update their own positions according to the positions of the discoverers and their own states; E. Judge the termination of iteration: If the iteration termination condition is reached, use the parameter values corresponding to the individual sparrow with the highest current fitness as the output parameter values; If the iteration termination condition is not reached, return to step B for the next round of iteration; F. Use the output parameter values obtained in step E as the parameter values of the extreme gradient boosting model, thereby constructing an initial geological disaster risk prediction model.

5. The geological disaster risk prediction method according to claim 4, wherein The SPM chaotic mapping scheme described in step A is used to optimize the initial population of the sparrow algorithm, which specifically includes the following steps: The following formula is used to perform SPM chaotic mapping to optimize the initial population of the sparrow algorithm: where x i is the position of the i-th sparrow; mod() is the modulo operator; η is a set first parameter, and η ∈ (0, 1); μ is a set second parameter, and μ ∈ (0, 1); tur is a random number, and tur ∈ (0, 1).

6. The geological disaster risk prediction method according to claim 5, characterized in that The calculation of the fitness value of the current population described in step B specifically includes the following steps: The accuracy of the extreme gradient boosting model is used as the fitness function, and the fitness value of the current population is calculated.

7. The geological hazard risk prediction method according to claim 6, wherein According to the fitness value obtained in step B, the sparrows are divided into discoverers and followers, which specifically includes the following steps: The fitness values of each individual in the current population obtained in step B are sorted in descending order, and the top n% of the individuals are selected as discoverers, and the remaining individuals are used as followers; n takes a positive value.

8. The geological disaster risk prediction method according to claim 7, characterized in that The discoverers update their own positions, which specifically includes the following steps: The discoverers use the following formula to update their own positions: Wherein: is the position of the i-th discoverer in the j-th dimension at the (t + 1)-th iteration; is the position of the i-th discoverer in the j-th dimension at the t-th iteration; rand is a random number uniformly distributed in the interval [0, 1]; T is the maximum number of iterations; ST is the set safety value; R2 is a random number with a value range of [0, 1]; Q is a random number of standard normal distribution; The followers update their own positions according to the positions of the discoverers and their own states, which specifically includes the following steps: The followers use the following formula to update their own positions: Wherein: is the position of the i1-th follower at the (t + 1)-th iteration in the j-th dimension; is the position of the i1-th follower at the t-th iteration in the j-th dimension; Q is a random number from the standard normal distribution; is the position of the individual corresponding to the global worst fitness; n is the number of joiners in the sparrow algorithm; is the position of the individual corresponding to the global best fitness; d is the number of spatial dimensions; A + is a random matrix of size 1×d, and each element in the matrix is randomly taken, and the value is 1 or -1.

9. The geological disaster risk prediction method according to claim 8, wherein The construction of the initial model for predicting geological disaster risks described in step F specifically includes the following steps: The constructed initial model for predicting geological disaster risks is expressed as In the formula is the logit of the initial disaster probability; γ is the logit of the proportion of the initial disaster samples; L(y i , γ) is the set loss function; N is the number of samples; i represents the sample index. The model update process is expressed as where is the predicted value after the t-th round of iteration; η is the learning rate; f t (x i ) is the predicted output value of the t-th tree; The final prediction formula is expressed as where T is the total number of trees; ηf t (x i ) is the weighted prediction value of the t-th tree.

10. A system for implementing the geological disaster risk prediction method according to any one of claims 1 to 9, characterized in that It includes a data acquisition module, an index selection module, a data determination module, a model construction module, a model training module, and a risk prediction module; the data acquisition module, the index selection module, the data determination module, the model construction module, the model training module, and the risk prediction module are connected in series in sequence; the data acquisition module is used to acquire existing geological disaster data information and upload the data information to the index selection module; The index selection module is used to select several geological factors as candidate indicators for geological disaster risks according to the received data information, and upload the data information to the data determination module; The data determination module is used to perform a correlation analysis on the selected candidate indicators for geological disaster risks according to the received data information to obtain the final geological disaster risk indicators, construct a training data set, and upload the data information to the model construction module; The model construction module is used to construct an initial model for predicting geological disaster risks based on the SPM chaotic mapping, the sparrow algorithm, and the extreme gradient boosting algorithm according to the received data information, and upload the data information to the model training module; The model training module is used to train the initial model for predicting geological disaster risks constructed in step S4 using the constructed training data set according to the received data information to obtain a trained model for predicting geological disaster risks, and upload the data information to the risk prediction module; The risk prediction module is used to perform actual geological disaster risk prediction for the target area using the obtained model for predicting geological disaster risks according to the received data information.

Images