Method and system for processing coal mine disaster data based on principal component analysis method

Abstract

The invention provides a method and system for processing coal mine disaster data based on a PCA (principal component analysis) method. The method comprises the following steps: preprocessing and standardizing working data to obtain effective data, performing dimension reduction analysis on the effective data by using PCA to convert the effective data into a plurality of feature vectors, and constructing a dimension reduction space of the effective data, and respectively calculating a reconstruction error and an interpretation variance ratio corresponding to each feature vector in the dimension reduction space, constructing dimension information corresponding to each feature vector to adjust the dimension reduction space, inputting effective data into the adjusted dimension reduction space for processing, generating dimension reduction data of the coal mine, and storing the dimension reduction data of the coal mine. And abnormal dimension reduction data segments which do not conform to the specified interpretation variance ratio are screened to obtain corresponding disaster data of the coal mine under different detection dimensions, so that the calculation steps can be effectively reduced, the analysis cost is reduced, whole-course visual analysis is realized, managers can conveniently understand the data, and the data processing efficiency is improved. And a powerful reference is provided for the safety of coal miners and the environment.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for processing coal mine disaster data based on principal component analysis. Background Art

[0002] Coal mine disasters are an important factor affecting the safety and efficiency of coal mine production. With the widespread application of sensor technology and data acquisition systems in modern coal mine production, coal mining companies can obtain a large amount of production monitoring data in real time, including geological conditions, equipment status, and environmental monitoring data. By observing and analyzing these data, potential dangers can be discovered in a timely manner and preventive measures can be taken. However, due to the large amount of data and its complexity, how to effectively use and analyze these data has become a huge challenge.

[0003] Traditional data analysis methods often have some limitations and shortcomings when processing coal mine disaster data, such as information loss, weak pattern recognition ability, overfitting, high computational complexity, and parameter sensitivity leading to suboptimal solutions. These problems restrict its effective analysis and application of coal mine disaster data. The specific examples are as follows: 1. Information loss and weak pattern recognition capabilities: Simple statistical methods, such as mean and standard deviation, cannot capture the multivariate relationships of the data, may ignore important information, and have difficulty extracting meaningful patterns and trends from complex data.

[0004] 2. Multicollinearity and dimensionality issues: Traditional data analysis methods often face problems such as instability, overfitting, high computational complexity, and poor effectiveness when dealing with multivariate data. For example, linear regression may become unstable due to multicollinearity; decision trees and random forests have poor interpretability in high-dimensional data and are prone to overfitting when data noise is high; support vector machines (SVMs) are very sensitive to parameter settings, and improper parameter selection can significantly affect model performance and result accuracy; neural networks are prone to overfitting when dealing with small sample data.

[0005] 3. Assumption limitation: Traditional data analysis methods mostly assume that there is a linear relationship between data before analyzing the data, and cannot handle nonlinear relationships. For example, linear regression methods cannot handle nonlinear relationships; although support vector machines (SVM) can use kernel functions to handle nonlinear data, it is extremely difficult to select appropriate kernel functions and parameters; time series analysis (ARIMA) is difficult to handle nonlinear data and requires data to be stationary, so it cannot be applied to all data types.

[0006] 4. High computational complexity: Coal mine disaster data involves complex interactions between multiple factors. Traditional data analysis methods often require a lot of computing resources and time for training when analyzing and calculating them, resulting in high computational complexity and high computational overhead.

[0007] 5. Poor interpretability: Traditional data analysis methods have poor interpretability for high-dimensional data, and it is difficult to understand the clustering features and the relationship between the features.

[0008] Therefore, the present invention provides a method and system for processing coal mine disaster data based on principal component analysis. Summary of the invention

[0009] The present invention provides a method and system for processing coal mine disaster data based on principal component analysis, which can effectively overcome many limitations of traditional data analysis methods. On the one hand, PCA can reduce the complexity of data by reducing dimensionality, thereby improving the efficiency and accuracy of analysis. On the other hand, PCA can extract the main features in the data, remove redundant information and noise, and make pattern recognition more accurate. In addition, PCA can reduce the computational complexity and avoid the problem of overfitting by projecting high-dimensional data into low-dimensional space. The characteristic that PCA does not rely on specific parameter settings also makes the analysis results more stable and reliable.

[0010] The present invention provides a method for processing coal mine disaster data based on principal component analysis, comprising: Step 1: Collect the working data of the coal mine, pre-process and standardize the working data to obtain the valid data of the coal mine; Step 2: Perform dimensionality reduction analysis on the valid data using PCA, convert the valid data into a number of feature vectors, and use the feature vectors to construct a dimensionality reduction space of the valid data; Step 3: Calculate the reconstruction error and explained variance ratio corresponding to each of the eigenvectors in the dimensionality reduction space respectively, and construct the dimensional information corresponding to each of the eigenvectors; Step 4: adjusting the dimension reduction space based on the dimension information, and inputting the valid data into the adjusted dimension reduction space for processing to generate dimension reduction data of the coal mine; Step 5: Screen out abnormal dimensionality reduction data segments that do not meet the specified explained variance ratio, optimize the abnormal dimensionality reduction data segments, and obtain the disaster data corresponding to the coal mine under different detection dimensions.

[0011] In one practicable manner, The step 1 comprises: Step 11: collecting real-time gas concentration information, real-time temperature information, real-time humidity information, real-time mineralization information, real-time water tank flow information, real-time grouting information and real-time nitrogen injection information of the coal mine, and generating working data of the coal mine by time sorting; Step 12: obtaining the mean of the working data, analyzing the standard deviation distance between each data point in the working data and the mean, screening the standard deviation distance, determining a number of abnormal data points contained in the working data, and performing cleaning preprocessing on the abnormal data points; Step 13: construct a box plot of the working data, identify noise data points in the working data in the box plot, perform noise reduction preprocessing on the noise data points, and regard the working data that has completed the cleaning preprocessing and noise reduction preprocessing as the cleaning data of the coal mine; Step 14: Standardize the cleaned data to obtain valid data with a mean of 0 and a variance of 1.

[0012] In one practicable manner, Also includes: Find the first outlier data point that exceeds the upper quartile + 1.5 times the interquartile range and the second outlier data point that exceeds the lower quartile - 1.5 times the interquartile range in the box plot; The first outlier data point and the second outlier data segment are regarded as noise data points in the working data.

[0013] In one practicable manner, The step 2 comprises: Step 21: draw the original space of the valid data, search for the first several original coordinate axes with mutually orthogonal features in the original space, search for the target coordinate axis with the largest variance with the valid data as the first coordinate axis, locate the first coordinate axis in the original space, and determine the search starting point; Step 22: starting from the search starting point, sequentially search for a second coordinate axis that is orthogonal to the first coordinate axis and has the largest variance with the first coordinate plane, and a third coordinate axis that is orthogonal to the first coordinate axis and the second coordinate axis plane and has the largest variance with the first-second coordinate plane, and obtain N coordinate axes after the search is completed; Step 23: establishing a variance change list of the original space according to the variances between adjacent coordinate axes, roughly screening the variance change list to obtain K valid coordinate axes whose variances are not 0, and constructing a corresponding feature vector according to each of the valid coordinate axes; Step 24: Establish a covariance matrix of the eigenvector, analyze the eigenvalues ​​corresponding to each eigenvector in the covariance matrix using eigendecomposition method, build a spatial framework based on the eigenvectors, and input the eigenvalues ​​into the spatial framework to generate a dimensionality reduction space for the valid data.

[0014] In one practicable manner, The step 3 comprises: Step 31: respectively obtain the explained variance ratio of each of the eigenvectors in the dimensionality reduction space, draw a visualization ratio diagram of the effective data, identify the importance corresponding to each of the eigenvectors in the visualization ratio diagram, and construct the first dimensional sub-information corresponding to the eigenvector; Step 32: construct corresponding feature data according to the feature vector, map the feature data to the original space for loss assessment, obtain data loss information corresponding to the feature vector, and construct the reconstruction error of the valid data; Step 33: construct the reconstruction error to determine the spatial structure error and spatial pattern error of the dimensionality reduction space, and generate the second dimensional sub-information corresponding to the feature vector; Step 34: Establish dimensional information corresponding to the feature vector in the dimensionality reduction space according to the first dimensional sub-information and the second dimensional sub-information.

[0015] In one practicable manner, The step 4 comprises: Step 41: adjusting the spatial structure and spatial pattern of the dimensionality reduction space based on the dimensionality information to obtain a stable dimensionality reduction space; Step 42: input the valid data into the stable dimensionality reduction space, perform dimensionality reduction training on the valid data in the stable dimensionality reduction space, and generate dimensionality reduction data of the coal mine.

[0016] In one practicable manner, The step 5 comprises: Step 51: Obtain the standard working range corresponding to each detection dimension in the coal mine, construct the prescribed explained variance ratio range of the corresponding detection dimension according to the standard working range, estimate the dimensional information corresponding to each detection dimension and construct a data partitioning rule; Step 52: dividing the dimension-reduced data into a plurality of dimension-reduced data segments based on the data partitioning rule, and identifying abnormal dimension data segments that do not meet the requirements using the corresponding specified explained variance rate range; Step 53: Mark the abnormal dimension data segment in the valid data, and track the source of each abnormal dimension data segment respectively, construct and display the disaster data corresponding to the coal mine under different detection dimensions.

[0017] In one practicable manner, The process of establishing the covariance matrix of the eigenvector includes: Count all eigenvectors and construct the eigenvector set X; According to formula (1), the mean value corresponding to each of the eigenvectors is calculated respectively; in, represents the mean of the first j vectors, n represents the total amount of the eigenvectors, Represents the i-th coordinate axis feature of the j-th eigenvector in the corresponding coordinate axis; According to formula (2), each of the feature vectors is decentralized to generate the target feature vector set ; Use formula (3) to establish the target feature vector set The covariance matrix of Among them, T represents the transpose symbol and C represents the covariance matrix symbol.

[0018] The present invention provides a system for processing coal mine disaster data based on principal component analysis method, comprising: A data processing module is used to collect the working data of the coal mine, pre-process and standardize the working data, and obtain the effective data of the coal mine; A dimensionality reduction analysis module, used to perform dimensionality reduction analysis on the valid data using PCA, convert the valid data into a plurality of eigenvectors, and construct a dimensionality reduction space of the valid data using the eigenvectors; An information construction module, used to respectively calculate the reconstruction error and explained variance ratio corresponding to each of the eigenvectors in the dimensionality reduction space, and construct the dimensional information corresponding to each of the eigenvectors; A dimension reduction execution module, used for adjusting the dimension reduction space based on the dimension information, and inputting the valid data into the adjusted dimension reduction space for processing to generate dimension reduction data of the coal mine; The data processing module is used to screen out abnormal dimensionality reduction data segments that do not meet the specified explained variance ratio, optimize the abnormal dimensionality reduction data segments, and obtain the disaster data corresponding to the coal mine under different detection dimensions.

[0019] In one practicable manner, The dimension reduction analysis module comprises: A space construction unit is used to draw the original space of the valid data, search for the first several original coordinate axes with mutually orthogonal characteristics in the original space, search for the target coordinate axis with the largest variance with the valid data as the first coordinate axis, locate the first coordinate axis in the original space, and determine the search starting point; A coordinate decomposition unit is used to sequentially search, starting from the search starting point, for a second coordinate axis that has an orthogonal relationship with the first coordinate axis and has the largest variance with the first coordinate plane, and a third coordinate axis that has an orthogonal relationship with the first coordinate axis and the second coordinate axis plane and has the largest variance with the first-second coordinate plane, and obtain N coordinate axes after completing the search; A coordinate screening unit, used to establish a variance change list of the original space according to the variance between adjacent coordinate axes, perform rough screening on the variance change list to obtain K valid coordinate axes whose variance is not 0, and construct a corresponding feature vector according to each of the valid coordinate axes; A space construction unit is used to establish a covariance matrix of the eigenvector, analyze the eigenvalues ​​corresponding to each eigenvector in the covariance matrix using the eigendecomposition method, build a space framework based on the eigenvectors, and input the eigenvalues ​​into the space framework to generate a dimensionality reduction space of the valid data.

[0020] The achievable beneficial effects of the above technical solution are as follows: since a large amount of data is generated during the operation of a coal mine, it is necessary to quickly screen the data when analyzing its disasters, so the working data is first preprocessed and standardized to make its mean 0 and variance 1 to eliminate the influence between different feature dimensions, and obtain valid data. Then, the principal component analysis method is used to reduce the dimension of the valid data, and several feature vectors are determined. At the same time, a reduced dimension space of the valid data is constructed, and the adjacent dimensional information of the features is further determined based on the reconstruction error and the explained variance ratio of the feature vector in the reduced dimension space, and the reduced dimension space is adjusted to realize the reduced dimension processing of the valid data, and the reduced dimension data of the coal mine is obtained. At this time, the reduced dimension data can be used in the reduced dimension data. By identifying abnormal data segments, the disaster data of the coal mine is determined. In this way, the collected raw data can be processed and converted so that the PCA method can effectively extract key features and reduce the dimension, screen out important features in the working data, reduce the impact of redundant data and redundant features on disaster analysis, and perform error adjustment during dimensionality reduction to avoid data loss. This method can analyze the working data of various types of coal mines and is a comprehensive analysis method. Since this method first screens the features and then analyzes them, it can effectively reduce the calculation steps and reduce the analysis cost. The whole process is visualized, which makes it easier for managers to understand the ins and outs of the data and provides a strong reference for the safety of coal miners and the environment.

[0021] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0022] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 A schematic diagram of a workflow of a method for processing coal mine disaster data based on principal component analysis in an embodiment of the present invention; Figure 2 The figure is a schematic diagram of the composition of a system for processing coal mine disaster data based on principal component analysis in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0025] Example 1 This embodiment provides a method and system for processing coal mine disaster data based on principal component analysis. Figure 1 As shown, including: Step 1: Collect the working data of the coal mine, pre-process and standardize the working data to obtain the valid data of the coal mine; Step 2: Perform dimensionality reduction analysis on the valid data using PCA, convert the valid data into a number of feature vectors, and use the feature vectors to construct a dimensionality reduction space of the valid data; Step 3: Calculate the reconstruction error and explained variance ratio corresponding to each of the eigenvectors in the dimensionality reduction space respectively, and construct the dimensional information corresponding to each of the eigenvectors; Step 4: adjusting the dimension reduction space based on the dimension information, and inputting the valid data into the adjusted dimension reduction space for processing to generate dimension reduction data of the coal mine; Step 5: Screen out abnormal dimensionality reduction data segments that do not meet the specified explained variance ratio, optimize the abnormal dimensionality reduction data segments, and obtain the disaster data corresponding to the coal mine under different detection dimensions.

[0026] In this instance, preprocessing refers to the process of cleaning the working data; In this example, the main idea of ​​PCA is to map n-dimensional features to k-dimensional features. This k-dimensional feature is a new orthogonal feature also called principal component, which is a k-dimensional feature reconstructed on the basis of the original n-dimensional feature. In this example, the feature vector represents the data vector presented by the valid data in different dimensions; In this example, the dimension reduction space refers to the space where the dimension-reduced data is located after the effective data is dimensionally reduced; In this example, the purpose of dimensionality reduction is: PCA projects high-dimensional data into a lower-dimensional space, extracts the main factors of multivariate things through spatial projection and linear transformation, and reveals their essential characteristics. PCA deletes redundant and repeated variables (closely related variables) for all the variables originally proposed, and establishes as few new variables as possible so that these new variables are uncorrelated with each other. These new variables are called principal components, which can well describe the sample information. The core of the PCA algorithm is to reduce the dimension and screen the main features. The purpose of principal component analysis is to simplify the model training process and improve the computational efficiency of the model by retaining the most important feature information and reducing the dimension of the data. Through PCA dimensionality reduction, the redundant information between features can be reduced, and more representative features can be extracted, which helps to improve the generalization ability and prediction accuracy of the model; In this example, the reconstruction error represents the error generated during the dimensionality reduction process; In this example, the explained variance ratio represents an indicator used to measure the proportion of each feature vector in the dimensionality reduction space; In this example, the dimension information represents the information contained in the feature vector under its corresponding dimension; In this example, the detection dimensions include: gas concentration dimension, temperature dimension, humidity dimension, mineralization dimension, water tank flow dimension, grouting dimension and nitrogen injection dimension.

[0027] The working principle and beneficial effects of the above technical solution: Since a large amount of data is generated during the operation of a coal mine, it is necessary to quickly screen the data when analyzing its disasters. Therefore, the working data is first preprocessed and standardized to make its mean 0 and variance 1 to eliminate the influence between different feature dimensions and obtain valid data. Then, the principal component analysis method is used to reduce the dimension of the valid data, determine several feature vectors, and construct a reduced dimension space for the valid data. Further, the adjacent dimensional information of the feature is determined based on the reconstruction error and explained variance ratio of the feature vector in the reduced dimension space, and the reduced dimension space is adjusted to achieve the reduced dimension processing of the valid data, and the reduced dimension data of the coal mine is obtained. At this time, the reduced dimension data can be used in the reduced dimension data. By identifying abnormal data segments, the disaster data of the coal mine is determined. In this way, the collected raw data can be processed and converted so that the PCA method can effectively extract key features and reduce the dimension, screen out important features in the working data, reduce the impact of redundant data and redundant features on disaster analysis, and perform error adjustment during dimensionality reduction to avoid data loss. This method can analyze the working data of various types of coal mines and is a comprehensive analysis method. Since this method first screens the features and then analyzes them, it can effectively reduce the calculation steps and reduce the analysis cost. The whole process is visualized, which makes it easier for managers to understand the ins and outs of the data and provides a strong reference for the safety of coal miners and the environment.

[0028] Example 2 On the basis of Example 1, the method for processing coal mine disaster data based on principal component analysis method, step 1, comprises: Step 11: collecting real-time gas concentration information, real-time temperature information, real-time humidity information, real-time mineralization information, real-time water tank flow information, real-time grouting information and real-time nitrogen injection information of the coal mine, and generating working data of the coal mine by time sorting; Step 12: obtaining the mean of the working data, analyzing the standard deviation distance between each data point in the working data and the mean, screening the standard deviation distance, determining a number of abnormal data points contained in the working data, and performing cleaning preprocessing on the abnormal data points; Step 13: construct a box plot of the working data, identify noise data points in the working data in the box plot, perform noise reduction preprocessing on the noise data points, and regard the working data that has completed the cleaning preprocessing and noise reduction preprocessing as the cleaning data of the coal mine; Step 14: Standardize the cleaned data to obtain valid data with a mean of 0 and a variance of 1.

[0029] In this example, the data collection module is a key link. It is responsible for collecting data on monitoring indicators and environmental factors related to coal mine disasters. Through reasonable data collection strategies and quality control measures, it ensures that the collected data has sufficient relevance and high quality, providing reliable input for subsequent data analysis. In this example, the standard distance difference indicates the relative position between the data value of the data point and the mean value, and is an indicator used to describe the degree of dispersion between the data point and the mean point; In this example, the box plot represents a tool used to display the distribution, symmetry, outliers, and median of the working data.

[0030] The working principle and beneficial effects of the above technical solution: In order to improve the efficiency of data analysis and avoid being affected by redundant data, the collected information is first sorted to construct the working data of the coal mine, and then the collected original data is cleaned according to different situations to remove noise and anomalies, and then standard processing is performed to eliminate the influence between different characteristic dimensions, so that the effective data obtained can not only represent the information of the coal mine, but also ensure the quality of subsequent work.

[0031] Example 3 Based on Example 2, the method for processing coal mine disaster data based on principal component analysis method further includes: Find the first outlier data point that exceeds the upper quartile + 1.5 times the interquartile range and the second outlier data point that exceeds the lower quartile - 1.5 times the interquartile range in the box plot; The first outlier data point and the second outlier data segment are regarded as noise data points in the working data.

[0032] The working principle and beneficial effects of the above technical solution are as follows: locating noise data points in valid data to facilitate noise elimination.

[0033] Example 4 On the basis of Example 1, the method for processing coal mine disaster data based on principal component analysis, step 2, comprises: Step 21: draw the original space of the valid data, search for the first several original coordinate axes with mutually orthogonal features in the original space, search for the target coordinate axis with the largest variance with the valid data as the first coordinate axis, locate the first coordinate axis in the original space, and determine the search starting point; Step 22: starting from the search starting point, sequentially search for a second coordinate axis that is orthogonal to the first coordinate axis and has the largest variance with the first coordinate plane, and a third coordinate axis that is orthogonal to the first coordinate axis and the second coordinate axis plane and has the largest variance with the first-second coordinate plane, and obtain N coordinate axes after the search is completed; Step 23: establishing a variance change list of the original space according to the variances between adjacent coordinate axes, roughly screening the variance change list to obtain K valid coordinate axes whose variances are not 0, and constructing a corresponding feature vector according to each of the valid coordinate axes; Step 24: Establish a covariance matrix of the eigenvector, analyze the eigenvalues ​​corresponding to each eigenvector in the covariance matrix using eigendecomposition method, build a spatial framework based on the eigenvectors, and input the eigenvalues ​​into the spatial framework to generate a dimensionality reduction space for the valid data.

[0034] In this example, the work of PCA is to sequentially find a set of mutually orthogonal coordinate axes from the original space. The selection of new coordinate axes is closely related to the data itself. Among them, the first new coordinate axis is selected in the direction with the largest variance in the original data, the second new coordinate axis is selected in the plane orthogonal to the first coordinate axis so that the variance is the largest, and the third axis is the plane orthogonal to the 1st and 2nd axes with the largest variance. And so on, n such coordinate axes can be obtained. The new coordinate axes obtained in this way; In this example, the reason for coarse screening is that most of the variance is contained in the first k coordinate axes, and the variance contained in the following coordinate axes is almost 0, so the remaining coordinate axes can be ignored, and only the first k coordinate axes containing most of the variance are retained. In fact, this is equivalent to retaining only the dimensional features containing most of the variance, and ignoring the feature dimensions containing almost 0 variance, thereby achieving dimensionality reduction processing of data features; In this instance, the spatial frame represents the appearance of the reduced-dimensional space; In this example, the mutually orthogonal feature indicates that the inner product of the normal vectors of the plane formed by the two original coordinate axes is 0.

[0035] The working principle and beneficial effects of the above technical solution are as follows: In order to achieve data dimensionality reduction, the original space of valid data is first drawn, and then the original coordinate axes with mutually orthogonal features in the original space are searched according to the rules, and then the meaningful eigenvectors are determined by rough screening. PCA not only simplifies the data structure of coal mine disasters, but also reduces the computational overhead, while retaining the amount of information of the original data as much as possible, improving the accuracy and interpretability of the analysis results, and further using the matrix method to express the distribution of eigenvectors, thereby building a spatial framework of eigenvectors and generating a dimensionality reduction space, so that the valid data can be reduced in dimension in this space, thereby improving the accuracy of the results.

[0036] Example 5 On the basis of Example 1, the method for processing coal mine disaster data based on principal component analysis, step 3, comprises: Step 31: respectively obtain the explained variance ratio of each of the eigenvectors in the dimensionality reduction space, draw a visualization ratio diagram of the effective data, identify the importance corresponding to each of the eigenvectors in the visualization ratio diagram, and construct the first dimensional sub-information corresponding to the eigenvector; Step 32: construct corresponding feature data according to the feature vector, map the feature data to the original space for loss assessment, obtain data loss information corresponding to the feature vector, and construct the reconstruction error of the valid data; Step 33: construct the reconstruction error to determine the spatial structure error and spatial pattern error of the dimensionality reduction space, and generate the second dimensional sub-information corresponding to the feature vector; Step 34: Establish dimensional information corresponding to the feature vector in the dimensionality reduction space according to the first dimensional sub-information and the second dimensional sub-information.

[0037] In this example, the visualization ratio plot shows the result of visualizing the explained variance ratio of the eigenvectors; In this example, the first dimension sub-information includes the importance and important content of the feature vector in the dimensionality reduction space; In this example, the second dimensional sub-information includes the structural influence and pattern influence of the feature vector on the dimensionality reduction space.

[0038] The working principle and beneficial effects of the above technical solution are as follows: the importance of corresponding feature neighbors and data loss are analyzed by using the explained variance ratio and reconstruction error of adjacent features in the dimensionality reduction space, and the dimensionality information of the feature vector is established by using the structure and pattern error of the dimensionality reduction space. This can evaluate whether the selected feature vector can retain sufficient original data variation information and check whether the data in the space after dimensionality reduction maintains the main structure and pattern of the original data to achieve the purpose of deep training.

[0039] Example 6 On the basis of Example 1, the method for processing coal mine disaster data based on principal component analysis, step 4, comprises: Step 41: adjusting the spatial structure and spatial pattern of the dimensionality reduction space based on the dimensionality information to obtain a stable dimensionality reduction space; Step 42: input the valid data into the stable dimensionality reduction space, perform dimensionality reduction training on the valid data in the stable dimensionality reduction space, and generate dimensionality reduction data of the coal mine.

[0040] The working principle and beneficial effects of the above technical solution are as follows: the spatial structure and pattern of the dimensionality reduction space are adjusted according to the reconstruction error, and then valid data is input into it for formal dimensionality reduction training. Since various errors in the stable dimensionality reduction space have been eliminated in the early stage, complete and accurate dimensionality reduction data can be obtained at this time.

[0041] Example 7 On the basis of Example 1, the method for processing coal mine disaster data based on principal component analysis, step 5, comprises: Step 51: Obtain the standard working range corresponding to each detection dimension in the coal mine, construct the prescribed explained variance ratio range of the corresponding detection dimension according to the standard working range, estimate the dimensional information corresponding to each detection dimension and construct a data partitioning rule; Step 52: dividing the dimension-reduced data into a plurality of dimension-reduced data segments based on the data partitioning rule, and identifying abnormal dimension data segments that do not meet the requirements using the corresponding specified explained variance rate range; Step 53: Mark the abnormal dimension data segment in the valid data, and track the source of each abnormal dimension data segment respectively, construct and display the disaster data corresponding to the coal mine under different detection dimensions.

[0042] In this example, the standard working range indicates the range of normal data under a detection dimension, for example: the normal working stability of a coal mine is not higher than 39°; In this example, the data partitioning rule represents the process of partitioning the dimension-reduced data into corresponding detection dimensions.

[0043] The working principle and beneficial effects of the above technical solution are as follows: by dividing the dimensionality reduction data into corresponding dimensionality reduction data segments, then identifying the abnormal dimensionality data segments therein, and tracing the abnormal data segments to construct the disaster data of the coal mine, the tracing and positioning of the disaster data are achieved, providing powerful data information for management personnel.

[0044] Example 8 On the basis of Example 4, the method for processing coal mine disaster data based on principal component analysis method, the process of establishing the covariance matrix of the eigenvector includes: Count all eigenvectors and construct the eigenvector set X; According to formula (1), the mean value corresponding to each of the eigenvectors is calculated respectively; in, represents the mean of the first j vectors, n represents the total amount of the eigenvectors, Represents the i-th coordinate axis feature of the j-th eigenvector in the corresponding coordinate axis; According to formula (2), each of the feature vectors is decentralized to generate the target feature vector set ; Among them, decentralization can be completed by subtracting the mean value of each eigenvector from its vector value; Use formula (3) to establish the target feature vector set The covariance matrix of Among them, T represents the transpose symbol and C represents the covariance matrix symbol.

[0045] The working principle and beneficial effects of the above technical solution are as follows: by analyzing the mean of the eigenvectors, then using the mean to perform decentralization, and finally constructing a matrix.

[0046] Example 9 This embodiment provides a system for processing coal mine disaster data based on principal component analysis. Figure 2 As shown, including: A data processing module is used to collect the working data of the coal mine, pre-process and standardize the working data, and obtain the effective data of the coal mine; A dimensionality reduction analysis module, used to perform dimensionality reduction analysis on the valid data using PCA, convert the valid data into a plurality of eigenvectors, and construct a dimensionality reduction space of the valid data using the eigenvectors; An information construction module, used to respectively calculate the reconstruction error and explained variance ratio corresponding to each of the eigenvectors in the dimensionality reduction space, and construct the dimensional information corresponding to each of the eigenvectors; A dimension reduction execution module, used for adjusting the dimension reduction space based on the dimension information, and inputting the valid data into the adjusted dimension reduction space for processing to generate dimension reduction data of the coal mine; The data processing module is used to screen out abnormal dimensionality reduction data segments that do not meet the specified explained variance ratio, optimize the abnormal dimensionality reduction data segments, and obtain the disaster data corresponding to the coal mine under different detection dimensions.

[0047] In this instance, preprocessing refers to the process of cleaning the working data; In this example, the main idea of ​​PCA is to map n-dimensional features to k-dimensional features. This k-dimensional feature is a new orthogonal feature also called principal component, which is a k-dimensional feature reconstructed on the basis of the original n-dimensional feature. In this example, the feature vector represents the data vector presented by the valid data in different dimensions; In this example, the dimension reduction space refers to the space where the dimension-reduced data is located after the effective data is dimensionally reduced; In this example, the purpose of dimensionality reduction is: PCA projects high-dimensional data into a lower-dimensional space, extracts the main factors of multivariate things through spatial projection and linear transformation, and reveals their essential characteristics. PCA deletes redundant and repeated variables (closely related variables) for all the variables originally proposed, and establishes as few new variables as possible so that these new variables are uncorrelated with each other. These new variables are called principal components, which can well describe the sample information. The core of the PCA algorithm is to reduce the dimension and screen the main features. The purpose of principal component analysis is to simplify the model training process and improve the computational efficiency of the model by retaining the most important feature information and reducing the dimension of the data. Through PCA dimensionality reduction, the redundant information between features can be reduced, and more representative features can be extracted, which helps to improve the generalization ability and prediction accuracy of the model; In this example, the reconstruction error represents the error generated during the dimensionality reduction process; In this example, the explained variance ratio represents an indicator used to measure the proportion of each feature vector in the dimensionality reduction space; In this example, the dimension information represents the information contained in the feature vector under its corresponding dimension; In this example, the detection dimensions include: gas concentration dimension, temperature dimension, humidity dimension, mineralization dimension, water tank flow dimension, grouting dimension and nitrogen injection dimension.

[0048] The working principle and beneficial effects of the above technical solution: Since a large amount of data is generated during the operation of a coal mine, it is necessary to quickly screen the data when analyzing its disasters. Therefore, the working data is first preprocessed and standardized to make its mean 0 and variance 1 to eliminate the influence between different feature dimensions and obtain valid data. Then, the principal component analysis method is used to reduce the dimension of the valid data, determine several feature vectors, and construct a reduced dimension space for the valid data. Further, the adjacent dimensional information of the feature is determined based on the reconstruction error and explained variance ratio of the feature vector in the reduced dimension space, and the reduced dimension space is adjusted to achieve the reduced dimension processing of the valid data, and the reduced dimension data of the coal mine is obtained. At this time, the reduced dimension data can be used in the reduced dimension data. By identifying abnormal data segments, the disaster data of the coal mine is determined. In this way, the collected raw data can be processed and converted so that the PCA method can effectively extract key features and reduce the dimension, screen out important features in the working data, reduce the impact of redundant data and redundant features on disaster analysis, and perform error adjustment during dimensionality reduction to avoid data loss. This method can analyze the working data of various types of coal mines and is a comprehensive analysis method. Since this method first screens the features and then analyzes them, it can effectively reduce the calculation steps and reduce the analysis cost. The whole process is visualized, which makes it easier for managers to understand the ins and outs of the data and provides a strong reference for the safety of coal miners and the environment.

[0049] Example 10

[0050] Based on Example 9, the system for processing coal mine disaster data based on principal component analysis method includes: A space construction unit is used to draw the original space of the valid data, search for the first several original coordinate axes with mutually orthogonal characteristics in the original space, search for the target coordinate axis with the largest variance with the valid data as the first coordinate axis, locate the first coordinate axis in the original space, and determine the search starting point; A coordinate decomposition unit is used to sequentially search, starting from the search starting point, for a second coordinate axis that has an orthogonal relationship with the first coordinate axis and has the largest variance with the first coordinate plane, and a third coordinate axis that has an orthogonal relationship with the first coordinate axis and the second coordinate axis plane and has the largest variance with the first-second coordinate plane, and obtain N coordinate axes after completing the search; A coordinate screening unit, used to establish a variance change list of the original space according to the variance between adjacent coordinate axes, perform rough screening on the variance change list to obtain K valid coordinate axes whose variance is not 0, and construct a corresponding feature vector according to each of the valid coordinate axes; A space construction unit is used to establish a covariance matrix of the eigenvector, analyze the eigenvalues ​​corresponding to each eigenvector in the covariance matrix using the eigendecomposition method, build a space framework based on the eigenvectors, and input the eigenvalues ​​into the space framework to generate a dimensionality reduction space of the valid data.

[0051] In this example, the work of PCA is to sequentially find a set of mutually orthogonal coordinate axes from the original space. The selection of new coordinate axes is closely related to the data itself. Among them, the first new coordinate axis is selected in the direction with the largest variance in the original data, the second new coordinate axis is selected in the plane orthogonal to the first coordinate axis so that the variance is the largest, and the third axis is the plane orthogonal to the 1st and 2nd axes with the largest variance. And so on, n such coordinate axes can be obtained. The new coordinate axes obtained in this way; In this example, the reason for coarse screening is that most of the variance is contained in the first k coordinate axes, and the variance contained in the following coordinate axes is almost 0, so the remaining coordinate axes can be ignored, and only the first k coordinate axes containing most of the variance are retained. In fact, this is equivalent to retaining only the dimensional features containing most of the variance, and ignoring the feature dimensions containing almost 0 variance, thereby achieving dimensionality reduction processing of data features; In this instance, the spatial frame represents the appearance of the reduced-dimensional space; In this example, the mutually orthogonal feature indicates that the inner product of the normal vectors of the plane formed by the two original coordinate axes is 0.

[0052] The working principle and beneficial effects of the above technical solution are as follows: In order to achieve data dimensionality reduction, the original space of valid data is first drawn, and then the original coordinate axes with mutually orthogonal features in the original space are searched according to the rules, and then the meaningful eigenvectors are determined by rough screening. PCA not only simplifies the data structure of coal mine disasters, but also reduces the computational overhead, while retaining the amount of information of the original data as much as possible, improving the accuracy and interpretability of the analysis results, and further using the matrix method to express the distribution of eigenvectors, thereby building a spatial framework of eigenvectors and generating a dimensionality reduction space, so that the valid data can be reduced in dimension in this space, thereby improving the accuracy of the results.

[0053] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for processing coal mine disaster data based on principal component analysis, characterized in that: include: Step 1: Collect the working data of the coal mine, pre-process and standardize the working data to obtain the valid data of the coal mine; Step 2: Perform dimensionality reduction analysis on the valid data using PCA, convert the valid data into a number of feature vectors, and use the feature vectors to construct a dimensionality reduction space of the valid data; Step 3: Calculate the reconstruction error and explained variance ratio corresponding to each of the eigenvectors in the dimensionality reduction space respectively, and construct the dimensional information corresponding to each of the eigenvectors; Step 4: adjusting the dimension reduction space based on the dimension information, and inputting the valid data into the adjusted dimension reduction space for processing to generate dimension reduction data of the coal mine; Step 5: Screen out abnormal dimensionality reduction data segments that do not meet the specified explained variance ratio, optimize the abnormal dimensionality reduction data segments, and obtain the disaster data corresponding to the coal mine under different detection dimensions.

2. A method for processing coal mine disaster data based on principal component analysis as claimed in claim 1, characterized in that: The step 1 comprises: Step 11: collecting real-time gas concentration information, real-time temperature information, real-time humidity information, real-time mineralization information, real-time water tank flow information, real-time grouting information and real-time nitrogen injection information of the coal mine, and generating working data of the coal mine by time sorting; Step 12: obtaining the mean of the working data, analyzing the standard deviation distance between each data point in the working data and the mean, screening the standard deviation distance, determining a number of abnormal data points contained in the working data, and performing cleaning preprocessing on the abnormal data points; Step 13: construct a box plot of the working data, identify noise data points in the working data in the box plot, perform noise reduction preprocessing on the noise data points, and regard the working data that has completed the cleaning preprocessing and noise reduction preprocessing as the cleaning data of the coal mine; Step 14: Standardize the cleaned data to obtain valid data with a mean of 0 and a variance of 1.

3. A method for processing coal mine disaster data based on principal component analysis as claimed in claim 2, characterized in that: Also includes: Find the first outlier data point that exceeds the upper quartile + 1.5 times the interquartile range and the second outlier data point that exceeds the lower quartile - 1.5 times the interquartile range in the box plot; The first outlier data point and the second outlier data segment are regarded as noise data points in the working data.

4. The method for processing coal mine disaster data based on principal component analysis as claimed in claim 1, characterized in that: The step 2 comprises: Step 21: draw the original space of the valid data, search for the first several original coordinate axes with mutually orthogonal features in the original space, search for the target coordinate axis with the largest variance with the valid data as the first coordinate axis, locate the first coordinate axis in the original space, and determine the search starting point; Step 22: starting from the search starting point, sequentially search for a second coordinate axis that is orthogonal to the first coordinate axis and has the largest variance with the first coordinate plane, and a third coordinate axis that is orthogonal to the first coordinate axis and the second coordinate axis plane and has the largest variance with the first-second coordinate plane, and obtain N coordinate axes after the search is completed; Step 23: establishing a variance change list of the original space according to the variances between adjacent coordinate axes, roughly screening the variance change list to obtain K valid coordinate axes whose variances are not 0, and constructing a corresponding feature vector according to each of the valid coordinate axes; Step 24: Establish a covariance matrix of the eigenvector, analyze the eigenvalues ​​corresponding to each eigenvector in the covariance matrix using eigendecomposition method, build a spatial framework based on the eigenvectors, and input the eigenvalues ​​into the spatial framework to generate a dimensionality reduction space for the valid data.

5. The method for processing coal mine disaster data based on principal component analysis as claimed in claim 1, characterized in that: The step 3 comprises: Step 31: respectively obtain the explained variance ratio of each of the eigenvectors in the dimensionality reduction space, draw a visualization ratio diagram of the effective data, identify the importance corresponding to each of the eigenvectors in the visualization ratio diagram, and construct the first dimensional sub-information corresponding to the eigenvector; Step 32: construct corresponding feature data according to the feature vector, map the feature data to the original space for loss assessment, obtain data loss information corresponding to the feature vector, and construct the reconstruction error of the valid data; Step 33: construct the reconstruction error to determine the spatial structure error and spatial pattern error of the dimensionality reduction space, and generate the second dimensional sub-information corresponding to the feature vector; Step 34: Establish dimensional information corresponding to the feature vector in the dimensionality reduction space according to the first dimensional sub-information and the second dimensional sub-information.

6. The method for processing coal mine disaster data based on principal component analysis as claimed in claim 1, characterized in that: The step 4 comprises: Step 41: adjusting the spatial structure and spatial pattern of the dimensionality reduction space based on the dimensionality information to obtain a stable dimensionality reduction space; Step 42: input the valid data into the stable dimensionality reduction space, perform dimensionality reduction training on the valid data in the stable dimensionality reduction space, and generate dimensionality reduction data of the coal mine.

7. The method for processing coal mine disaster data based on principal component analysis as claimed in claim 1, characterized in that: The step 5 comprises: Step 51: Obtain the standard working range corresponding to each detection dimension in the coal mine, construct the prescribed explained variance ratio range of the corresponding detection dimension according to the standard working range, estimate the dimensional information corresponding to each detection dimension and construct a data partitioning rule; Step 52: dividing the dimension-reduced data into a plurality of dimension-reduced data segments based on the data partitioning rule, and identifying abnormal dimension data segments that do not meet the requirements using the corresponding specified explained variance rate range; Step 53: Mark the abnormal dimension data segment in the valid data, and track the source of each abnormal dimension data segment respectively, construct and display the disaster data corresponding to the coal mine under different detection dimensions.

8. The method for processing coal mine disaster data based on principal component analysis as claimed in claim 4, characterized in that: The process of establishing the covariance matrix of the eigenvector includes: Count all eigenvectors and construct the eigenvector set X; According to formula (1), the mean value corresponding to each of the eigenvectors is calculated respectively; （1） in, represents the mean of the first j vectors, n represents the total amount of the eigenvectors, Represents the i-th coordinate axis feature of the j-th eigenvector in the corresponding coordinate axis; According to formula (2), each of the feature vectors is decentralized to generate the target feature vector set ; Use formula (3) to establish the target feature vector set The covariance matrix of Among them, T represents the transpose symbol and C represents the covariance matrix symbol.

9. A system for processing coal mine disaster data based on principal component analysis, characterized in that: include: A data processing module is used to collect the working data of the coal mine, pre-process and standardize the working data, and obtain the effective data of the coal mine; A dimensionality reduction analysis module, used to perform dimensionality reduction analysis on the valid data using PCA, convert the valid data into a plurality of eigenvectors, and construct a dimensionality reduction space of the valid data using the eigenvectors; An information construction module, used to respectively calculate the reconstruction error and explained variance ratio corresponding to each of the eigenvectors in the dimensionality reduction space, and construct the dimensional information corresponding to each of the eigenvectors; A dimension reduction execution module, used for adjusting the dimension reduction space based on the dimension information, and inputting the valid data into the adjusted dimension reduction space for processing to generate dimension reduction data of the coal mine; The data processing module is used to screen out abnormal dimensionality reduction data segments that do not meet the specified explained variance ratio, optimize the abnormal dimensionality reduction data segments, and obtain the disaster data corresponding to the coal mine under different detection dimensions.

10. A system for processing coal mine disaster data based on principal component analysis as claimed in claim 9, characterized in that: The dimension reduction analysis module comprises: A space construction unit is used to draw the original space of the valid data, search for the first several original coordinate axes with mutually orthogonal characteristics in the original space, search for the target coordinate axis with the largest variance with the valid data as the first coordinate axis, locate the first coordinate axis in the original space, and determine the search starting point; A coordinate decomposition unit is used to sequentially search, starting from the search starting point, for a second coordinate axis that has an orthogonal relationship with the first coordinate axis and has the largest variance with the first coordinate plane, and a third coordinate axis that has an orthogonal relationship with the first coordinate axis and the second coordinate axis plane and has the largest variance with the first-second coordinate plane, and obtain N coordinate axes after completing the search; A coordinate screening unit, used to establish a variance change list of the original space according to the variance between adjacent coordinate axes, perform rough screening on the variance change list to obtain K valid coordinate axes whose variance is not 0, and construct a corresponding feature vector according to each of the valid coordinate axes; A space construction unit is used to establish a covariance matrix of the eigenvector, analyze the eigenvalues ​​corresponding to each eigenvector in the covariance matrix using the eigendecomposition method, build a space framework based on the eigenvectors, and input the eigenvalues ​​into the space framework to generate a dimensionality reduction space of the valid data.

Images