Mine rockburst engineering risk prediction method based on random forest and analytic hierarchy process

Abstract

The invention belongs to the field of engineering risk prediction, and provides a mine rockburst engineering risk prediction method based on a random forest and an analytic hierarchy process, and the method comprises the steps: carrying out the data collection of a to-be-predicted sample, and carrying out the calculation of a comprehensive attribution degree formula; the construction process of the comprehensive attribution degree formula comprises the steps of rock burst evaluation index determination, rock burst engineering case database construction, index calculation set acquisition, attribution degree data set acquisition, data preprocessing, random forest model prediction, weight calculation and parameter fusion. According to the method, through the random forest model and the analytic hierarchy process, the prediction weight of each index is obtained, the accuracy of the comprehensive attribution degree formula is improved, and a more accurate method is provided for evaluation and judgment of the engineering rockburst risk grade.

Description

Technical Field

[0001] The present invention relates to the field of engineering risk prediction, and in particular to a method for predicting the risk of mine rockburst engineering based on random forest and analytic hierarchy process. Background Art

[0002] Rockburst is a special rock mass destruction phenomenon, which usually occurs in rock mass or geological structures with high ground stress or limit equilibrium state. Under the disturbance of excavation activities, the stress energy stored in the rock mass will be released instantly, causing part of the rock mass around the excavation space to protrude or eject sharply and violently from the parent rock mass. With the increase of mining depth, ground stress and rock strength gradually increase, which increases the possibility and harm of rockburst. In addition, factors such as the physical and mechanical properties of the rock mass, geological structure characteristics and groundwater will also affect the occurrence of rockburst.

[0003] However, traditional methods generally only use one method for prediction and judgment, resulting in poor prediction of mine rockburst engineering risks. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a mine rockburst engineering risk prediction method based on random forest and hierarchical analysis method. Through the random forest model and hierarchical analysis method, the prediction weights of each indicator are obtained to improve the accuracy of the comprehensive attribution formula.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A mine rockburst engineering risk prediction method based on random forest and analytic hierarchy process, including:

[0007] Collect sample data to be predicted;

[0008] Calculate the sample data to be predicted using a preset comprehensive attribution degree formula to obtain a target prediction result;

[0009] The process of obtaining the comprehensive attribution degree formula includes:

[0010] Determine rockburst evaluation indicators;

[0011] Build a rockburst engineering case database based on historical data, field research, and expert consultation;

[0012] Calculating the rockburst evaluation index according to the rockburst engineering case database to obtain an index calculation set;

[0013] Inputting the indicator calculation set into a preset forward Gaussian cloud generator to obtain an attribution degree data set;

[0014] Cleaning, converting and standardizing the attribution degree data set to obtain preprocessed data;

[0015] Inputting the preprocessed data into a preset random forest model for prediction to obtain predicted data;

[0016] A judgment matrix is ​​constructed according to the hierarchical analysis method, and the prediction data is weighted by using the judgment matrix to obtain a rockburst index evaluation weight set;

[0017] The rockburst index evaluation weight set and the attribution degree data set are fused to obtain the comprehensive attribution degree formula.

[0018] Preferably, the rockburst evaluation indexes include: maximum tangential stress of surrounding rock, elastic energy index, rock brittleness index and Russense criterion.

[0019] Preferably, the rockburst engineering case database includes: engineering background, burial depth, rock sample type, maximum shear stress of surrounding rock, uniaxial compressive strength, uniaxial tensile strength, elastic energy index and rockburst risk level.

[0020] Preferably, the rockburst index evaluation weight set and the attribution degree data set are fused to obtain the comprehensive attribution degree formula, which also includes

[0021] Using a preset verification set to verify the comprehensive attribution degree formula, to obtain a verification error set;

[0022] The parameters of the random forest model or the judgment matrix are optimized according to the validation error set.

[0023] Preferably, the comprehensive attribution degree formula is: F is the calculation result of the comprehensive attribution formula; μ p The attribution data of the pth indicator in the indicator calculation set; w p is the weight data of the pth indicator in the indicator calculation set; m is the total number of indicators in the indicator calculation set.

[0024] Preferably, the indicator calculation set is input into a preset forward Gaussian cloud generator to obtain an attribution degree data set, including:

[0025] The rockburst-free level is calculated using the semi-cloud digital characteristic formula to obtain the risk level attribution calculation result; the calculation formula of the semi-cloud digital characteristic formula is: Among them, C max is the first maximum rockburst level; C min is the first minimum rockburst level; k h is the first constant; E xh is the first expectation; E nhis the first entropy; H eh is the first super entropy;

[0026] The full-cloud digital characteristic formula is used to calculate the data in the rockburst engineering case database except the rockburst risk level to obtain the remaining attribution calculation results; the calculation formula of the full-cloud digital characteristic formula is: Among them, E xq is the second expectation; E nq is the second entropy; H eq is the second super entropy; C maxq The second largest rockburst level; C minq is the second minimum rockburst level; k q is the second constant;

[0027] The risk level attribution degree calculation result and the other attribution degree calculation results are integrated to obtain the attribution degree data set.

[0028] The present invention discloses the following technical effects:

[0029] The present invention provides a method for predicting the risk of mine rockburst engineering based on random forest and analytic hierarchy process. Through the random forest model and analytic hierarchy process, the problem of poor accuracy of the comprehensive attribution degree formula is solved, and high-precision acquisition of the prediction weights of each indicator is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0031] Figure 1 A schematic diagram of a mine rockburst engineering risk prediction process based on random forest and analytic hierarchy process provided in an embodiment of the present invention;

[0032] Figure 2 A flowchart of the random forest-hierarchy analysis method implementation provided by the embodiment of the present invention;

[0033] Figure 3 The evaluation indicators provided in the embodiment of the present invention correspond to the attribution cloud diagrams of different rockburst levels. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] The purpose of the present invention is to provide a method for predicting the risk of mine rockburst engineering based on random forest and analytic hierarchy process. The prediction weights of various indicators are obtained through random forest model and analytic hierarchy process to improve the accuracy of comprehensive attribution formula.

[0036] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0037] Figure 1 A schematic diagram of a mine rockburst engineering risk prediction process based on random forest and analytic hierarchy process provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the present invention provides a method for predicting the risk of mine rock burst engineering based on random forest and hierarchical analysis method, including:

[0038] Step 100: Collect sample data to be predicted;

[0039] Step 200: Calculate the sample data to be predicted using a preset comprehensive attribution degree formula to obtain a target prediction result;

[0040] The process of obtaining the comprehensive attribution formula includes:

[0041] Determine rockburst evaluation indicators;

[0042] Build a rockburst engineering case database based on historical data, field research, and expert consultation;

[0043] The rockburst evaluation indexes are calculated according to the rockburst engineering case database to obtain the index calculation set;

[0044] Input the indicator calculation set into the preset forward Gaussian cloud generator to obtain the attribution degree data set;

[0045] Clean, transform and standardize the attribution data set to obtain preprocessed data;

[0046] Input the preprocessed data into the preset random forest model for prediction to obtain the predicted data;

[0047] According to the hierarchical analysis method, a judgment matrix is ​​constructed, and the judgment matrix is ​​used to calculate the weight of the predicted data to obtain the rockburst index evaluation weight set;

[0048] The rockburst index evaluation weight set and the attribution degree data set are fused to obtain the comprehensive attribution degree formula.

[0049] Specifically, the rockburst evaluation indicators include: maximum tangential stress of surrounding rock, elastic energy index, rock brittleness index and Russense criterion.

[0050] Specifically, the rock burst engineering case database includes: engineering background, burial depth, rock sample type, maximum shear stress of surrounding rock, uniaxial compressive strength, uniaxial tensile strength, elastic energy index and rock burst risk level.

[0051] Preferably, the rockburst index evaluation weight set and the attribution degree data set are fused to obtain a comprehensive attribution degree formula, which also includes

[0052] The comprehensive attribution degree formula is verified using a preset verification set to obtain a verification error set;

[0053] The parameters or judgment matrix of the random forest model are optimized based on the validation error set.

[0054] Furthermore, the comprehensive attribution formula is: F is the calculation result of the comprehensive attribution formula; μ p is the attribution data of the pth indicator in the indicator calculation set; w p is the weight data of the pth indicator in the indicator calculation set; m is the total number of indicators in the indicator calculation set.

[0055] Specifically, the indicator calculation set is input into the preset forward Gaussian cloud generator to obtain the attribution degree data set, including:

[0056] The non-rockburst grade is calculated using the semi-cloud digital characteristic formula to obtain the risk grade attribution calculation result; the calculation formula of the semi-cloud digital characteristic formula is: Among them, C max is the first maximum rockburst level; C min is the first minimum rockburst level; k h is the first constant; E xh is the first expectation; E nh is the first entropy; H eh is the first super entropy;

[0057] The full-cloud digital characteristic formula is used to calculate the data in the rockburst engineering case database except the rockburst risk level to obtain the remaining attribution calculation results; the calculation formula of the full-cloud digital characteristic formula is: Among them, E xq is the second expectation; E nq is the second entropy; H eq is the second super entropy; C maxq The second largest rockburst level; Cminq is the second minimum rockburst level; k q is the second constant;

[0058] The risk level attribution degree calculation results and the other attribution degree calculation results are integrated to obtain the attribution degree data set.

[0059] Preferably, the indicators selected for rockburst evaluation must be the core factors controlling the occurrence of rockburst and have a significant impact on the occurrence of rockburst. Each indicator should be independent of each other to ensure the objectivity and accuracy of the evaluation indicator system, and the selected evaluation indicators should be common, universal, and quantitatively reflect the tendency of rockburst, avoid overly complex evaluation indicators, improve the operability of the evaluation indicator system, and facilitate data collection and statistics. Therefore, this embodiment takes into account the inherent mechanical properties, stress state, and performance of the rock under a specific stress state and the characteristics of energy storage and release, and selects four indicators, namely, the maximum tangential stress of the surrounding rock, the elastic energy index, the rock brittleness index, and the Russense criterion, to reflect the rockburst tendency of the rock from different angles, so as to improve the accuracy and reliability of the prediction.

[0060] Furthermore, data collection and rockburst engineering case analysis were carried out through historical case collection, on-site investigation and expert consultation. Based on existing rockburst engineering cases and rockburst evaluation indicators, the engineering background, burial depth H, rock sample type, and surrounding rock maximum shear stress σ were selected. θ , uniaxial compressive strengthσ c , uniaxial tensile strength σ t , elastic energy index W et , rock burst risk level to build a rock burst engineering case database. The case database of this embodiment contains 295 groups of engineering data (44 groups are listed in Table 1), and the parameters in the database are used to provide data support for the preset random forest-hierarchy analysis method rock burst risk assessment.

[0061] Table 1

[0062]

[0063]

[0064]

[0065] refer to Figure 2Random forest and analytic hierarchy process are two completely different data analysis methods. The former is an integrated learning method in machine learning, while the latter is a decision-making tool used to deal with complex multi-criteria decision-making problems. Combining the two, the feature importance evaluation function of random forest can be used to determine the degree of influence of each indicator on the prediction target. These objective data-driven results can be used as a reference for the construction of the judgment matrix in the analytic hierarchy process. At the same time, the introduction of expert knowledge and experience in the analytic hierarchy process can reasonably explain and verify the prediction results of the random forest model. Random forest-analytic hierarchy process implementation flow chart reference Figure 2 , the specific steps are as follows:

[0066] 1) Problem definition and data collection: clarify the problem to be solved, collect relevant data, and determine the indicators or features for analysis and modeling;

[0067] 2) Data preprocessing: cleaning, converting and standardizing the collected data to facilitate subsequent analysis;

[0068] 3) Random forest model construction: Use the processed data to train the random forest model. Adjust the model parameters to optimize the model performance. Use the feature importance function of the random forest model to analyze the contribution of each feature to the model prediction;

[0069] 4) Application of the Analytical Hierarchy Process (AHP): Construct an indicator hierarchy and determine the corresponding evaluation objectives, criteria, and schemes. Establish a judgment matrix based on the importance of the features of random forests, calculate the weights of each level, and perform consistency checks;

[0070] 5) Model validation and optimization: Use an independent validation set to validate the performance of the model. Adjust model parameters or readjust AHP weights based on the validation results to optimize model performance.

[0071] Preferably, the cloud model theory is used to describe complex systems of uncertainty and ambiguity, and is a mathematical model based on probability statistics. In the cloud model, each element is regarded as a cloud droplet, and the cloud composed of these cloud droplets can represent the mapping process from a qualitative concept to a quantitative description. The core idea of ​​the cloud model is to divide uncertainty into three aspects: certainty, randomness, and fuzziness, and to describe these uncertainty characteristics by establishing membership functions and cloud functions. This embodiment introduces the cloud model theory, combined with the random forest-hierarchical analysis algorithm, to map different rock burst intensity levels into Gaussian clouds, and determine the degree of attribution of the sample to a certain rock burst intensity level based on this, as follows:

[0072] 1) Select the index σ for evaluating rockburst θ , W et , B, R, and refer to the rockburst intensity standard to calculate the cloud numerical characteristics of each indicator and obtain the entropy and super entropy values;

[0073] 2) Input the entropy and super entropy values ​​into the forward Gaussian cloud generator to generate Gaussian cloud droplets, and then calculate the degree of attribution of each evaluation index relative to different rockburst intensities;

[0074] 3) Use the random forest-hierarchy analysis algorithm to evaluate the weights of rockburst indicators;

[0075] 4) Read the example data of rockburst prediction, calculate the weight of each evaluation index and their degree of attribution to different rockburst levels; multiply the weight by the degree of attribution to obtain a comprehensive degree of attribution; the rockburst intensity level of the predicted sample is determined according to the rockburst level corresponding to the maximum value of the comprehensive degree of attribution obtained by different evaluation indicators. Generate a Gaussian cloud map according to the grade standards of different rockburst evaluation indicators listed in Table 1. To ensure that the Gaussian cloud map is in the positive interval of the horizontal coordinate, the half-cloud digital characteristic formula is used for the interval without rockburst level data, and the full-cloud digital characteristic formula is used for other rockburst level data intervals. The calculation results are shown in Table 2. The calculation formula includes:

[0076] Full cloud digital feature type:

[0077]

[0078] Digital characteristic formula of semi-falling clouds:

[0079]

[0080] Table 2

[0081]

[0082] Furthermore, Table 3 shows the grade standards of different rockburst evaluation indicators:

[0083] Table 3

[0084]

[0085]

[0086] C max , C min are the maximum and minimum values ​​of each indicator corresponding to different rockburst levels; E x is the expectation, the expectation of the distribution of cloud droplets in space, which is the point that can represent the qualitative concept and is the most typical sample for the quantification of this concept; n is entropy, representing the range of cloud droplets that can be accepted by the concept; H e is the super entropy, representing the discrete degree of cloud droplets, and the constant k is usually taken as 0.01. The calculation results of risk level attribution degree and other attribution degree calculation results are integrated to obtain the attribution degree data set.

[0087] refer to Figure 3 , the maximum tangential stress of surrounding rock σ θ , elastic energy index W et , rock brittleness index B, Russenes criterion R corresponding to the degree of cloud map reference Figure 3 According to the results of random forest algorithm, the importance of different evaluation indicators to rockburst risk level is calculated according to the random forest algorithm, and the results are shown in Table 4. As shown in Table 4, the importance of the four rockburst evaluation indicators is: θ >W et >B>R. A hierarchical analysis matrix is ​​constructed according to the importance of the evaluation indicators, and the weights of different indicators are calculated.

[0088] Table 4

[0089]

[0090] Specifically, σ θ , W et , B, R evaluation index weight calculation process:

[0091]

[0092] Characteristic root λ calculation:

[0093]

[0094] Among them, n is the number of indicators, W i is the eigenvector, (BW) i is the i-th component of the vector BW;

[0095] Consistency index CI test:

[0096]

[0097] Consistency ratio CR test:

[0098]

[0099] The consistency test is passed, σ θ , W et The corresponding weights of the A, B, and R evaluation indicators are 0.546, 0.268, 0.125, and 0.061, respectively.

[0100] Furthermore, combined with the weight values ​​of the rockburst evaluation indicators, the following formula is used to obtain the comprehensive attribution degree:

[0101]

[0102] F is the calculation result of the comprehensive attribution formula; μ pis the attribution data of the pth indicator in the indicator calculation set; w p is the weight data of the pth indicator in the indicator calculation set; m is the total number of indicators in the indicator calculation set.

[0103] Preferably, the effectiveness of the rockburst risk prediction model is verified: In the rockburst prediction model of the random forest-hierarchy analysis method, since the prediction results have four levels (no rockburst, slight rockburst, moderate rockburst, and severe rockburst), the traditional binary classification confusion matrix cannot be directly applied to calculate true positive examples, false positive examples, true negative examples, and false negative examples. In multi-classification problems, these indicators need to be calculated separately for each category, and the problem is converted into multiple binary classification problems. This embodiment takes the "no rockburst" level as an example: first, no rockburst is regarded as a positive class, and all other categories (slight rockburst, moderate rockburst, and severe rockburst) are regarded as negative classes, and then a binary classification confusion matrix is ​​constructed for this category, which contains the following four contents:

[0104] True positive TP: the number of samples that are actually “no rock burst” and predicted as “no rock burst”;

[0105] False positive examples FP: the number of samples that are actually other rockburst levels but are predicted as “no rockburst”;

[0106] True negative TN: the number of samples that are actually other rockburst levels and are also predicted to be other rockburst levels;

[0107] False negative FN: the number of samples that are actually “no rockburst” but predicted to be other rockburst levels;

[0108] The confusion matrix parameters of other rockburst levels are also calculated according to the above definition process. In order to verify the effectiveness of the random forest-hierarchy analysis method rockburst prediction model, it is necessary to select multiple indicators to comprehensively evaluate the performance of the model: accuracy, precision, recall and F1 value. A confusion matrix is ​​established based on the prediction results to show the classification ability of the model for different categories. The details are as follows:

[0109] 1) Accuracy

[0110] Accuracy is the ratio of the number of samples correctly predicted by the model to the total number of samples. A high accuracy indicates that the model is able to classify samples well overall.

[0111]

[0112] 2) Precision

[0113] Precision refers to the proportion of samples that are actually positive when the model predicts them to be positive. A high precision indicates that when the model predicts that a sample is positive, the prediction is likely to be correct.

[0114]

[0115] 3) Recall

[0116] Recall refers to the proportion of actual positive samples that are correctly predicted by the model. A high recall rate indicates that the model can find most of the true positive samples.

[0117]

[0118] 4) F1 value (F 1-score )

[0119] The F1 value is the harmonic mean of precision and recall, and is used to comprehensively evaluate the performance of the model. A high F1 value indicates that the model performs well in both precision and recall, that is, if the F1 value is close to 1, then the model is both accurate and comprehensive in predicting positive examples.

[0120]

[0121] This embodiment selects 50 groups of actual engineering points and 50 engineering points. By comparing the deep mine rockburst risk prediction level with the actual engineering risk level (Table 5), the accuracy, precision, recall and F1 value corresponding to different rockburst level evaluations are calculated (Table 6). By comparing the prediction results of the random forest-hierarchy analysis method rockburst prediction and the hierarchical analysis method rockburst prediction in Table 6, the accuracy evaluation index of the former is greater than that of the latter, that is, the accuracy of rockburst risk prediction has been greatly improved after the introduction of the random forest algorithm. Therefore, the rockburst risk prediction model established in this embodiment has good adaptability to the deep engineering rockburst risk prediction and high accuracy.

[0122] Table 5

[0123]

[0124]

[0125]

[0126]

[0127] Table 6

[0128]

[0129] Among them, ① represents the random forest-hierarchy analysis method rockburst prediction, and ② represents the hierarchy analysis method rockburst prediction.

[0130] Specifically, this embodiment uses machine learning technology to select four indicators, namely, the maximum tangential stress of the surrounding rock, the elastic energy index, the rock brittleness index, and the Russenes criterion, for importance and objective quantitative analysis, and establishes a rockburst evaluation indicator importance analysis model based on random forests; combines the cloud model theory to determine the rockburst intensity level attribution of different evaluation indicators, and finally obtains the rockburst intensity level; combines 50 groups of actual engineering point information to calculate the rockburst occurrence risk, introduces four machine learning evaluation indicators, namely, accuracy, precision, recall rate, and F1 value, to verify the accuracy and practicality of the rockburst prediction model established in this embodiment. This embodiment explains the definition and basic principles of random forests, the basic process of step-by-step calculation, decision tree learning, and the construction of random forests; explains the basic principles and distribution calculation process of the analytic hierarchy process, combines the importance evaluation of the random forest algorithm with the analytic hierarchy process weight calculation, and proposes a random forest-analytic hierarchy process calculation algorithm. This embodiment obtains the importance scores of four rockburst evaluation indicators through the rockburst evaluation indicator importance analysis model of random forests, which are σ θ , W et , B and R; then, the analysis matrix in the hierarchical analysis method is constructed according to the importance of different indicators, and the hierarchical analysis method is optimized; finally, the cloud model theory is introduced to judge the degree of attribution of different evaluation indicators to obtain the final rockburst intensity level. This embodiment uses the random forest-hierarchical analysis method to calculate the rockburst risk of 50 sets of actual engineering point information, introduces four machine learning accuracy evaluation indicators of accuracy, precision, recall rate and F1 value to calculate, and obtains the prediction accuracy of each rockburst intensity. The accuracy of the random forest-hierarchical analysis rockburst prediction method and the hierarchical analysis rockburst prediction method are compared and analyzed to ensure the accuracy and practicality of the rockburst prediction model established in this embodiment.

[0131] The beneficial effects of the present invention are as follows:

[0132] The present invention combines random forest with analytic hierarchy process and utilizes the advantages of the two methods to jointly decide the risk level. Random forest and analytic hierarchy process are two completely different data analysis methods, the former is an integrated learning method in machine learning, and the latter is a decision-making tool for dealing with complex multi-criteria decision-making problems. By combining the two, the feature importance evaluation function of random forest can be used to determine the influence of each indicator on the prediction target, and these objective data-driven results can be used as a reference for the construction of the judgment matrix in the analytic hierarchy process. At the same time, the introduction of expert knowledge and experience in the analytic hierarchy process can reasonably explain and verify the prediction results of the random forest model.

[0133] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0134] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A mine rockburst engineering risk prediction method based on random forest and hierarchical analysis method, characterized in that: include: Collect sample data to be predicted; Calculate the sample data to be predicted using a preset comprehensive attribution degree formula to obtain a target prediction result; The process of obtaining the comprehensive attribution degree formula includes: Determine rockburst evaluation indicators; Build a rockburst engineering case database based on historical data, field research, and expert consultation; Calculating the rockburst evaluation index according to the rockburst engineering case database to obtain an index calculation set; Inputting the indicator calculation set into a preset forward Gaussian cloud generator to obtain an attribution degree data set; Cleaning, converting and standardizing the attribution degree data set to obtain preprocessed data; Inputting the preprocessed data into a preset random forest model for prediction to obtain predicted data; A judgment matrix is ​​constructed according to the hierarchical analysis method, and the prediction data is weighted by using the judgment matrix to obtain a rockburst index evaluation weight set; The rockburst index evaluation weight set and the attribution degree data set are fused to obtain the comprehensive attribution degree formula.

2. A method for predicting the risk of mine rockburst engineering based on random forest and analytic hierarchy process according to claim 1, characterized in that: The rockburst evaluation indexes include: maximum tangential stress of surrounding rock, elastic energy index, rock brittleness index and Russense criterion.

3. The method for predicting the risk of mine rockburst engineering based on random forest and analytic hierarchy process according to claim 1 is characterized in that: The rock burst engineering case database includes: engineering background, burial depth, rock sample type, maximum shear stress of surrounding rock, uniaxial compressive strength, uniaxial tensile strength, elastic energy index and rock burst risk level.

4. The method for predicting the risk of mine rockburst engineering based on random forest and analytic hierarchy process according to claim 1, characterized in that: The rockburst index evaluation weight set and the attribution degree data set are merged to obtain the comprehensive attribution degree formula, which also includes Using a preset verification set to verify the comprehensive attribution degree formula, to obtain a verification error set; The parameters of the random forest model or the judgment matrix are optimized according to the validation error set.

5. The method for predicting the risk of mine rockburst engineering based on random forest and analytic hierarchy process according to claim 1, characterized in that: The comprehensive attribution degree formula is: F is the calculation result of the comprehensive attribution formula; μ p Calculating the attribution data of the pth indicator in the indicator calculation set; w p is the weight data of the pth indicator in the indicator calculation set; m is the total number of indicators in the indicator calculation set.

6. The method for predicting the risk of mine rockburst engineering based on random forest and analytic hierarchy process according to claim 3 is characterized in that: The indicator calculation set is input into a preset forward Gaussian cloud generator to obtain a belonging degree data set, including: The rockburst-free level is calculated using the semi-cloud digital characteristic formula to obtain the risk level attribution calculation result; the calculation formula of the semi-cloud digital characteristic formula is: Among them, C max is the first maximum rockburst level; C min is the first minimum rockburst level; k h is the first constant; E xh is the first expectation; E nh is the first entropy; H eh is the first super entropy; The full-cloud digital characteristic formula is used to calculate the data in the rockburst engineering case database except the rockburst risk level to obtain the remaining attribution calculation results; the calculation formula of the full-cloud digital characteristic formula is: Among them, E xq is the second expectation; E nq is the second entropy; H eq is the second super entropy; C maxq The second largest rockburst level; C minq is the second minimum rockburst level; k q is the second constant; The risk level attribution degree calculation result and the remaining attribution degree calculation results are integrated to obtain the attribution degree data set.

Images