Training method of mine pressure data classification model and mine pressure data classification method

Abstract

The invention provides a training method of a mine pressure data classification model and a mine pressure data classification method. The method comprises the following steps: acquiring mine pressure data and mine pressure text data corresponding to the same mine pressure event; extracting mine pressure data features of the mine pressure data and mine pressure text features of the mine pressure text data, and splicing the mine pressure data features and the mine pressure text features to obtain sample feature data; associating the sample feature data with the sample mine pressure event to generate a training sample; according to the training sample, the initial mine pressure data classification model is trained, and the trained mine pressure data classification model is obtained, so that the training sample can be obtained without a large amount of labeled data, the initial mine pressure data classification model is trained based on the training sample, the efficiency in the training process of the mine pressure data classification model is improved, and the training efficiency of the mine pressure data classification model is improved. And the mine pressure data is classified based on the mine pressure data classification model to obtain the mine pressure event, so that the accuracy and reliability of obtaining the mine pressure event are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of mining, and in particular to a training method for a mine pressure data classification model and a mine pressure data classification method. Background Art

[0002] During mining, the occurrence of mine pressure incidents at the working face may pose a major threat to the personal safety of miners and the normal production of coal mines. Mine pressure monitoring is an important link in ensuring safe mining.

[0003] In related technologies, classification often relies on manual experience or specific sensor data threshold settings, resulting in limited classification accuracy. With the development of artificial intelligence technology, especially the application of deep learning methods, mine pressure classification has been able to be performed in an automated manner. However, most existing methods require a large amount of labeled data to train the model. Due to the diversity and complexity of mine pressure events, it is often difficult to collect enough sample data. Therefore, how to train a mine pressure data classification model when samples are scarce, and classify the mine pressure data based on the trained mine pressure data classification model to obtain mine pressure events has become an urgent problem to be solved. Summary of the invention

[0004] The present application aims to solve one of the technical problems in the related art at least to some extent.

[0005] According to the first aspect of the present application, a training method for a mine pressure data classification model is provided, including obtaining mine pressure data and mine pressure text data corresponding to the same mine pressure event; extracting mine pressure data features of the mine pressure data and mine pressure text features of the mine pressure text data, splicing the mine pressure data features and the mine pressure text features to obtain sample feature data; associating the sample feature data with sample mine pressure events to generate training samples; and training an initial mine pressure data classification model based on the training samples to obtain a trained mine pressure data classification model.

[0006] According to the second aspect of the present application, a mine pressure data classification method is provided, including: obtaining target mine pressure data and target mine pressure text data to be classified; extracting target mine pressure data features of the target mine pressure data and target mine pressure text features of the target mine pressure text data, splicing the target mine pressure data features and the target mine pressure text features to obtain target feature data; inputting the target feature data into a mine pressure data classification model to obtain a target mine pressure event, wherein the mine pressure data classification model is a model trained using the training method described in the first aspect.

[0007] According to a third aspect of the present application, a training device for a mine pressure data classification model is provided, comprising: a first acquisition module, used to acquire mine pressure data and mine pressure text data corresponding to the same mine pressure event; a second acquisition module, used to extract mine pressure data features of the mine pressure data and mine pressure text features of the mine pressure text data, and splice the mine pressure data features and the mine pressure text features to obtain sample feature data; a third acquisition module, used to associate the sample feature data with sample mine pressure events to generate training samples; and a training module, used to train an initial mine pressure data classification model according to the training samples to obtain a trained mine pressure data classification model.

[0008] According to the fourth aspect of the present application, a mine pressure data classification device is provided, comprising: an acquisition module, used to acquire target mine pressure data and target mine pressure text data to be classified; a feature extraction module, used to extract target mine pressure data features of the target mine pressure data and target mine pressure text features of the target mine pressure text data, and splice the target mine pressure data features and the target mine pressure text features to obtain target feature data; a classification module, used to input the target feature data into a mine pressure data classification model to obtain a target mine pressure event, wherein the mine pressure data classification model is a model trained using the training method described in the first aspect.

[0009] In order to achieve the above-mentioned purpose, the fifth aspect of the present application proposes an electronic device, characterized in that it includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the training method of the mine pressure data classification model described in the first aspect or the mine pressure data classification method described in the second aspect.

[0010] In order to achieve the above-mentioned purpose, the sixth aspect of the present application proposes a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the training method of the mine pressure data classification model described in the first aspect or the mine pressure data classification method described in the second aspect.

[0011] In order to achieve the above-mentioned purpose, the seventh aspect of the present application proposes a computer program product, including a computer program, which, when executed by a processor, implements the training method of the mine pressure data classification model described in the first aspect or the mine pressure data classification method described in the second aspect.

[0012] The technical solution provided by the embodiments of the present application includes at least the following beneficial effects: The present application provides a training method for a mine pressure data classification model, which can obtain training samples without a large amount of labeled data, and trains an initial mine pressure data classification model based on the training samples, thereby improving the efficiency of the mine pressure data classification model training process, and classifies the mine pressure data based on the mine pressure data classification model to obtain mine pressure events, thereby improving the accuracy and reliability of obtaining mine pressure events.

[0013] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present application. Figure 1 A schematic flow chart of a method for training a mine pressure data classification model provided in an embodiment of the present application; Figure 2 A schematic diagram of a flow chart of a mine pressure data classification method provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of a training device for a mine pressure data classification model provided in an embodiment of the present application; Figure 4 A schematic diagram of the structure of a mine pressure data classification device provided in an embodiment of the present application; Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0015] The following is a description of exemplary embodiments of the present application in conjunction with the accompanying drawings, including various details of the embodiments of the present application to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0016] The following uses an embodiment to explain in detail the training method of the mine pressure data classification model of the present application.

[0017] Figure 1 A flowchart of a method for training a mine pressure data classification model provided in an embodiment of the present application.

[0018] like Figure 1 As shown, the training method of the mine pressure data classification model proposed in this embodiment specifically includes the following steps: S101. Obtaining mine pressure data and mine pressure text data corresponding to the same mine pressure event.

[0019] It should be noted that mine pressure events include but are not limited to initial pressure, periodic pressure and large-scale pressure.

[0020] For example, the mine pressure data and mine pressure text data corresponding to the first pressure, the mine pressure data and mine pressure text data corresponding to periodic pressure, and the mine pressure data and mine pressure text data corresponding to large-area pressure, etc. can be obtained.

[0021] The mine pressure data refers to the mine pressure data during mining, including but not limited to pressure sensor data and vibration data, for example: working resistance data of working face supports.

[0022] In an embodiment of the present application, original mine pressure data corresponding to the same mine pressure event can be obtained, the time span of the original mine pressure data and a preset time span threshold can be obtained, and the original mine pressure data with a time span greater than or equal to the time span threshold can be used as the first mine pressure data. The first mine pressure data is preprocessed to obtain the mine pressure data. Optionally, the tensile strength of the rock formation, the thickness of the key layer and the load of the key layer can be obtained, and the number of coal mining machine feeds and the number of daily cycles of the coal mining machine can be obtained. Based on the tensile strength of the rock formation, the thickness of the key layer, the load of the key layer, the number of coal mining machine feeds and the number of daily cycles of the coal mining machine, the time span of the original mine pressure data can be obtained.

[0023] For example, the time span of the original mine pressure data can be obtained using the following formula:

[0024] in, is the time span of the original mine pressure data, in days, is the tensile strength of the rock formation, is the key layer thickness, is the key layer load, is the number of times the coal mining machine cuts, It is the number of daily cycles of the coal mining machine.

[0025] It should be noted that the specific method of preprocessing the first mine pressure data to obtain the mine pressure data is not limited in the present application and can be selected according to actual conditions.

[0026] Optionally, in order to improve the quality and availability of the data, duplicate items in the support data can be removed. For missing values ​​in the first mine pressure data, if there are many missing values ​​or the missing values ​​are continuous, they will be deleted. If there are a small number of missing values ​​and they are discontinuous, the missing values ​​will be filled by interpolation to ensure the consistency of the time series intervals. For outliers, the deletion and filling method is used to process them. Data from multiple sources are integrated, the data format is unified, and redundant fields and conflicting data are removed to facilitate subsequent feature extraction and model training.

[0027] For example, at the time point There are missing values , using adjacent data points and Data and Perform linear interpolation: .

[0028] S102, extracting the mine pressure data features of the mine pressure data and the mine pressure text features of the mine pressure text data, and concatenating the mine pressure data features and the mine pressure text features to obtain sample feature data.

[0029] In an embodiment of the present application, a pre-trained long short-term memory auto encoder (LSTM-AE) model can be obtained, and the mine pressure data is input into the LSTM-AE model. The encoder in the LSTM-AE model extracts features from the mine pressure data to obtain mine pressure data features.

[0030] In an embodiment of the present application, a pre-constructed mine pressure-specific vocabulary library can be obtained, and the mine pressure text data can be segmented according to the mine pressure-specific vocabulary library to obtain standard mine pressure text data. The standard mine pressure text data is input into a fine-tuned Bert model, and the Bert model is used to extract features from the standard mine pressure text data to obtain mine pressure text features.

[0031] Optionally, sample standard mine pressure text data may be obtained to fine-tune the Bert model to obtain a fine-tuned Bert model.

[0032] In an embodiment of the present application, it is possible to obtain mine pressure text data from multiple sources, preprocess the mine pressure text data from multiple sources to obtain target mine pressure text data, perform keyword extraction on the target mine pressure text data, obtain target keywords in the target mine pressure text data, and construct a mine pressure-specific vocabulary based on the target keywords.

[0033] It should be noted that the multi-source mine pressure text data includes but is not limited to mine pressure log text data, expert knowledge text data, mine pressure related papers and patent data.

[0034] Optionally, preprocessing operations such as word segmentation, part-of-speech tagging and stop word removal can be performed on multi-source mine pressure text data to obtain target mine pressure text data, and keywords can be extracted from the target mine pressure text data. Based on the extracted target keywords, a mine pressure-specific vocabulary library is constructed to ensure that the mine pressure data classification model can effectively identify terms related to mine pressure events.

[0035] It should be noted that the present application does not limit the specific method of extracting keywords from the target mine pressure text data, and the method can be selected according to actual conditions.

[0036] Optionally, the keyword extraction algorithm Text Rank can be used to extract keywords from the target mine pressure text data, construct a word graph, and calculate the iterative update weight:

[0037] in, For keywords The weight value, d is the learning rate, which can be set to 0.85, For all pointing keywords A collection of keywords, Indicates from keywords Point to keywords The weight of the edge, Indicates from keywords The specified node collection.

[0038] It should be noted that after obtaining the weight value of the keyword, the weight value of the keyword can be sorted in descending order, and the keywords corresponding to the first K weight values ​​​​in the sorting results are selected as target keywords, and the target keywords are verified to obtain the final target keywords. According to the final target keywords, a special vocabulary library for mine pressure is constructed.

[0039] In an embodiment of the present application, after obtaining the mine pressure data features and the mine pressure text features, the mine pressure data features and the mine pressure text features can be spliced ​​to obtain sample feature data, that is, the sample feature data includes feature data of text mode and time series mode.

[0040] S103: Associating sample feature data with sample mine pressure events to generate training samples.

[0041] The sample mine pressure event is a label corresponding to the sample characteristic data. The sample characteristic data corresponds to the sample mine pressure event one by one, that is, each sample characteristic data has a corresponding sample mine pressure event.

[0042] In the embodiment of the present application, after the sample characteristic data is obtained, the sample mine pressure event corresponding to the sample data can be obtained, and the sample characteristic data and the sample mine pressure event can be associated to generate a training sample.

[0043] S104. According to the training samples, the initial mine pressure data classification model is trained to obtain a trained mine pressure data classification model.

[0044] Optionally, the initial mine pressure data classification model may be a support vector machine (Support Vector Machines) learning model.

[0045] In an embodiment of the present application, sample feature data is input into an initial mine pressure data classification model to output predicted mine pressure events, obtain loss functions of predicted mine pressure events and sample mine pressure events, adjust model parameters of the initial mine pressure data classification model according to the loss function, and continue training until the termination conditions are met to obtain a trained mine pressure data classification model.

[0046] Optionally, a cross entropy loss function can be obtained based on predicted mine pressure events and sample mine pressure events, the gradient can be calculated based on the cross entropy loss function, the model parameters can be updated by the back propagation algorithm, the model parameters of the initial mine pressure data classification model can be adjusted, and the performance of the mine pressure data classification model can be tested using a cross-validation evaluation method. The model parameters can be adjusted based on the test results, and training can be continued until the termination conditions are met to obtain a trained mine pressure data classification model.

[0047] It should be noted that the present disclosure does not limit the setting of the training end condition, which can be set according to actual conditions.

[0048] Optionally, the training end condition may be set to that the loss value is less than a loss threshold; optionally, the number of adjustments of the model parameters of the initial mine pressure data classification model may be set to reach a preset number threshold.

[0049] The training method of the mine pressure data classification model provided in the present application obtains the mine pressure data and the mine pressure text data corresponding to the same mine pressure event, extracts the mine pressure data features of the mine pressure data and the mine pressure text features of the mine pressure text data, splices the mine pressure data features and the mine pressure text features to obtain sample feature data, associates the sample feature data with the sample mine pressure event, generates training samples, and trains the initial mine pressure data classification model according to the training samples to obtain a trained mine pressure data classification model. Therefore, the present application can obtain the training samples without a large amount of labeled data, and trains the initial mine pressure data classification model based on the training samples, thereby improving the efficiency of the mine pressure data classification model training process, and classifies the mine pressure data based on the mine pressure data classification model to obtain mine pressure events, and can accurately classify the unseen mine pressure events without a large amount of labeled data, can effectively solve the problems existing in the traditional mine pressure classification method, improve the accuracy and reliability of obtaining mine pressure events, and can be quickly deployed in a new mining environment without additional training time, thereby improving the efficiency of obtaining mine pressure events. It should be noted that after obtaining the trained mine pressure data classification model, the mine pressure data can be classified based on the mine pressure data classification model to obtain mine pressure events.

[0050] The following uses an example to explain in detail the mine pressure data classification method of the present application.

[0051] Figure 2 A schematic diagram of the flow chart of the mine pressure data classification method provided in an embodiment of the present application.

[0052] like Figure 2 As shown, the mine pressure data classification method proposed in this embodiment specifically includes the following steps: S201. Obtain target mine pressure data and target mine pressure text data to be classified.

[0053] In the embodiment of the present application, target mine pressure data and target mine pressure text data to be classified can be obtained during the real-time mining process.

[0054] S202, extracting target mine pressure data features of the target mine pressure data and target mine pressure text features of the target mine pressure text data, and concatenating the target mine pressure data features and the target mine pressure text features to obtain target feature data.

[0055] In an embodiment of the present application, after acquiring the target mine pressure data and the target mine pressure text data, the target mine pressure data is input into the LSTM-AE model, and the encoder in the LSTM-AE model performs feature extraction on the target mine pressure data to obtain the target mine pressure data features, and the target mine pressure text data is segmented according to the mine pressure-specific vocabulary to obtain the target standard mine pressure text data, and the target standard mine pressure text data is input into the fine-tuned Bert model, and the Bert model performs feature extraction on the target standard mine pressure text data to obtain the target mine pressure text features.

[0056] In the embodiment of the present application, after the target mine pressure data features and the target mine pressure text features are acquired, the target mine pressure data features and the target mine pressure text features are spliced ​​to obtain the target feature data.

[0057] S203, inputting the target characteristic data into the mine pressure data classification model to obtain the target mine pressure event.

[0058] Among them, the mine pressure data classification model is a model trained using the above-mentioned model training method.

[0059] In the embodiment of the present application, the target characteristic data may be input into a mine pressure data classification model to output a target mine pressure event corresponding to the target characteristic data.

[0060] Among them, target mine pressure events include but are not limited to initial pressure, periodic pressure and large-area pressure.

[0061] The mine pressure data classification method provided in the present application obtains target mine pressure data and target mine pressure text data to be classified, extracts target mine pressure data features of the target mine pressure data and target mine pressure text features of the target mine pressure text data, splices the target mine pressure data features and the target mine pressure text features to obtain target feature data, inputs the target feature data into a mine pressure data classification model, and obtains target mine pressure events. Therefore, the present application classifies the target feature data through the mine pressure data classification model to obtain target mine pressure events, thereby improving the accuracy and reliability of obtaining target mine pressure events, and can better ensure the personal safety of miners and the normal production of coal mines.

[0062] In order to implement the above embodiment, this embodiment provides a training device for a mine pressure data classification model. Figure 3 A schematic diagram of the structure of a training device for a mine pressure data classification model provided in an embodiment of the present application.

[0063] like Figure 3 As shown, the training device 1000 for the mine pressure data classification model includes: a first acquisition module 110 , a second acquisition module 120 , a third acquisition module 130 and a training module 140 .

[0064] A first acquisition module 110 is used to acquire the mine pressure data and mine pressure text data corresponding to the same mine pressure event; A second acquisition module 120 is used to extract the mine pressure data features of the mine pressure data and the mine pressure text features of the mine pressure text data, and to combine the mine pressure data features and the mine pressure text features to obtain sample feature data; A third acquisition module 130 is used to associate the sample feature data with the sample mine pressure event to generate a training sample; The training module 140 is used to train the initial mine pressure data classification model according to the training samples to obtain a trained mine pressure data classification model.

[0065] According to one embodiment of the present application, the first acquisition module 110 is also used to: obtain original mine pressure data corresponding to the same mine pressure event; obtain the time span of the original mine pressure data and a preset time span threshold, and use the original mine pressure data with a time span greater than or equal to the time span threshold as the first mine pressure data; and preprocess the first mine pressure data to obtain the mine pressure data. In one embodiment of the present application, the first acquisition module 110 is also used to: obtain the tensile strength of the rock formation, the thickness of the key layer and the load of the key layer, and obtain the number of coal mining machine feeds and the number of daily cycles of the coal mining machine; obtain the time span of the original mine pressure data based on the tensile strength of the rock formation, the thickness of the key layer, the load of the key layer, the number of coal mining machine feeds and the number of daily cycles of the coal mining machine.

[0066] In one embodiment of the present application, the second acquisition module 120 is also used to: obtain a pre-trained long short-term memory autoencoder LSTM-AE model; input the mine pressure data into the LSTM-AE model, and the encoder in the LSTM-AE model performs feature extraction on the mine pressure data to obtain the mine pressure data features.

[0067] In one embodiment of the present application, the second acquisition module 120 is also used to: obtain a pre-constructed mine pressure-specific vocabulary library; perform word segmentation processing on the mine pressure text data according to the mine pressure-specific vocabulary library to obtain standard mine pressure text data; input the standard mine pressure text data into a fine-tuned Bert model, and use the Bert model to perform feature extraction on the standard mine pressure text data to obtain the mine pressure text features.

[0068] In one embodiment of the present application, the process of constructing the mine pressure-specific vocabulary library includes: acquiring mine pressure text data from multiple sources; preprocessing the mine pressure text data from multiple sources to obtain target mine pressure text data; performing keyword extraction on the target mine pressure text data to obtain target keywords in the target mine pressure text data; and constructing the mine pressure-specific vocabulary library based on the target keywords.

[0069] In one embodiment of the present application, the training module 140 is also used to: input the sample feature data into the initial mine pressure data classification model to output a predicted mine pressure event; obtain the loss function of the predicted mine pressure event and the sample mine pressure event, adjust the model parameters of the initial mine pressure data classification model according to the loss function, and continue training until the end condition is met to obtain the trained mine pressure data classification model.

[0070] The training device for the mine pressure data classification model provided in the present application obtains mine pressure data and mine pressure text data corresponding to the same mine pressure event, extracts the mine pressure data features of the mine pressure data and the mine pressure text features of the mine pressure text data, splices the mine pressure data features and the mine pressure text features to obtain sample feature data, associates the sample feature data with the sample mine pressure event, generates training samples, and trains the initial mine pressure data classification model according to the training samples to obtain a trained mine pressure data classification model. Therefore, the present application can obtain training samples without a large amount of labeled data, and trains the initial mine pressure data classification model based on the training samples, thereby improving the efficiency of the mine pressure data classification model training process, and classifies the mine pressure data based on the mine pressure data classification model to obtain mine pressure events, thereby improving the accuracy and reliability of obtaining mine pressure events, and can be quickly deployed in new mining environments without the need for additional training time, thereby improving the efficiency of obtaining mine pressure events.

[0071] In order to implement the above embodiment, this embodiment provides a mine pressure data classification device. Figure 4 A schematic diagram of the structure of a mine pressure data classification device provided in an embodiment of the present application.

[0072] like Figure 4 As shown, the mine pressure data classification device 2000 includes: an acquisition module 210 , a feature extraction module 220 and a classification module 230 .

[0073] An acquisition module 210 is used to acquire target mine pressure data and target mine pressure text data to be classified; A feature extraction module 220 is used to extract target mine pressure data features of the target mine pressure data and target mine pressure text features of the target mine pressure text data, and to concatenate the target mine pressure data features and the target mine pressure text features to obtain target feature data; The classification module 230 is used to input the target feature data into a mine pressure data classification model to obtain a target mine pressure event, wherein the mine pressure data classification model is a model trained using the training method described in the first aspect.

[0074] The mine pressure data classification device provided by the present application obtains target mine pressure data and target mine pressure text data to be classified, extracts target mine pressure data features of the target mine pressure data and target mine pressure text features of the target mine pressure text data, splices the target mine pressure data features and the target mine pressure text features to obtain target feature data, inputs the target feature data into a mine pressure data classification model, and obtains target mine pressure events. Therefore, the present application classifies the target feature data through the mine pressure data classification model to obtain target mine pressure events, thereby improving the accuracy and reliability of obtaining target mine pressure events, and can better ensure the personal safety of miners and the normal production of coal mines.

[0075] In order to implement the above embodiment, the present application also proposes an electronic device 3000, such as Figure 5 As shown, it includes: a memory 310, a processor 320, and a computer program stored in the memory 310 and executable on the processor 320. When the processor executes the program, the training method of the mine pressure data classification model described in the first aspect or the mine pressure data classification method described in the second aspect is implemented.

[0076] In order to implement the above embodiments, the present application proposes a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the training method of the mine pressure data classification model described in the first aspect or the mine pressure data classification method described in the second aspect.

[0077] In order to implement the above embodiments, the present application also proposes a computer program product, including a computer program, which, when executed by a processor, implements the training method of the mine pressure data classification model described in the first aspect or the mine pressure data classification method described in the second aspect.

[0078] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this application can be executed in parallel, sequentially or in different orders, as long as the expected results of the technical solution disclosed in this application can be achieved, and this document is not limited here.

[0079] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A training method for a mine pressure data classification model, characterized in that: The method comprises: Obtaining the mine pressure data and mine pressure text data corresponding to the same mine pressure event; Extracting the mine pressure data features of the mine pressure data and the mine pressure text features of the mine pressure text data, and splicing the mine pressure data features and the mine pressure text features to obtain sample feature data; Associating the sample characteristic data with sample mine pressure events to generate training samples; The initial mine pressure data classification model is trained according to the training samples to obtain a trained mine pressure data classification model.

2. The method according to claim 1, characterized in that The obtaining of the mine pressure data corresponding to the same mine pressure event includes: Obtaining original mine pressure data corresponding to the same mine pressure event; Acquire the time span of the original mine pressure data and a preset time span threshold, and use the original mine pressure data whose time span is greater than or equal to the time span threshold as the first mine pressure data; The first mine pressure data is preprocessed to obtain the mine pressure data.

3. The method according to claim 2, characterized in that The time span for obtaining the original mine pressure data includes: Obtain the tensile strength of rock formations, thickness of key layers and load of key layers, as well as the number of shearer feeds and daily cycles; The time span of the original mine pressure data is obtained according to the tensile strength of the rock formation, the thickness of the key layer, the load of the key layer, the number of cutting times of the coal mining machine and the number of daily cycles of the coal mining machine.

4. The method according to claim 1, characterized in that: The extracting of the mine pressure data features of the mine pressure data comprises: Get the pre-trained long short-term memory autoencoder LSTM-AE model; The mine pressure data is input into the LSTM-AE model, and the encoder in the LSTM-AE model performs feature extraction on the mine pressure data to obtain the mine pressure data features.

5. The method according to claim 1, characterized in that: The step of extracting the mine pressure text features of the mine pressure text data comprises: Get pre-built mining pressure-specific vocabulary; According to the mine pressure special word library, the mine pressure text data is segmented to obtain standard mine pressure text data; The standard mine pressure text data is input into the fine-tuned Bert model, and the Bert model is used to extract features of the standard mine pressure text data to obtain the mine pressure text features.

6. The method according to claim 5, characterized in that The construction process of the mine pressure special word library includes: Obtaining mine pressure text data from multiple sources; Preprocessing the multi-source mine pressure text data to obtain target mine pressure text data; Performing keyword extraction on the target mine pressure text data to obtain target keywords in the target mine pressure text data; The mine pressure-specific vocabulary is constructed according to the target keywords.

7. The method according to claim 1, characterized in that The initial mine pressure data classification model is trained according to the training samples to obtain a trained mine pressure data classification model, including: Inputting the sample characteristic data into the initial mine pressure data classification model to output a predicted mine pressure event; The loss functions of the predicted mine pressure event and the sample mine pressure event are obtained, model parameters of the initial mine pressure data classification model are adjusted according to the loss function, and training is continued until the end condition is met to obtain the trained mine pressure data classification model.

8. A mine pressure data classification method, characterized in that: The method comprises: Obtaining target mine pressure data to be classified and target mine pressure text data; Extracting target mine pressure data features of the target mine pressure data and target mine pressure text features of the target mine pressure text data, and splicing the target mine pressure data features and the target mine pressure text features to obtain target feature data; The target characteristic data is input into a mine pressure data classification model to obtain a target mine pressure event, wherein the mine pressure data classification model is a model trained by using the training method described in any one of claims 1 to 7.

9. A training device for a mine pressure data classification model, characterized in that: The device comprises: A first acquisition module is used to acquire the mine pressure data and mine pressure text data corresponding to the same mine pressure event; A second acquisition module is used to extract the mine pressure data features of the mine pressure data and the mine pressure text features of the mine pressure text data, and to splice the mine pressure data features and the mine pressure text features to obtain sample feature data; A third acquisition module is used to associate the sample feature data with the sample mine pressure event to generate a training sample; The training module is used to train the initial mine pressure data classification model according to the training samples to obtain a trained mine pressure data classification model.

10. A mine pressure data classification device, characterized in that: The device comprises: An acquisition module, used to acquire target mine pressure data and target mine pressure text data to be classified; A feature extraction module is used to extract target mine pressure data features of the target mine pressure data and target mine pressure text features of the target mine pressure text data, and to concatenate the target mine pressure data features and the target mine pressure text features to obtain target feature data; A classification module is used to input the target feature data into a mine pressure data classification model to obtain a target mine pressure event, wherein the mine pressure data classification model is a model trained using the training method described in any one of claims 1-7.

Images