CN113762394B - Blasting block prediction method - Google Patents
Blasting block prediction method Download PDFInfo
- Publication number
- CN113762394B CN113762394B CN202111054833.2A CN202111054833A CN113762394B CN 113762394 B CN113762394 B CN 113762394B CN 202111054833 A CN202111054833 A CN 202111054833A CN 113762394 B CN113762394 B CN 113762394B
- Authority
- CN
- China
- Prior art keywords
- random forest
- model
- training
- layer
- blasting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000005422 blasting Methods 0.000 title claims abstract description 52
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000007637 random forest analysis Methods 0.000 claims abstract description 63
- 238000012549 training Methods 0.000 claims abstract description 58
- 238000003066 decision tree Methods 0.000 claims abstract description 31
- 238000012952 Resampling Methods 0.000 claims abstract description 4
- 238000012360 testing method Methods 0.000 claims description 20
- 230000011218 segmentation Effects 0.000 claims description 12
- 239000011435 rock Substances 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000005553 drilling Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 6
- 230000000694 effects Effects 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 230000000903 blocking effect Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000011065 in-situ storage Methods 0.000 claims description 4
- 238000004088 simulation Methods 0.000 claims description 3
- 238000004880 explosion Methods 0.000 description 7
- 238000002474 experimental method Methods 0.000 description 3
- 230000010365 information processing Effects 0.000 description 3
- 241001529734 Ocimum Species 0.000 description 2
- 235000010676 Ocimum basilicum Nutrition 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000013138 pruning Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035939 shock Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Evolutionary Computation (AREA)
- General Business, Economics & Management (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Marketing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Agronomy & Crop Science (AREA)
- Health & Medical Sciences (AREA)
- Mining & Mineral Resources (AREA)
- Primary Health Care (AREA)
- Marine Sciences & Fisheries (AREA)
- Animal Husbandry (AREA)
- General Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a blasting block size prediction method, and belongs to the technical field of blasting block size prediction. Firstly, constructing a CART decision tree model; then resampling the training sample set; calculating the relation between different decision tree trees (Ntree) and model Mean Square Error (MSE) by utilizing MATLAB script language writing codes; then adopting GINI coefficients as attribute splitting criteria for the decision tree, and selecting nodes based on the criteria; based on the training set of the nodes, different decision trees are obtained, and a first layer of random forest is finally generated; generating a second layer random forest according to the training residual error of the first layer random forest as a training sample set of the second layer random forest; and sequentially superposing the outputs of the first layer random forest model and the second layer random forest model to obtain the final output of the double-layer random forest prediction model, and predicting the blasting block according to the final output. Compared with the traditional prediction method, the method has higher reliability and stability.
Description
Technical Field
The invention relates to a blasting block size prediction method, and belongs to the technical field of blasting block size prediction.
Background
The drilling and blasting method is one of the rock breaking methods frequently adopted in mining and tunnel construction, and the blasting effect brings about a plurality of negative effects such as blasting vibration, air shock waves, flying stones, noise, dust and the like while breaking the rock, wherein the blasting vibration hazard is particularly remarkable. The explosion regulations in China adopt the speed to measure the vibration intensity, accurately predict the particle vibration speed caused by explosion and effectively control the harm of explosion vibration.
The accuracy and reliability of the blasting block prediction method in the prior art are low.
Disclosure of Invention
The invention aims to solve the technical problems that: the blasting block degree prediction method is provided, the problem of inaccurate and fuzzy information processing which needs to consider a plurality of factors and conditions simultaneously is solved, and the accuracy and reliability of blasting block degree prediction are improved.
The technical scheme adopted by the invention is as follows: a blasting block prediction method, comprising the steps of:
1) Forming a training data set;
2) Selecting a split node;
3) Generating a first layer of random forest;
4) Generating a second layer of random forest;
5) And sequentially superposing the outputs of the first layer of random forest and the second layer of random forest to obtain the final output of the double-layer random forest prediction model.
Specifically, the step 1) includes the steps of:
Step1.1: taking the blasting average block size value actually measured in a blasting test as an output variable (label value) of a model, acquiring a resistance line distance (B), a drilling row distance (S), a step height (H), a drilling diameter (d), a blocking length (L), a rock mass elastic modulus (E), an in-situ rock block size (x) and a blasting unit consumption (q) index as input variables (attribute values) of the model by corresponding blasting test occasions, and forming a training data set by the output variables and the input variables
Step1.2: the data set obtained by using Step1.1 is subjected to resampling of training sample set by adopting Bootstrap method, m data samples are extracted, and n training data sets are randomly generatedD= { x i1,xi2,…,xin,yi } (i e1, m).
Based on a Bootstrap data set, a Bagging method is adopted to randomly select training data with a put back, then a classifier is constructed, and finally the overall acquisition effect is increased by combining the learned models.
Step1.3: using a decision tree algorithm in each subset, selecting an optimal mode to split nodes according to a 'minimum coefficient criterion', and setting the prediction result of a single decision tree predictor f (x, theta k) as f i (x) without pruning in the splitting process, wherein the final prediction result of a random forest regression model is expressed as that a parameter set RFP of a modeling process of the random forest regression algorithm is defined in a formula (1); equation (2) is used to make a prediction of blasting block based on random forest regression.
RFP＝{Ntree,Mtry} (1)
Where x represents the input vector, θ k is the vector representing the generation of each tree growth path, and F (x) represents the predicted blast block. Ntree is the number of decision trees in the model, and Ntree values affect the training degree and accuracy of the random forest model. In order to obtain the best evaluation result, codes are written by using MATLAB script language, and the relation between different decision tree trees (Ntree) and the model Mean Square Error (MSE) is simulated and calculated. Mtry is the number of features randomly extracted from the features. The Mtry value controls the disturbance degree of the random forest model attribute and is an important parameter in the model. The Mtry value directly affects the accuracy of the model. The Mtry values were calculated according to the following empirical formula:
Mtry&＝[log2M] (3)
Mtry&＝[M/3 (4)
wherein: m is the number of model input parameters; [] Representing a rounding down operation.
Specifically, the step 2) includes the steps of:
step2.1: a binary tree form decision tree is adopted, and a binary recursion is utilized to divide the data space into different subsets. In classification, assuming that there are K classes, the base index of the probability distribution is:
where p represents the probability of the sample point and p k represents the probability that the sample point belongs to the K-th class.
Step2.2: the decision tree uses GINI coefficients as criteria for attribute splitting. The feature with the lowest coefficient of the base is selected as the root node. And so on, selecting the leaf node with the smallest characteristic coefficient.
Step2.3: from the training dataset, starting from the root node, step 3 is recursively performed for each node, constructing a binary decision tree.
Specifically, the step 3) includes the steps of:
step3.1: the subset D is divided into two parts D 1 and D 2 according to whether the feature a takes a certain possible value a, D 1＝{(x,y)∈D|A(x)＝a},D2＝D-D1, and under the condition of the feature a, the base index of the set D is defined as:
where D 1 represents the training set of nodes, and the Basil index of the existing feature to the dataset is calculated.
Step3.2: for each feature a, for each possible value a, the base index at a=a is calculated by dividing D into two parts D 1 and D 2, depending on whether the test of the sample point pair a=a is yes or no. Among all possible features a and all possible cut points a, the cut point with the smallest keni index is selected as the cut point. Generating two child nodes from the current segmentation point according to the optimal characteristics and the optimal segmentation point, and distributing the training data set to two leaf nodes according to the characteristics.
Step3.3: step2.3 to Step3.2 are invoked recursively for both leaf nodes.
Step3.4: repeating the steps 3.1 to 3.3 to obtain different decision trees, recursively branching and growing each decision tree from top to bottom, stopping growing the regression tree after the segmentation termination condition is met, and finally combining all the regression trees together to form a random forest.
Step3.5: the input vector X in the prediction set X is input into a first layer random forest, and arithmetic average is carried out on the predicted value y i of the single regression tree T i by using a simple average method
Specifically, the step 4) includes the steps of:
step4.1: calculating the output of a layer 1 random forest on a training sample set Subtracting the output value/>, from the actual value y i of the training sampleTraining residual error/>, can be obtainedTraining residual/>Substituting the original training sample to construct a new data set as a training sample of a layer 2 random forest, so that the input is x i, and the expected output is/>
Step4.2: because the training residual value obtained by Step4.1 is too small, the training residual needs to be normalized:
Wherein: l in is the original input data; l min is the minimum of the same kind of data in the original input data; l max is the maximum value of the same kind of data in the original input data; l is the input data after normalization processing. And after normalization treatment, a training data set of the layer 2 random forest is formed.
Step4.3: and (3) establishing a layer 2 random forest model according to Step1.1 to Step4.1 to obtain a prediction result of the layer 2 random forest model.
The beneficial effects of the invention are as follows: the invention is particularly suitable for dealing with imprecise and ambiguous information processing problems requiring simultaneous consideration of many factors and conditions, and has higher reliability and stability than conventional prediction methods.
Drawings
FIG. 1 is a flowchart of a blast block predictive value calculation based on random forest regression;
FIG. 2 is a schematic diagram of a step blasting parameter;
FIG. 3 is a plot of a matrix of raw data set distribution points;
FIG. 4 is a comparison of measured and predicted blocking values for a burst test;
FIG. 5 shows the relative error between measured and predicted block sizes for a burst test.
Detailed Description
The invention will be further described with reference to the drawings and the specific embodiments.
Example 1: the random forest is used as an advanced algorithm tool, is particularly suitable for processing the imprecise and fuzzy information processing problem which needs to consider a plurality of factors and conditions at the same time, introduces a random forest regression method into the blasting vibration speed prediction, and compares the result of the random forest model prediction with the result of the traditional method prediction so as to provide an optimal prediction method for blasting block monitoring.
As shown in fig. 1, a blasting block prediction method includes the following steps:
1) Forming a training data set;
2) Selecting a split node;
3) Generating a first layer of random forest;
4) Generating a second layer of random forest;
5) And sequentially superposing the outputs of the first layer of random forest and the second layer of random forest to obtain the final output of the double-layer random forest prediction model.
The specific implementation process is as follows: in order to solve the problems of low model prediction precision, poor model generalization capability and the like in the research on the prediction of the block degree of blasting materials at the present stage, the block degree of the pile is difficult to accurately control, and the block degree meets the dam-loading conditions of the pile for blasting mining. Aiming at the defects of the current blasting prediction model, the blasting block size of the rock-fill dam material is effectively controlled, and the blasting block size prediction model is established based on a random forest regression method.
Further, the step 1) includes the steps of:
Step1.1: taking the blasting average block size value actually measured in a blasting test as an output variable (label value) of a model, acquiring a resistance line distance (B), a drilling row distance (S), a step height (H), a drilling diameter (d), a blocking length (L), a rock mass elastic modulus (E), an in-situ rock block size (x) and a blasting unit consumption (q) index as input variables (attribute values) of the model by corresponding blasting test occasions, and forming a training data set by the output variables and the input variables The step blasting parameter is schematically shown in fig. 2.
Step1.2: the data set obtained by using Step1.1 is subjected to resampling of training sample set by adopting Bootstrap method, m data samples are extracted, and n training data sets are randomly generatedD= { x i1,xi2,…,xin,yi } (i e1, m).
Based on a Bootstrap data set, a Bagging method is adopted to randomly select training data with a put back, then a classifier is constructed, and finally the overall acquisition effect is increased by combining the learned models.
Step1.3: using a decision tree algorithm in each subset, selecting an optimal mode to split nodes according to a 'minimum coefficient criterion', and setting the prediction result of a single decision tree predictor f (x, theta k) as f i (x) without pruning in the splitting process, wherein the final prediction result of a random forest regression model is expressed as that a parameter set RFP of a modeling process of the random forest regression algorithm is defined in a formula (1); equation (2) is used to make a prediction of blasting block based on random forest regression.
RFP＝{Ntree,Mtry} (1)
Where x represents the input vector, θ k is the vector representing the generation of each tree growth path, and F (x) represents the predicted blast block. Ntree is the number of decision trees in the model, and Ntree values affect the training degree and accuracy of the random forest model. In order to obtain the best evaluation result, codes are written by using MATLAB script language, and the relation between different decision tree trees (Ntree) and the model Mean Square Error (MSE) is simulated and calculated. Mtry is the number of features randomly extracted from the features. The Mtry value controls the disturbance degree of the random forest model attribute and is an important parameter in the model. The Mtry value directly affects the accuracy of the model. The Mtry values were calculated according to the following empirical formula:
Mtry&＝[log2M] (3)
Mtry&＝[M/3 (4)
wherein: m is the number of model input parameters; [] Representing a rounding down operation.
Further, the step 2) includes the steps of:
step2.1: a binary tree form decision tree is adopted, and a binary recursion is utilized to divide the data space into different subsets. In classification, assuming that there are K classes, the base index of the probability distribution is:
where p represents the probability of the sample point and p k represents the probability that the sample point belongs to the K-th class.
Step2.2: the decision tree uses GINI coefficients as criteria for attribute splitting. The feature with the lowest coefficient of the base is selected as the root node. And so on, selecting the leaf node with the smallest characteristic coefficient.
Step2.3: from the training dataset, starting from the root node, step 3 is recursively performed for each node, constructing a binary decision tree.
Further, the step 3) includes the steps of:
step3.1: the subset D is divided into two parts D 1 and D 2 according to whether the feature a takes a certain possible value a, D 1＝{(x,y)∈D|A(x)＝a},D2＝D-D1, and under the condition of the feature a, the base index of the set D is defined as:
where D 1 represents the training set of nodes, and the Basil index of the existing feature to the dataset is calculated.
Step3.2: for each feature a, for each possible value a, the base index at a=a is calculated by dividing D into two parts D 1 and D 2, depending on whether the test of the sample point pair a=a is yes or no. Among all possible features a and all possible cut points a, the cut point with the smallest keni index is selected as the cut point. Generating two child nodes from the current segmentation point according to the optimal characteristics and the optimal segmentation point, and distributing the training data set to two leaf nodes according to the characteristics.
Step3.3: step2.3 to Step3.2 are invoked recursively for both leaf nodes.
Step3.4: repeating the steps 3.1 to 3.3 to obtain different decision trees, recursively branching and growing each decision tree from top to bottom, stopping growing the regression tree after the segmentation termination condition is met, and finally combining all the regression trees together to form a random forest.
Step3.5: the input vector X in the prediction set X is input into a first layer random forest, and arithmetic average is carried out on the predicted value y i of the single regression tree by using a simple average method
Further, the step 4) includes the steps of:
step4.1: calculating the output of a layer 1 random forest on a training sample set Subtracting the output value/>, from the actual value y i of the training sampleTraining residual error/>, can be obtainedTraining residual/>Substituting the original training sample to construct a new data set as a training sample of a layer 2 random forest, so that the input is x i, and the expected output is/>
Step4.2: because the training residual value obtained by Step4.1 is too small, the training residual needs to be normalized:
Wherein: l in is the original input data; l min is the minimum of the same kind of data in the original input data; l max is the maximum value of the same kind of data in the original input data; l is the input data after normalization processing. And after normalization treatment, a training data set of the layer 2 random forest is formed.
Step4.3: and (3) establishing a layer 2 random forest model according to Step1.1 to Step4.1 to obtain a prediction result of the layer 2 random forest model.
Further, according to the statistical analysis result of the training sample set, the Range (Range), average value (Mean) and standard deviation (std. Dev) indexes of each modeling parameter are respectively given in table 1. As can be seen from fig. 2, the obvious correlation between the parameters is not shown in the figure, which indicates that the selected parameters have independence.
TABLE 1 random forest regression model training sample input and output parameter descriptive statistics
Case:
In order to further illustrate the accuracy and reliability of the method of the present invention, a step blasting experiment is taken as an example. The median block sizes of each field are respectively 0.200m,0.135m,0.105m,0.115m and 0.122m through a screening test after the field blasting test and the actual measurement of the blasting block size grading curve of each field; and calculating according to the explosion design parameters, the rock information parameters and the explosive information parameters adopted in the field explosion test to obtain the input parameters of the field explosion test model, wherein the input parameters are shown in table 2.
Table 2 in-situ burst test model input parameters
According to the established blasting block prediction system based on the RF-R method, 10 simulation prediction calculations are respectively carried out on the 5 groups of blasting tests of the experiment, and finally the prediction block average value of each test field is obtained, and the result is shown in Table 3.
TABLE 3 simulation experiment prediction blockiness results
Through calculation, the relative errors of the predicted block value and the actually measured block value of each blasting test field are respectively as follows: 0.60%, 1.04%, 12.76%, 2.63%, 2.80%, wherein the maximum error is 12.76%, the minimum error is 0.60%, and the average error is 3.96%.
The relative error between the predicted and measured block values for each burst test run is shown in fig. 5, for example, in fig. 4.
In conclusion, the explosion block prediction precision based on the RF-R model is higher than that of other prediction models, and the maximum relative error aiming at the experiment is not more than 15%, so that the method meets the engineering requirements. Therefore, by means of the quantitative relation between the blasting parameters and the blasting block sizes, the blasting parameters can be properly adjusted according to the blasting block sizes predicted by the RF-R model and the system and the actual conditions so as to meet the dam feeding block size requirements.
While the present invention has been described in detail with reference to the drawings, the present invention is not limited to the above embodiments, and various changes can be made without departing from the spirit of the present invention within the knowledge of those skilled in the art.
Claims (1)
1. A blasting block prediction method is characterized in that: the method comprises the following steps:
1) Forming a training data set;
2) Selecting a split node;
3) Generating a first layer of random forest;
4) Generating a second layer of random forest;
5) Sequentially superposing the output of the first layer random forest and the output of the second layer random forest to obtain the final output of the double-layer random forest prediction model;
the step 1) comprises the following steps:
Step1.1: taking the blasting average block value actually measured by a blasting test as an output variable of a model, acquiring a resistance line distance, a drilling row distance, a step height, a drilling diameter, a blocking length, a rock mass elastic modulus, an in-situ rock block and a blasting unit consumption index of a corresponding blasting test field as input variables of the model, and forming a training data set by the output variables and the input variables;
step1.2: the data set obtained by using Step1.1 is subjected to resampling of training sample set by adopting Bootstrap method, m data samples are extracted, and n training data sets are randomly generated D= { x b1,xb2,…,xbn,yb } (b e1, m ]);
based on a Bootstrap data set, a Bagging method is adopted to randomly select training data with a put back, then a classifier is constructed, and finally the overall acquisition effect is increased by combining the learned models;
Step1.3: and (3) using a decision tree algorithm in each subset, selecting an optimal mode to split nodes according to a minimum coefficient criterion, and if the prediction result of a single decision tree predictor f (x, theta k) is f i (x) in the splitting process, the final prediction result of a random forest regression model is expressed as: equation (1) defines a random forest regression algorithm modeling process parameter set RFP; the formula (2) is used for carrying out blasting block prediction based on random forest regression;
RFP＝{Ntree,Mtry} (1)
wherein x represents an input vector, θ k represents a vector for generating a growth path of each tree, F (x) represents predicted blasting block, ntree represents a decision tree number in a model, a MATLAB script language is used for writing codes and carrying out relation simulation calculation between different decision tree numbers Ntree and model mean square errors MSEs, mtry represents feature numbers randomly extracted from features, mtry values control disturbance degrees of random forest model attributes, and Mtry values are calculated according to the following empirical formula:
Mtry&＝[log2M] (3)
Mtry&＝[M/3] (4)
wherein: m is the number of model input parameters; [] Representing a downward rounding operation;
said step 2) comprises the steps of:
Step2.1: adopting a binary tree form decision tree, continuously dividing a data space into different subsets by utilizing binary recursion, and assuming K classes in classification, the radix index of probability distribution is:
wherein p represents the probability of the sample point, and p j represents the probability that the sample point belongs to the K-th class;
Step2.2: the decision tree adopts GINI coefficients as the attribute splitting standard, selects the feature with the lowest coefficient of the key as the root node, and so on, and selects the leaf node with the smallest coefficient of the rest features;
step2.3: recursively performing step 3) on each node from the root node according to the training data set, and constructing a binary decision tree;
said step 3) comprises the steps of:
Step3.1: the subset D is divided into two parts according to whether the feature A takes a certain value a;
Step3.2: for each feature A, taking a value a of each feature A, dividing D into two parts of D 1 and D 2 according to whether the test of the sample point pair A=a is yes or no, and calculating a base index when A=a; selecting a segmentation point with the minimum base index from all the features A and all the segmentation points a as a segmentation point, generating two child nodes from the current segmentation point according to the optimal features and the optimal segmentation point, and distributing a training data set into two leaf nodes according to the features;
step3.3: recursively invoking step2.3 to step3.2 for both leaf nodes;
Step3.4: repeating the steps 3.1 to 3.3 to obtain different decision trees, recursively branching and growing each decision tree from top to bottom, stopping growing the regression tree after the segmentation termination condition is met, and finally combining all the regression trees together to form a random forest;
step3.5: inputting an input vector X in a prediction set X into a first layer random forest, and calculating an arithmetic average value of predicted values of a single regression tree T i by using a simple average method;
The step 4) comprises the following steps:
Step4.1: calculating the output of a first layer of random forest on a training sample set Subtracting the output value/>, from the actual value y i of the training sampleTraining residual error/>, can be obtainedTraining residual/>Substituting the original training sample to construct a new data set as a training sample of a second layer random forest, so that the input is x i, and the expected output is/>
Step4.2: because the training residual value obtained by Step4.1 is too small, the training residual needs to be normalized:
Wherein: l in is the original input data; l min is the minimum of the same kind of data in the original input data; l max is the maximum value of the same kind of data in the original input data; l is input data after normalization processing, and a training data set of a second layer random forest is formed after normalization processing;
Step4.3: and establishing a second layer random forest model according to Step1.1 to Step4.1 to obtain a prediction result of the second layer random forest model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111054833.2A CN113762394B (en) | 2021-09-09 | 2021-09-09 | Blasting block prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111054833.2A CN113762394B (en) | 2021-09-09 | 2021-09-09 | Blasting block prediction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113762394A CN113762394A (en) | 2021-12-07 |
| CN113762394B true CN113762394B (en) | 2024-04-26 |
Family
ID=78794212
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111054833.2A Active CN113762394B (en) | 2021-09-09 | 2021-09-09 | Blasting block prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113762394B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115659263B (en) * | 2022-10-14 | 2023-08-08 | 长江三峡通航管理局 | Ship control behavior risk assessment system and method based on big data |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105574544A (en) * | 2015-12-16 | 2016-05-11 | 平安科技（深圳）有限公司 | Data processing method and device |
| CN109146195A (en) * | 2018-09-06 | 2019-01-04 | 北方爆破科技有限公司 | A kind of blast fragmentation size prediction technique based on cart tree regression algorithm |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10885259B2 (en) * | 2019-08-30 | 2021-01-05 | Intel Corporation | Random forest model for prediction of chip layout attributes |
-
2021
- 2021-09-09 CN CN202111054833.2A patent/CN113762394B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105574544A (en) * | 2015-12-16 | 2016-05-11 | 平安科技（深圳）有限公司 | Data processing method and device |
| CN109146195A (en) * | 2018-09-06 | 2019-01-04 | 北方爆破科技有限公司 | A kind of blast fragmentation size prediction technique based on cart tree regression algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113762394A (en) | 2021-12-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2020125668A1 (en) | Method and system for automatically identifying surrounding rock level by applying while-drilling parameters | |
| CN110442666B (en) | Mineral resource prediction method and system based on neural network model | |
| CN112529341B (en) | Drilling well leakage probability prediction method based on naive Bayesian algorithm | |
| Arbin et al. | Comparative analysis between k-means and k-medoids for statistical clustering | |
| CN110135167B (en) | Edge computing terminal security level evaluation method for random forest | |
| CN112989708B (en) | Well logging lithology identification method and system based on LSTM neural network | |
| CN109409647A (en) | A kind of analysis method of the salary level influence factor based on random forests algorithm | |
| CN103617147A (en) | Method for identifying mine water-inrush source | |
| CN108280289B (en) | Rock burst danger level prediction method based on local weighted C4.5 algorithm | |
| CN110344824B (en) | Acoustic curve generation method based on random forest regression | |
| CN113705110B (en) | Blasting vibration speed prediction method based on double random forest regression method | |
| CN113221960B (en) | Construction method and collection method of high-quality vulnerability data collection model | |
| CN113762394B (en) | Blasting block prediction method | |
| CN114662405A (en) | Rock burst prediction method based on few-sample measurement and ensemble learning | |
| CN104915679A (en) | Large-scale high-dimensional data classification method based on random forest weighted distance | |
| Chu et al. | Co-training based on semi-supervised ensemble classification approach for multi-label data stream | |
| CN105740521A (en) | Small grid removing method and apparatus in oil reservoir numerical simulation system solution process | |
| CN114091333A (en) | Shale gas content artificial intelligence prediction method based on machine learning | |
| CN102621953B (en) | Automatic online quality monitoring and prediction model updating method for rubber hardness | |
| CN117541095A (en) | Agricultural land soil environment quality classification method | |
| CN115809601A (en) | Sedimentary rock structure background distinguishing method | |
| CN118133104A (en) | Rapid identification method for lithofacies of deep sea-phase shale gas well | |
| CN111985782A (en) | Automatic tramcar driving risk assessment method based on environment perception | |
| CN114372640A (en) | Wind power prediction method based on fluctuation sequence classification correction | |
| CN111125082B (en) | Cement process parameter data analysis method based on association rule mining |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |