A coal mine safety risk early warning method based on incremental extreme learning machine

Abstract

The present invention belongs to the field of computer / coal mine safety, and specifically relates to a coal mine safety risk early warning method based on an incremental extreme learning machine. The method comprises: constructing a coal mine safety risk prediction model based on an incremental extreme learning machine; acquiring coal mine data in real time, preprocessing the data, and inputting the preprocessed data into a trained coal mine safety risk prediction model based on an incremental extreme learning machine to obtain the coal mine safety risk prediction result; and issuing an early warning to the coal mine according to the coal mine safety risk prediction result. The model prediction output of the present invention has a better fit with the actual accident occurrence situation, and the prediction output has better stability without large fluctuations, that is, the present invention is more reliable and has more advantages in predicting coal mine risks.

Description

Technical Field

[0001] The present invention belongs to the field of computer / coal mine safety, and in particular relates to a coal mine safety risk early warning method based on an incremental extreme learning machine. Background Art

[0002] As my country strengthens its supervision and inspection of coal mines, the comprehensive use of cloud computing, artificial intelligence, machine learning and other technologies for detection and early warning is becoming an increasingly important solution to ensure safe production in coal mines.

[0003] In the process of informatization of coal mine production, although a large amount of coal mine safety production data has been accumulated, there are few corresponding methods or models based on safety production, and there is no unified definition of accident safety hazards. Many rich coal mine production data cannot provide reliable decision-making information support for coal mine safety management. How to analyze and calculate these massive coal mine risk big data, discover the potential hidden dangers in coal mine production, and how to effectively model coal mine safety risks in a complex and changeable coal mine production environment are issues that need to be solved urgently. Summary of the invention

[0004] In order to solve the above existing problems, the present invention proposes a coal mine safety risk early warning method based on an incremental extreme learning machine, the method comprising: constructing a coal mine safety risk prediction model based on an incremental extreme learning machine; acquiring coal mine data in real time, preprocessing the data, inputting the preprocessed data into a trained coal mine safety risk prediction model based on an incremental extreme learning machine, and obtaining the coal mine safety risk prediction result; and issuing an early warning to the coal mine according to the coal mine safety risk prediction result;

[0005] The process of training the coal mine safety risk prediction model based on incremental extreme learning machine includes:

[0006] S1: Obtain historical coal mine data, pre-process the acquired data, and obtain regular hidden danger time series data; coal mine data includes the number of daily hidden dangers in coal mines, hidden danger types, hidden danger levels, historical accident records of mines, and basic mine data;

[0007] S2: Set the time granularity t, divide the historical coal mine data according to the set time granularity, and obtain training samples;

[0008] S3: Construct a nonlinear coal mine production safety risk model based on extreme learning machine and set the MSE threshold during model training;

[0009] S4: Initialize the model's hyperparameters and use genetic algorithm to select the optimal model hyperparameters;

[0010] S5: Put the data set into the coal mine production safety risk model, perform initial training on the model according to the selected hyperparameters, obtain the model output layer weights through the generalized inverse matrix, and select the time granularity t with the lowest MSE according to the initial training results;

[0011] S6: Incrementally train the model according to the time granularity t under the lowest MSE, and when the predicted MSE is higher than the MSE threshold set in step S3, incrementally update the model parameters;

[0012] S7: Update the number of hidden layers of the model to complete the model training process.

[0013] Preferably, the process of preprocessing the data includes: for the daily hidden danger inspection records of a mining area, extracting non-missing hidden danger data in the same time period according to the hidden danger type, hidden danger level, and inspection unit, calculating the mean and variance, and using the calculated mean and variance to fill in all missing values ​​to obtain a regular multi-granularity hidden danger monitoring time series.

[0014] The optimal hyperparameters of the coal mine safety risk prediction model include: the number of hidden layers L, the optimal time granularity t, the model weight W, the hidden layer bias and the model output layer weight β i .

[0015] Preferably, the process of selecting the optimal model hyperparameters using a genetic algorithm includes:

[0016] S41: Initialize the population number, population size and number of iterations; and randomly select the weight W and bias;

[0017] S42: Calculate the fitness of the population. The fitness function is expressed as:

[0018] F=min||MSE Trainingset ,MSE Testingset ||

[0019] Where MSE Trainingset is the mean square error of the training set, MSE Testingset is the mean square error of the test set;

[0020] S43: Perform selection, crossover and mutation operations on the population to generate new sub-populations;

[0021] S44: Calculate the fitness of the new sub-population. If the fitness function value is less than the previous fitness value, update the model weight W and the hidden layer bias.

[0022] S45: When the maximum number of iterations is reached, the optimal W and bias are obtained.

[0023] Preferably, the process of initially training the model according to the selected hyperparameters includes:

[0024] S51: Define the training loss function Function(H,T)=min||Hβ-T|| 2 ,in

[0025]

[0026]

[0027]

[0028] Among them, H is the output matrix, β is the output layer weight, T is the actual value, and g represents the activation function. Indicates Input weights, x N represents the Nth variable, Indicates A bias, Indicates output weights, represents the Nth label value, N represents the number of training sets, Represents the number of hidden layers; S52: The process of model training is to solve the output layer weight β when the loss function reaches the minimum value. The output weight of the model is calculated by the least squares method in matrix form, and its expression is:

[0029] β=(H T H) -1 H T T

[0030] S53: Predict the test set through T=Hβ, and select the optimal time granularity t according to the mean square error MSE.

[0031] Preferably, the process of incrementally training the model includes:

[0032] S61: When the predicted MSE is higher than the set MSE threshold, define the loss function after adding training data H0 is the last hidden layer output matrix, H1 is the output matrix of the newly added data; T0 represents the last label value, T1 represents the newly added label value, and β1 represents the new output weight that needs to be solved;

[0033] S62: According to the calculation method of generalized inverse matrix, we can get in: K1 represents the product of the output matrix and its transposed matrix;

[0034] S63: Convert the formula after adding samples into a function represented by known constants β0, K1, H1 and T1, and obtain the recursive formula of the incremental extreme learning machine training process; its expression is:

[0035]

[0036]

[0037] Among them, K k+1 represents the result of transposing and multiplying the hidden layer output matrix H during the k+1th incremental training; β k+1 represents the result of the new output layer weights during the k+1th incremental training; T k+1 Represents the training label during the k+1th incremental training;

[0038] S64: Predict the test set based on the result β of the updated output layer weight, and the prediction result is T=Hβ.

[0039] Preferably, the process of updating the hidden layer of the model includes:

[0040] S71: Update the model weight W and bias value. The weight update formula is: W l+1 =W l w, the bias update formula is b l+1 =b l b, where W l and b l is the weight and bias of l hidden layer nodes before the update, w is the newly added weight, b is the newly added bias, and w and b are arbitrary values;

[0041] S72: Calculate the output layer matrix according to the updated model weights and biases; the calculation expression is:

[0042] H l+1 = g([W l W l+1 ]X+[b l b l+1 ])=[H l h]

[0043] Among them, g represents the activation function, X represents the feature value, and W l represents the weight before update, W l+1 represents the newly added weight, b l represents the bias before updating, b l+1 Indicates the added bias, H l represents the hidden layer output matrix, h represents the newly added output matrix;

[0044] S73: The output layer weight after updating the hidden layer according to the output layer matrix is ​​expressed as:

[0045] β l+1 =(H l+1 ) + Y=([H l h] T [H l h]) -1 [H l h] T Y

[0046] Among them, (H l+1 ) + Represents the output matrix H l+1 The reversible matrix of l represents the original output matrix, and h represents the newly added output matrix.

[0047] Preferably, the process of processing data by the coal mine safety risk prediction model based on the incremental extreme learning machine includes:

[0048]

[0049]

[0050] Among them, L represents the number of hidden layers, g is the activation function, and W i represents the weight in the model, bias is the hidden layer bias, t represents the time window, x t represents the relevant characteristic value in the t time window, n represents the number of mines in the coal mine, β i Represents the output layer weight.

[0051] To achieve the above objectives, the present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, it implements any of the above-mentioned coal mine safety risk early warning methods based on an incremental extreme learning machine.

[0052] To achieve the above-mentioned purpose, the present invention also provides a coal mine safety risk warning device based on an incremental extreme learning machine, comprising a processor and a memory; the memory is used to store a computer program; the processor is connected to the memory, and is used to execute the computer program stored in the memory, so that the coal mine safety risk warning device based on an incremental extreme learning machine executes any of the above-mentioned coal mine safety risk warning methods based on an incremental extreme learning machine.

[0053] Beneficial effects of the present invention:

[0054] The present invention uses variable granularity to extract the spatiotemporal characteristics of coal mine hidden dangers, and outputs the prediction results after fusion of the characteristics, and finally trains the entire model for coal mine safety risk prediction. The model prediction output of the present invention has a better fit with the actual accident situation, and the prediction output has good stability without large fluctuations, that is, the present invention is more reliable and has more advantages in predicting coal mine risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flow chart of the method for predicting coal mine safety risks based on the incremental extreme learning machine of the present invention;

[0056] Figure 2 This is a rendering of the incremental extreme learning machine model implemented in the present invention. DETAILED DESCRIPTION

[0057] 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.

[0058] A specific implementation method of a coal mine safety risk early warning method based on an incremental extreme learning machine, the method comprising: constructing a coal mine safety risk prediction model based on an incremental extreme learning machine; acquiring coal mine data in real time, preprocessing the data, and inputting the preprocessed data into a trained coal mine safety risk prediction model based on an incremental extreme learning machine to obtain a coal mine safety risk prediction result; and issuing an early warning to the coal mine according to the coal mine safety risk prediction result.

[0059] The process of training the coal mine safety risk prediction model based on incremental extreme learning machine includes:

[0060] S1: Obtain historical coal mine data, pre-process the acquired data, and obtain regular hidden danger time series data; coal mine data includes the number of daily hidden dangers in coal mines, hidden danger types, hidden danger levels, historical accident records of mines, and basic mine data;

[0061] S2: Set the time granularity t, divide the historical coal mine data according to the set time granularity, and obtain training samples;

[0062] S3: Construct a nonlinear coal mine production safety risk model based on extreme learning machine and set the MSE threshold during model training;

[0063] S4: Initialize the model's hyperparameters and use genetic algorithm to select the optimal model hyperparameters;

[0064] S5: Put the data set into the coal mine production safety risk model, perform initial training on the model according to the selected hyperparameters, obtain the model output layer weights through the generalized inverse matrix, and select the time granularity t with the lowest MSE according to the initial training results;

[0065] S6: Incrementally train the model according to the time granularity t under the lowest MSE, and when the predicted MSE is higher than the MSE threshold set in step S3, incrementally update the model parameters;

[0066] S7: Update the number of hidden layers of the model to complete the model training process.

[0067] A specific implementation method of a coal mine safety risk early warning method based on an incremental extreme learning machine is provided. The method predicts the coal mine accident risk based on the improved incremental extreme learning machine model, and then provides decision support for coal mine hidden danger management and accident prevention, thereby achieving a breakthrough in key technologies for coal mine safety risk prevention. The method proposed in the present invention may include the following steps, such as Figure 1 As shown,

[0068] Step S1: for a certain mine, obtaining selected hidden danger monitoring data;

[0069] Step S2, selecting indicators according to the distribution rules of various hidden dangers;

[0070] Step S3, data preprocessing and missing data filling: For the daily hidden danger inspection records of a mining area, extract the non-missing hidden danger data in the same time period according to the hidden danger type, hidden danger level, and inspection unit, calculate the mean and variance, and use the calculated mean and variance to fill in all missing values ​​to obtain a regular multi-granularity hidden danger monitoring time series.

[0071] Step S4, sample set generation: select the specified time granularity T, the hidden danger monitoring data at the spatial granularity as the feature x, the coal mine risk value of the day as the prediction target Y, and generate training samples one by one in a single-point sequential moving manner to obtain the sample set X;

[0072] Step S5: For missing data, the mean of adjacent data is used to fill in. For hidden danger inspection data, it records the inspection results of coal mines by relevant inspectors every day. In order to facilitate model learning, it needs to be normalized to obtain key spatiotemporal features. Normalization can be expressed as:

[0073]

[0074] Among them, X new represents the normalized data, X i represents the i-th data, X minRepresents the minimum value of all data, X max Indicates the maximum value among all data.

[0075] Step S6: construct a coal mine safety risk model, which can be expressed as:

[0076]

[0077] Where L represents the number of hidden layers, g is the activation function, and the sigmoid function is selected, which can be expressed as:

[0078]

[0079] W i =[W1,W2,...W n ] T represents the weight in the model, bias is the hidden layer bias, t represents the time window, x t =(x 1t ,x 2t ,...,x nt ) T is the relevant characteristic value in the t time window, n represents the number of mines in the coal mine, this parameter is used to exclude the large difference in indicator data caused by different mine scales, β i is the output layer weight.

[0080] Step S7: Set the number of hidden layers L, weights W, and hidden layer bias of the initial model.

[0081] The process of using genetic algorithms to select the optimal model hyperparameters includes:

[0082] Step 1: Initialize the population number, population size, and number of iterations; and randomly select values ​​for the weight W and bias;

[0083] Step 2: Calculate the fitness of the population. The fitness function is expressed as:

[0084] F=min||MSE Trainingset ,MSE Testingset ||

[0085] Where MSE Trainingset is the mean square error of the training set, MSE Testingset is the mean square error of the test set;

[0086] Step 3: Perform selection, crossover and mutation operations on the population to generate new sub-populations;

[0087] Step 4: Calculate the fitness of the new sub-population. If the fitness function value is less than the previous fitness value, update the model weight W and the hidden layer bias.

[0088] Step 5: When the maximum number of iterations is reached, the optimal weights W and bias are obtained.

[0089] Step S8: Perform initial training on the model; initial training can be defined as:

[0090] Function(H,T)=min||Hβ-T|| 2

[0091]

[0092]

[0093]

[0094] Among them, H represents the output matrix of the hidden layer corresponding to the training data set, β is the weight from the hidden layer to the output layer, and the training process is equivalent to finding the optimal β to make the above formula valid.

[0095] The process of model training is to solve the output layer weight β when the loss function reaches the minimum value. The output weight of the model is calculated by the least squares method in matrix form, and its expression is:

[0096] β=(H T H) -1 H T T

[0097] Predict the data: T = Hβ. According to the training structure MSE, select the time granularity t with the lowest MSE.

[0098] Step S9, performing incremental training on the model;

[0099] When the predicted MSE is higher than the set MSE threshold, the loss function after adding training data is defined, and its expression is:

[0100]

[0101] H0 is the previous hidden layer output matrix, and H1 is the output matrix of the newly added data.

[0102] According to the calculation method of generalized inverse matrix, we can get: in:

[0103] Convert the formula after adding samples into a function of known constants β0, K1, H1 and T1, and get the recursive formula of the incremental extreme learning machine training process, that is, K1 can be expressed as: and We can get: Thus, the recursive formula of the incremental extreme learning machine training process can be obtained:

[0104]

[0105]

[0106] Among them, K k+1 represents the result of transposing and multiplying the hidden layer output matrix H during the k+1th incremental training; β k+1 represents the result of the new output layer weights during the k+1th incremental training; T k+1 Represents the training label during the k+1th incremental training; based on the updated output layer matrix β, the test set is predicted: the prediction result T = Hβ.

[0107] Step S10: Update the model hidden layer, and the updated weight W l+1 =W l w, bias b l+1 =b l b,W l and b l is the weight and bias of the number of hidden layer nodes before the update. The output weight of the newly added hidden layer node and the bias of the newly added hidden layer node are both random values. The specific steps are:

[0108] Step 1: Update the model's weight W and bias values. The weight update formula is: l+1 =W l w, the bias update formula is b l+1 =b l b, where W l and b l is the weight and bias of l hidden layer nodes before the update, w is the newly added weight, b is the newly added bias, and w and b are arbitrary values.

[0109] Step 2: Calculate the output layer matrix based on the updated model weights and biases; the calculation expression is:

[0110] H l+1 = g([W l W l+1 ]X+[b l b l+1 ])=[H l h]

[0111] Among them, g represents the activation function, X represents the eigenvalue, and W l represents the weight before update, W l+1 represents the newly added weight, b l represents the bias before updating, b l+1 Indicates the added bias, H lrepresents the hidden layer output matrix, and h represents the newly added output matrix.

[0112] Step 3: Update the output layer weights after the hidden layer according to the output layer matrix, and its expression is:

[0113] β l+1 =(H l+1 ) + Y=([H l h] T [H l h]) -1 [H l h] T Y

[0114] Among them, (H l+1 ) + Represents the output matrix H l+1 The reversible matrix of l represents the original output matrix, and h represents the newly added output matrix.

[0115] like Figure 2 The model prediction effect is shown in the following table:

[0116]

[0117] It can be concluded that this method has the best effect on coal mine risk fitting in a 21-day time window.

[0118] In one embodiment of the present invention, the present invention also includes a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, any of the above-mentioned coal mine safety risk warning methods based on an incremental extreme learning machine is implemented.

[0119] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

[0120] A coal mine safety risk warning device based on an incremental extreme learning machine comprises a processor and a memory; the memory is used to store a computer program; the processor is connected to the memory and is used to execute the computer program stored in the memory, so that the coal mine safety risk warning device based on an incremental extreme learning machine executes any of the above-mentioned coal mine safety risk warning methods based on an incremental extreme learning machine.

[0121] Specifically, the memory includes: ROM, RAM, magnetic disk, USB flash drive, memory card or optical disk, etc., which can store program codes.

[0122] Preferably, the processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0123] The above embodiments further illustrate the purpose, technical solutions and advantages of the present invention in detail. It should be understood that the above embodiments are only preferred implementation modes of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made to the present invention within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A coal mine safety risk early warning method based on incremental extreme learning machine, characterized in that: include: Construct a coal mine safety risk prediction model based on incremental extreme learning machine; Acquire coal mine data in real time, preprocess the data, and input the preprocessed data into a trained coal mine safety risk prediction model based on an incremental extreme learning machine to obtain the coal mine safety risk prediction result; Provide early warning to the coal mine based on the coal mine safety risk prediction results; The process of training the coal mine safety risk prediction model based on incremental extreme learning machine includes: S1: Obtain historical coal mine data, pre-process the acquired data, and obtain regular hidden danger time series data; coal mine data includes the number of daily hidden dangers in coal mines, hidden danger types, hidden danger levels, historical accident records of mines, and basic mine data; S2: Set the time granularity t, divide the historical coal mine data according to the set time granularity, and obtain training samples; S3: Construct a nonlinear coal mine production safety risk model based on extreme learning machine and set the MSE threshold during model training; S4: Initialize the model's hyperparameters and use genetic algorithm to select the optimal model hyperparameters; S41: Initialize the population number, population size and number of iterations; and randomly select the weight W and bias; S42: Calculate the fitness of the population. The fitness function is expressed as: F=min||MSE Trainingset ,MSE Testingset || Where MSE Trainingset is the mean square error of the training set, MSE Testingset is the mean square error of the test set; S43: Perform selection, crossover and mutation operations on the population to generate new sub-populations; S44: Calculate the fitness of the new sub-population. If the fitness function value is less than the previous fitness value, update the model weight W and the hidden layer bias. S45: When the maximum number of iterations is reached, the optimal W and bias are obtained; S5: Put the data set into the coal mine production safety risk model, perform initial training on the model according to the selected hyperparameters, obtain the model output layer weights through the generalized inverse matrix, and select the time granularity t with the lowest MSE according to the initial training results; S51: Define the training loss function Function(H,T)=min‖Hβ-T‖ 2 ,in H is the output matrix, β is the output layer weight, T is the actual value, g represents the activation function, Indicates Input weights, x N represents the Nth variable, Indicates A bias, Indicates output weights, represents the Nth label value, N represents the number of training sets, represents the number of hidden layers; S52: The process of model training is to solve the output layer weight β when the loss function reaches the minimum value. The output weight of the model is calculated by the least squares method in matrix form, and its expression is: β=(H T H) -1 H T T S53: predict the test set through T = Hβ, and select the optimal time granularity t according to the mean square error MSE; S6: Incrementally train the model according to the time granularity t under the lowest MSE, and when the predicted MSE is higher than the MSE threshold set in step S3, incrementally update the model parameters; S7: Update the number of hidden layers of the model to complete the model training process.

2. A coal mine safety risk early warning method based on incremental extreme learning machine according to claim 1, characterized in that: The process of data preprocessing includes: for a mining area's daily hidden danger inspection records, extracting non-missing hidden danger data in the same time period according to hidden danger type, hidden danger level, and inspection unit, calculating the mean and variance, and using the calculated mean and variance to fill in all missing values ​​to obtain a regular multi-granularity hidden danger monitoring time series.

3. The coal mine safety risk early warning method based on incremental extreme learning machine according to claim 1 is characterized in that: The hyperparameters of the coal mine safety risk prediction model include: the number of hidden layers L, the optimal time granularity t, the model weight W, the hidden layer bias and the model output layer weight β i .

4. The coal mine safety risk early warning method based on incremental extreme learning machine according to claim 1, characterized in that: The process of incrementally training a model involves: S61: When the predicted MSE is higher than the set MSE threshold, define the loss function after adding training data H0 is the last hidden layer output matrix, H1 is the output matrix of the newly added data; T0 represents the last label value, T1 represents the newly added label value, and β1 represents the new output weight that needs to be solved; S62: According to the calculation method of generalized inverse matrix, we can get in: K1 represents the product of the output matrix and its transposed matrix; S63: Convert the formula after adding samples into a function represented by known constants β0, K1, H1 and T1, and obtain the recursive formula of the incremental extreme learning machine training process; its expression is: Among them, K k+1 represents the result of transposing and multiplying the hidden layer output matrix H during the k+1th incremental training; β k+1 represents the result of the new output layer weights during the k+1th incremental training; T k+1 Represents the training label during the k+1th incremental training; S64: Predict the test set based on the result β of the updated output layer weight, and the prediction result is T=Hβ.

5. The coal mine safety risk early warning method based on incremental extreme learning machine according to claim 1, characterized in that: The process of updating the number of hidden layers of the model includes: S71: Update the model weight W and bias value. The weight update formula is: W l+1 =W l w, the bias update formula is b l+1 =b l b, where W l and b l is the weight and bias of l hidden layer nodes before the update, w is the newly added weight, b is the newly added bias, and w and b are arbitrary values; S72: Calculate the output layer matrix according to the updated model weights and biases; the calculation expression is: H l+1 =g([W l W l+1 ]X+[b l b l+1 ])=[H l h] Among them, g represents the activation function, X represents the feature value, and W l represents the weight before update, W l+1 represents the newly added weight, b l represents the bias before updating, b l+1 Indicates the added bias, H l represents the hidden layer output matrix, and h represents the newly added output matrix; S73: The output layer weight after updating the hidden layer according to the output layer matrix is ​​expressed as: β l+1 =(H l+1 ) + Y=([H l h] T [H l h]) -1 [H l h] T Y Among them, (H l+1 ) + Represents the output matrix H l+1 The reversible matrix of l represents the original output matrix, and h represents the newly added output matrix.

6. The coal mine safety risk early warning method based on incremental extreme learning machine according to claim 1, characterized in that: The process of processing data in the coal mine safety risk prediction model based on incremental extreme learning machine includes: Among them, L represents the number of hidden layers, g is the activation function, and W i represents the weight in the model, bias is the hidden layer bias, t represents the time window, x t represents the relevant characteristic value in the t time window, n represents the number of mines in the coal mine, β i Represents the output layer weights.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the coal mine safety risk early warning method based on an incremental extreme learning machine as claimed in any one of claims 1 to 6.

8. A coal mine safety risk early warning device based on an incremental extreme learning machine, characterized in that: It includes a processor and a memory; the memory is used to store computer programs; the processor is connected to the memory and is used to execute the computer programs stored in the memory, so that the coal mine safety risk warning device based on an incremental extreme learning machine executes any one of claims 1 to 6 of the coal mine safety risk warning method based on an incremental extreme learning machine.

Images