Coal gangue classification label generation method and system based on clustering algorithm

Abstract

The invention discloses a coal gangue classification label generation method and system based on a clustering algorithm. The method comprises the steps of obtaining N pieces of feature data related to coal gangue spontaneous combustion, and screening out M different features from the N pieces of feature data; the method comprises the following steps of: investigating a large number of documents related to spontaneous combustion of coal gangue, capturing the characteristics of high occurrence frequency related to spontaneous combustion of coal gangue in all documents, and screening out L characteristics through expert judgment by combining M characteristics; summarizing feature data belonging to features in L features included in each piece of the N pieces of feature data, preprocessing any piece of feature data having a missing value after summarization to obtain H pieces of feature data, each piece of feature data including L pieces of feature data; calculating a correlation coefficient between every two features in the L features, and deleting and optimizing the L features according to the correlation coefficients to obtain I features; calculating the weight of each feature; and clustering the H pieces of feature data subjected to negative correlation feature data inversion processing by adopting a clustering model in combination with the weight to obtain a final coal gangue spontaneous combustion category label of each piece of the H pieces of feature data.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and resource and environmental technology, and in particular to a method and system for generating coal gangue classification labels based on a clustering algorithm. Background Art

[0002] In the current national bulk solid waste treatment environment, coal gangue is one of the industrial wastes with the largest emission and cumulative volume, and the most serious hazards in my country. The existing gangue in my country has accumulated more than 6 billion tons (of which North China accounts for about 1 / 2), forming 1,600-1,800 gangue mountains, occupying and polluting a large amount of land resources. Untreated spontaneous combustion and dumping, coupled with long-term heat storage, have caused more than 300 of them to be in a state of long-term spontaneous combustion. The long-term spontaneous combustion of gangue mountains will emit 10.8gCO, about 2gH2S and NO per square meter of burning area. X , about 6.5g SO2, as well as polycyclic aromatic hydrocarbons (PAHs) such as benzopyrene and benzofluoranthene, and particulate matter, polluting the atmosphere and endangering health. Over 50 explosions have occurred, including nearly 10 major and serious ones. Spontaneously combusting gangue piles, eroded by heavy rainfall and experiencing thermal expansion during the rainy season, can easily lead to landslides, collapses, and explosions. These hazards not only affect the physical and mental health of local residents but also pose a serious threat to coal mines and society. Therefore, the treatment and disposal of coal gangue in my country is urgent.

[0003] Due to the ecological damage caused by high-intensity, large-scale coal mining in my country's major coal mining areas, which has led to problems such as surface subsidence and collapse, and groundwater drainage, a "underground filling + surface backfill" treatment method has been adopted to reduce gangue volume. This treatment method primarily considers the impact of gangue spontaneous combustion on the design and effectiveness of the treatment plan. Therefore, it is necessary to pre-classify gangue according to its spontaneous combustion characteristics and then assign different backfill plans to gangue of different combustion levels, thereby saving manpower, material and financial resources and improving efficiency.

[0004] However, current methods for identifying the spontaneous combustion tendency of coal gangue primarily rely on experimental data and empirical formulas. For example, one method for identifying the spontaneous combustion tendency of coal gangue derives a formula based on the sulfur content and crossover temperature of coal gangue samples. However, these methods suffer from issues such as incomplete data collection, inaccurate feature selection, and insufficient model validation, making them difficult to meet the high-precision requirements of practical applications. Furthermore, the Polish PSO / Z method (Particle Swarm Optimization and Z-Score Method) is widely used to predict the spontaneous combustion tendency of coal gangue. The PSO / Z method combines the particle swarm optimization algorithm (PSO) with the Z-Score analysis method. In predicting the spontaneous combustion tendency of coal gangue, PSO is used to optimize the parameters of the prediction model to enhance its predictive power, while the Z-Score analysis method is used to assess the spontaneous combustion risk of different gangue samples. Specifically, PSO is used to optimize the hyperparameters of neural networks, support vector machines (SVMs), or other machine learning algorithms to improve the accuracy of spontaneous combustion predictions. The Z-Score method can effectively eliminate noisy or extreme data through standardization and anomaly detection, thereby improving the stability of the prediction model. By combining PSO and Z-Score, researchers can extract key features that influence gangue spontaneous combustion from large amounts of raw data, thereby establishing a more accurate prediction model. Although the PSO / Z method has improved prediction accuracy to a certain extent, there is still room for improvement when dealing with complex and dynamically changing environmental factors. Furthermore, the method studies a relatively small number of gangue characteristics, which has certain limitations.

[0005] To this end, a machine learning algorithm is used to classify gangue based on its spontaneous combustion characteristics. However, the classification requires pre-determining the category labels for each piece of data. Therefore, the present invention uses a clustering algorithm to integrate multiple gangue component attribute characteristics to cluster the gangue. The clustering results are verified using the clustering algorithm's internal evaluation indicators and a visual scatter plot of the clustering results to obtain the final gangue spontaneous combustion category label, providing prerequisite technical support for subsequent labeled gangue classification prediction and backfill plan issuance. Summary of the Invention

[0006] In view of the problems and shortcomings of the prior art, the present invention provides a method and system for generating coal gangue classification labels based on a clustering algorithm.

[0007] The present invention solves the above technical problems through the following technical solutions:

[0008] The present invention provides a method for generating coal gangue classification labels based on a clustering algorithm, which is characterized by comprising the following steps:

[0009] S1. Preliminarily collect multiple feature data related to gangue spontaneous combustion from each of N historical gangue spontaneous combustion related materials to obtain N pieces of feature data, and screen out M different features based on the N pieces of feature data;

[0010] S2. By researching a large number of literature related to gangue spontaneous combustion, we captured the features with high frequency of occurrence related to gangue spontaneous combustion in all the literature. At the same time, we combined M features and screened out L features through expert evaluation, where L < M;

[0011] S3. Summarize the feature data of each of the N feature data pieces that belong to the L features, and perform preprocessing of deleting or filling in any feature data with missing values in the N feature data pieces after aggregation, to obtain H feature data pieces, each of which includes L feature data;

[0012] S4. Calculate the correlation coefficient between each pair of L features using H feature data, and perform optimization on the L features based on the correlation coefficient to obtain I features, where I≤L, and each feature data contains I feature data;

[0013] S5. Research K articles related to spontaneous combustion of coal gangue, count the frequency of each feature in I in the K articles, and calculate the feature weight of each feature;

[0014] Among the S6 and H feature data, the feature data that is negatively correlated with the spontaneous combustion of coal gangue is inverted to obtain the feature data that is positively correlated. Combined with the feature weights, the weighted WK-means clustering model is used to cluster the H feature data after inversion processing, and cluster them into three clusters: not easy to spontaneously combust, easy to spontaneously combust, and very easy to spontaneously combust, and the clustering results are obtained;

[0015] S7. Perform multiple verifications on the clustering results to obtain the final category label of gangue spontaneous combustion for each feature data in the H feature data.

[0016] The present invention also provides a method for generating coal gangue classification labels based on a clustering algorithm, which is characterized by comprising:

[0017] a data collection module for pre-collecting multiple feature data related to gangue spontaneous combustion from N historical gangue spontaneous combustion related materials, obtaining N pieces of feature data, and screening out M different features based on the N pieces of feature data;

[0018] A feature screening module is used to investigate a large number of literature related to gangue spontaneous combustion, capture the features with high frequency of occurrence related to gangue spontaneous combustion in all literature, and combine M features and screen out L features through expert evaluation, where L < M;

[0019] a preprocessing module for summarizing feature data belonging to the L features included in each of the N feature data pieces, and performing preprocessing of deleting or filling in any feature data having missing values in the N feature data pieces after summarization, to obtain H feature data pieces, each of which includes the L feature data pieces;

[0020] a correlation calculation module, configured to calculate the correlation coefficients between each pair of L features using H pieces of feature data, and to perform reduction and optimization on the L features based on the correlation coefficients to obtain I features, where I≤L, and each piece of feature data contains I feature data;

[0021] A weight calculation module is used to investigate K articles related to gangue spontaneous combustion, and to count the frequency of each feature in I in the K articles, so as to calculate the feature weight of each feature;

[0022] A clustering module is used to invert the feature data of H pieces of feature data that are negatively correlated with the spontaneous combustion of coal gangue to obtain feature data that are positively correlated. Combined with the feature weights, the weighted WK-means clustering model is used to cluster the H pieces of feature data after inversion processing into three clusters: not easy to spontaneously combust, easy to spontaneously combust, and very easy to spontaneously combust, thereby obtaining a clustering result.

[0023] A category acquisition module is used to perform multiple verifications on the clustering results and obtain the final category label of coal gangue spontaneous combustion for each feature data in H feature data.

[0024] The positive progress effect of the present invention is:

[0025] The present invention proposes a method and system for generating coal gangue classification labels based on a clustering algorithm, which breaks the classification limitations of traditional classification standards for single features, avoids the subjectivity and limitations of manual clustering, and provides a leading condition for the construction of subsequent classification prediction models and the issuance of backfill plans.

[0026] The present invention collects characteristic data related to gangue spontaneous combustion by searching well-known literature databases and combining them with historical experimental data. The collected data is then subjected to feature screening to identify characteristic data highly correlated with gangue spontaneous combustion. The screened data is then preprocessed, including deleting rows with severe data omissions and predicting missing values. The finalized data is then evaluated for feature relevance. A simple frequency-based weighting method is used to calculate the weight of each feature relative to gangue spontaneous combustion. The gangue data is then divided into three clusters using a combination of weights and a weighted WK-means clustering algorithm. The clustering model performance is evaluated using visualization of clustering results using scatter plots and the Calinski-Harabasz index. This method provides effective technical support for the establishment of data labels based on gangue spontaneous combustion classification, and also provides a foundation for the subsequent construction of a prediction model based on gangue component attribute characteristics and the determination of backfilling plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a method for generating coal gangue classification labels based on a clustering algorithm according to a preferred embodiment of the present invention.

[0028] Figure 2 This is a scatter plot showing the visualization of clustering results of the method for generating coal gangue classification labels based on clustering algorithm in a preferred embodiment of the present invention.

[0029] Figure 3 This is a Calinski-Harabasz index histogram of clustering results of the coal gangue classification label generation method based on clustering algorithm in a preferred embodiment of the present invention.

[0030] Figure 4 This is a block diagram of a coal gangue classification label generation system based on a clustering algorithm in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0031] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0032] For ease of description, only the parts related to the present invention are shown in the accompanying drawings. The first, second, etc. involved in the present invention are only for the convenience of describing the technical solution of the present invention and do not have a specific limiting effect. They are all general references and do not constitute a limiting effect on the technical solution of the present invention.

[0033] like Figure 1As shown, an embodiment of the present invention provides a method for generating coal gangue classification labels based on a clustering algorithm, comprising the following steps:

[0034] Step 101: pre-collect multiple feature data related to gangue spontaneous combustion from N historical gangue spontaneous combustion related materials, obtain N feature data, and screen out M different features based on the N feature data.

[0035] Where N is the number of historical gangue spontaneous combustion related materials collected, and M is the number of features collected related to gangue spontaneous combustion. One piece of historical gangue spontaneous combustion related material collects one piece of feature data, and each piece of feature data includes multiple feature data.

[0036] In step 101, a search is conducted on published documents in various existing databases using keywords related to gangue spontaneous combustion. Multiple feature data related to gangue spontaneous combustion are found from each retrieved document. The multiple feature data obtained from any one document are used as one feature data item. For example, using keywords such as "gangue," "gangue spontaneous combustion," and "comprehensive utilization of gangue," a comprehensive and in-depth literature review is conducted on published documents in well-known databases such as Web of Science, Wanfang, HowNet, and Scopus. Based on the keywords, multiple features related to spontaneous combustion are found from a large number of relevant documents. After screening, 135 feature data items are screened out, including 14 features.

[0037] Experimental data of multiple coal mines were obtained from the Ecological Environment Bureau. The experimental data included multiple characteristic data related to spontaneous combustion of coal gangue. The multiple characteristic data obtained based on any experimental data were used as one characteristic data.

[0038] Obtain N pieces of feature data and filter out M different features based on the N pieces of feature data. For example, N is 135 and M is 14.

[0039] Step 102: By investigating a large number of documents related to spontaneous combustion of coal gangue, the features with high occurrence frequency related to spontaneous combustion of coal gangue in all documents are captured. At the same time, M features are combined and judged by experts to select L features, where L<M, for example, L is 9.

[0040] In this step, by investigating a large number of documents related to gangue spontaneous combustion, the features with high frequency of occurrence related to gangue spontaneous combustion in all documents are captured. At the same time, combined with the collected features related to gangue spontaneous combustion, 9 features that are more important for gangue spontaneous combustion are screened out as the final input data.

[0041] Step 103: Summarize the feature data belonging to the L features included in each of the N pieces of feature data. Perform preprocessing to delete or fill in any missing value in the N pieces of feature data, thereby obtaining H pieces of feature data, each of which includes L features. In each of the H pieces of feature data, data exists for all L features included, with no missing values.

[0042] In step 103, the feature data of each feature in the N feature data items that belongs to the L features are summarized to construct a first Excel table containing N rows and L columns of feature data. For any row, the feature data in the row is analyzed to see if there are any missing features. If there are no missing features, the row is retained. If there are missing features, further analysis is performed. If there are two or more missing feature data, the row is directly deleted. If there is only one missing feature data, the regression filling method is used to predict and fill in the missing feature data. Based on this, a second Excel table containing H rows and L columns of feature data is obtained, for example, L is 9 and H is 85.

[0043] The regression filling method includes:

[0044] S31. Select all row feature data without missing values from the first Excel table as training samples, use the missing feature of the row as the target variable and the other L-1 features as explanatory variables, and establish a regression model.

[0045] Common regression models include linear regression, ridge regression, and Lasso regression. This example uses a linear regression model. Based on the training set (data without missing values), the target variable (sulfur trioxide SO3) is predicted using selected explanatory variables (fixed carbon, sulfur, moisture, ash, volatile matter, Al2O3, SiO2, and CaO). The specific steps are as follows:

[0046] With the known explanatory variables (X n ) to predict the value of the target variable (Y), train the regression model, and fit an equation: Y = β0 + β1*X1 + β2*X2 + ... + β n *X n .

[0047] S32. The feature data corresponding to the explanatory variable of each row in the training sample is used as input, and the feature data corresponding to the target variable is used as output. The training sample is used to train the regression model to obtain a trained regression model.

[0048] S33. Substitute the other L-1 feature data of the row into the trained regression model to predict the target variable value. The target variable value is used to fill the feature data of this feature that is missing in the row.

[0049] For target variables with missing values, we use regression models and known explanatory variables (Xn ) to predict the missing values of the target variable (Y: SO3); fill the missing positions in the original data set with the predicted missing values.

[0050] In this step, for data items with serious missing feature values (two or more of the nine features of each data item are missing), the rows containing missing values are deleted; for the remaining data items with fewer missing features (only one of the nine features of each data item is missing), the regression imputation method is used to predict and fill the missing values, and finally 85 data items containing 9 features (fixed carbon, sulfur, moisture, ash, volatile matter, Al2O3, SiO2, CaO, SO3) are obtained.

[0051] Step 104: Calculate the correlation coefficient between each pair of the L features using the H feature data. Delete and optimize the L features based on the correlation coefficient to obtain I features, where I≤L, and each feature data contains I feature data.

[0052] In step 104, an H*L matrix is constructed for H pieces of feature data. The H*L matrix is input into the Origin professional function drawing software. The correlation coefficient between each pair of features is calculated using the Pearson Correlation Coefficient formula. Each calculated correlation coefficient is analyzed. When any correlation coefficient is greater than a set threshold (such as 0.9), the two features corresponding to the correlation coefficient are further analyzed to see whether they are positively correlated or negatively correlated with the spontaneous combustion of coal gangue. If the two features are respectively positively correlated and negatively correlated with the spontaneous combustion of coal gangue, the two features are retained. If both features are positively correlated or negatively correlated with the spontaneous combustion of coal gangue, one of the features is deleted. Thus, each piece of feature data contains L-1 features. After the above operations, this embodiment does not delete any of the 9 features, and ultimately there are still 9 features, that is, I=L.

[0053] In this step, the Pearson correlation coefficient is used to evaluate the feature correlation of the 9 features. The Pearson correlation coefficient is used to measure the linear relationship between two continuous variables, and the value range is -1 to 1 (+1 indicates a perfect positive correlation; -1 indicates a perfect negative correlation; 0 indicates no linear relationship).

[0054] Pearson correlation coefficient formula:

[0055]

[0056] x i ,y i is the sample value, is the sample mean.

[0057] Step 105: Research K articles related to gangue spontaneous combustion, count the frequency of each of the L features in the K articles, and calculate the feature weight of each feature for subsequent weighting processing. For example, K is 50.

[0058] For example: 50 papers on spontaneous combustion of coal gangue were surveyed, and the frequency of occurrence of 9 characteristics was counted (fixed carbon 36 times, sulfur 43 times, moisture 38 times, ash 12 times, volatile matter 17 times, Al2O3 6 times, SiO2 6 times, CaO 6 times, SO3 3 times). The statistical data were input into SPSS, and the weight of each characteristic relative to spontaneous combustion of coal gangue was obtained through a simple weighted method based on frequency (fixed carbon: 0.215, sulfur: 0.257, moisture: 0.228, ash: 0.072, volatile matter: 0.102, Al2O3: 0.036, SiO2: 0.036, CaO: 0.036, SO3: 0.018).

[0059] Step 106: Among the H pieces of feature data, the feature data that is negatively correlated with the spontaneous combustion of coal gangue is inverted to obtain feature data that is positively correlated. Combined with the feature weights, the weighted WK-means clustering model is used to cluster the H pieces of feature data after inversion processing, and the data are clustered into three clusters: not easy to spontaneously combust, easy to spontaneously combust, and extremely easy to spontaneously combust, thereby obtaining a clustering result.

[0060] In step 106, S61, for the second Excel table containing H rows and L columns of feature data, the feature data of the feature that is negatively correlated with the spontaneous combustion of coal gangue is inverted using the inverted maximum subtraction method to obtain feature data that is positively correlated, the feature weight of each feature is read, and the feature data of the H rows and L columns after the inversion processing is weighted by the feature weight to obtain the weighted feature value of the H rows and L columns.

[0061] S62. Use the data standardization (StandardScaler) method to standardize the H rows and L columns of weighted eigenvalues to obtain H rows and L columns of standardized eigenvalues; the standardization process unifies the scale of each feature to avoid the weighted WK-means clustering algorithm being affected by scale differences.

[0062] S63. Set the number of clusters to 3 and the random seed to 42 to ensure the reproducibility of the results. Input the standardized eigenvalues of H rows and L columns into the weighted WK-means clustering model for clustering to obtain the clustering result label of each row. The clustering result labels are not easy to spontaneously combust, easy to spontaneously combust, and extremely easy to spontaneously combust. The three clusters are simplified to 0, 1, and 2, respectively. Add the clustering result labels as the L+1 column in the second Excel table, and add the simplified clustering result labels as the L+2 column, forming H rows (L+2) columns of data.

[0063] For example: the two negatively correlated features "ash content (%)" and "CaO (%)" are reversed so that they have the same change trend as other positively correlated features (in this embodiment, the negatively correlated feature inversion method adopts the reverse maximum subtraction method, that is, for the two feature values of "ash content (%)" and "CaO (%)" of each sample, the maximum value of the feature is subtracted from the value of the sample); each column of features in the data is traversed, and if the feature is in the weight dictionary, it is multiplied by the corresponding weight to obtain the weighted feature value.

[0064] Step 107: Multiple verifications are performed on the clustering results using a clustering result visualization scatter plot, Calinski-Harabasz index, etc., to obtain a final category label of gangue spontaneous combustion for each of the H feature data.

[0065] In step 107,

[0066] S71. Obtain the coordinates of the center point of each cluster from the weighted WK-means clustering model after clustering. These center points reflect the average level of the characteristics of each cluster.

[0067] S72. Output the complete data containing the clustering results (H rows and (L+2) columns of data) and the cluster center coordinates of each cluster, and save the complete data with the clustering results into a new Excel file for subsequent viewing and analysis.

[0068] S73. Use the t-SNE algorithm to reduce the dimensionality of the H rows and L columns of data to a two-dimensional space for visualization. Add the t-SNE results to a new Excel file to form new "t-SNE_1" and "t-SNE_2" columns. Based on the t-SNE results, mark the data points of different clusters with different colors to visually demonstrate the clustering effect. Draw the cluster center points on the scatter plot and mark them with red cross shapes to highlight the core positions of each cluster.

[0069] S74. Use the pip install command to install the matplotlib, seaborn, and scikit-learn libraries, which are used for data visualization and clustering result analysis.

[0070] S75. Use the calinski_harabasz_score function to calculate the Calinski-Harabasz index to measure the variance ratio of the clustering. The larger the value, the better the clustering effect.

[0071] S76. Display the clustered results in a scatter plot, such as Figure 2 As shown; the calculated Calinski-Harabasz index performance indicators are displayed in the form of a bar chart, as shown Figure 3 The Calinski-Harabasz index (CHIndex) is 20.06587142368264, while the CH Index of other clustering models is 10.17225. These two indicators indicate that this method has a good clustering effect and provide label data for the subsequent classification of coal gangue.

[0072] like Figure 4 As shown, an embodiment of the present invention also provides a method for generating coal gangue classification labels based on a clustering algorithm, including a data collection module 1, a feature screening module 2, a preprocessing module 3, a correlation calculation module 4, a weight calculation module 5, a clustering module 6 and a category acquisition module 7.

[0073] The data collection module 1 is used to pre-collect multiple feature data related to gangue spontaneous combustion in each of N historical gangue spontaneous combustion related materials, obtain N feature data, and screen out M different features based on the N feature data.

[0074] The feature screening module 2 is used to investigate a large number of documents related to spontaneous combustion of coal gangue, capture the features with high occurrence frequency related to spontaneous combustion of coal gangue in all documents, and at the same time combine M features and screen out L features through expert evaluation, where L<M.

[0075] The preprocessing module 3 is used to summarize the feature data of each feature in the N feature data belonging to the L features, and perform preprocessing on the feature data with missing values in any of the N feature data after summary to obtain H feature data, each of which includes L feature data.

[0076] Correlation calculation module 4 is used to calculate the correlation coefficient between each pair of L features using H feature data. Based on the correlation coefficient, the L features are pruned and optimized to obtain I features, where I≤L, and each feature data piece contains I feature data. In this embodiment, no feature is pruned from the 9 features, and ultimately, there are still 9 features, i.e., I=L.

[0077] The weight calculation module 5 is used to investigate K documents related to spontaneous combustion of coal gangue, and to count the frequency of each of the L features appearing in the K documents, so as to calculate the feature weight of each feature.

[0078] Clustering module 6 is used to invert the feature data of H pieces of feature data that are negatively correlated with spontaneous combustion of coal gangue to obtain feature data that are positively correlated. Combined with the feature weights, the weighted WK-means clustering model is used to cluster the H pieces of feature data after inversion processing, and cluster them into three clusters: not easy to spontaneously combust, easy to spontaneously combust, and extremely easy to spontaneously combust, to obtain the clustering results.

[0079] The category acquisition module 7 is used to perform multiple verifications on the clustering results to obtain the final category label of gangue spontaneous combustion for each feature data in the H feature data.

[0080] An embodiment of the present invention further provides an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0081] An embodiment of the present invention further provides a computer-readable storage medium having computer program instructions stored thereon, and the computer program instructions implement the aforementioned method when executed by a processor.

[0082] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0083] Although specific embodiments of the present invention have been described above, those skilled in the art will appreciate that these are merely illustrative and that the scope of the present invention is defined by the appended claims. Those skilled in the art may make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, and such changes and modifications are intended to fall within the scope of the present invention.

Claims

1. A method for generating coal gangue classification labels based on clustering algorithm, characterized in that: include: S1. Preliminarily collect multiple feature data related to gangue spontaneous combustion from each of N historical gangue spontaneous combustion related materials to obtain N pieces of feature data, and screen out M different features based on the N pieces of feature data; S2. By researching a large number of literature related to gangue spontaneous combustion, we captured the features with high frequency of occurrence related to gangue spontaneous combustion in all the literature. At the same time, we combined M features and screened out L features through expert evaluation, where L < M; S3. Summarize the feature data of each of the N feature data pieces that belong to the L features, and perform preprocessing of deleting or filling in any feature data with missing values in the N feature data pieces after aggregation, to obtain H feature data pieces, each of which includes L feature data; S4. Calculate the correlation coefficient between each pair of L features using H feature data, and perform optimization on the L features based on the correlation coefficient to obtain I features, where I≤L, and each feature data contains I feature data; S5. Research K articles related to spontaneous combustion of coal gangue, count the frequency of each feature in I in the K articles, and calculate the feature weight of each feature; Among the S6 and H feature data, the feature data that is negatively correlated with the spontaneous combustion of coal gangue is inverted to obtain the feature data that is positively correlated. Combined with the feature weights, the weighted WK-means clustering model is used to cluster the H feature data after inversion processing, and cluster them into three clusters: not easy to spontaneously combust, easy to spontaneously combust, and very easy to spontaneously combust, and the clustering results are obtained; S7. Perform multiple verifications on the clustering results to obtain the final category label of gangue spontaneous combustion for each feature data in the H feature data.

2. The method for generating coal gangue classification labels based on clustering algorithm according to claim 1, wherein In S1, the words related to gangue spontaneous combustion are used as keywords to search the published documents in various existing databases, and multiple feature data related to gangue spontaneous combustion are found from each retrieved document. The multiple feature data obtained based on any document are regarded as one feature data; Obtain experimental data from multiple coal mines from the Ecological Environment Bureau. The experimental data includes multiple characteristic data related to spontaneous combustion of coal gangue. The multiple characteristic data obtained based on any experimental data are regarded as one characteristic data item. Obtain N pieces of feature data, and filter out M different features based on the N pieces of feature data.

3. The method for generating coal gangue classification labels based on clustering algorithm according to claim 1, characterized in that: In S3, the feature data of each of the N feature data belonging to the L features is summarized to construct a first Excel table containing N rows and L columns of feature data. For any row, the feature data of the row is analyzed to see if there are any missing features. If there are no missing features, the row is retained. If there are missing features, further analysis is performed. If there are two or more missing feature data, the row is directly deleted. If there is only one missing feature data, the missing feature data is filled in. Based on this, a second Excel table containing H rows and L columns of feature data is obtained.

4. The method for generating coal gangue classification labels based on clustering algorithm according to claim 3, characterized in that: In S3, there is only one missing feature data in this row, and the regression filling method is used to predict and fill in the missing feature data; The regression filling method includes: S31. Select all row feature data without missing values from the first Excel table as training samples, use the missing feature of the row as the target variable and the other L-1 features as explanatory variables, and establish a regression model; S32, taking the feature data corresponding to the explanatory variable of each row in the training sample as input and the feature data corresponding to the target variable as output, and using the training sample to train the regression model to obtain a trained regression model; S33. Substitute the other L-1 feature data of the row into the trained regression model to predict the target variable value. The target variable value is used to fill the feature data of this feature that is missing in the row.

5. The method for generating coal gangue classification labels based on clustering algorithm according to claim 1, wherein In S4, for H pieces of feature data, an H*L matrix is constructed, and the H*L matrix is input into the Origin professional function drawing software. The correlation coefficient between each pair of features is calculated using the Pearson correlation coefficient formula, and each calculated correlation coefficient is analyzed. When any correlation coefficient is greater than the set threshold, the two features corresponding to the correlation coefficient are further analyzed to see whether they are positively correlated or negatively correlated with the spontaneous combustion of coal gangue. If these two features are respectively positively correlated and negatively correlated with the spontaneous combustion of coal gangue, the two features are retained. If both features are positively correlated or negatively correlated with the spontaneous combustion of coal gangue, one of the features is deleted. Therefore, each piece of feature data contains L-1 features. After this operation, no feature is deleted, and there are still L features in the end, that is, I=L.

6. The method for generating coal gangue classification labels based on clustering algorithm according to claim 5, characterized in that: In S6, S61, for a second Excel table containing H rows and L columns of feature data, invert the feature data of the feature that is negatively correlated with spontaneous combustion of coal gangue using inverted maximum subtraction to obtain feature data that is positively correlated, read the feature weight of each feature, and weight the inverted feature data of H rows and L columns to obtain a weighted feature value of H rows and L columns; S62, using a data standardization method to standardize the weighted eigenvalues of H rows and L columns to obtain H rows and L columns of standardized eigenvalues; S63. Set the number of clusters to 3 and the random seed to 42. Input the standardized eigenvalues of H rows and L columns into the weighted WK-means clustering model for clustering to obtain the clustering result label of each row. The clustering result labels are not easy to spontaneously combust, easy to spontaneously combust, and extremely easy to spontaneously combust. Add the clustering result labels as the L+1th column in the second Excel table to form H rows (L+1) columns of data.

7. The method for generating coal gangue classification labels based on clustering algorithm according to claim 6, characterized in that: The three clusters are simplified as 0, 1, and 2 respectively; In S63 , the simplified labels of the clustering results are added as the L+2th column to form H rows (L+2) columns of data.

8. The method for generating coal gangue classification labels based on clustering algorithm according to claim 7, wherein In S7, S71, obtaining the center point coordinates of each cluster from the weighted WK-means clustering model after clustering; S72. Output H rows (L+2) columns of data containing the clustering results and the coordinates of the cluster centers of each cluster, and save the H rows (L+2) columns of data containing the clustering results into a new Excel file; S73. Use the t-SNE algorithm to reduce the dimensionality of the H rows and L columns of data to a two-dimensional space. Add the t-SNE results to a new Excel file to form new columns "t-SNE_1" and "t-SNE_2". Based on the t-SNE results, mark the data points of different clusters with different colors to visually demonstrate the clustering effect. Draw the cluster centers on the scatter plot and mark them with red crosses to highlight the core positions of each cluster. S74. Install matplotlib, seaborn, and scikit-learn libraries using the pip install command. S75. Calculate the Calinski-Harabasz index using the calinski_harabasz_score function; S76. Display the clustering results in a scatter plot and the Calinski-Harabasz index performance indicator in the form of a bar chart.

9. A method for generating coal gangue classification labels based on a clustering algorithm, characterized by comprising: a data collection module for pre-collecting multiple feature data related to gangue spontaneous combustion from N historical gangue spontaneous combustion related materials, obtaining N pieces of feature data, and screening out M different features based on the N pieces of feature data; A feature screening module is used to investigate a large number of literature related to gangue spontaneous combustion, capture the features with high frequency of occurrence related to gangue spontaneous combustion in all literature, and combine M features and screen out L features through expert evaluation, where L < M; a preprocessing module for summarizing feature data belonging to the L features included in each of the N feature data pieces, and performing preprocessing of deleting or filling in any feature data having missing values in the N feature data pieces after summarization, to obtain H feature data pieces, each of which includes the L feature data pieces; a correlation calculation module, configured to calculate the correlation coefficients between each pair of L features using H pieces of feature data, and to perform reduction and optimization on the L features based on the correlation coefficients to obtain I features, where I≤L, and each piece of feature data contains I feature data; A weight calculation module is used to investigate K articles related to gangue spontaneous combustion, and to count the frequency of each feature in I in the K articles, so as to calculate the feature weight of each feature; A clustering module is used to invert the feature data of H pieces of feature data that are negatively correlated with the spontaneous combustion of coal gangue to obtain feature data that are positively correlated. Combined with the feature weights, the weighted WK-means clustering model is used to cluster the H pieces of feature data after inversion processing into three clusters: not easy to spontaneously combust, easy to spontaneously combust, and very easy to spontaneously combust, thereby obtaining a clustering result. A category acquisition module is used to perform multiple verifications on the clustering results and obtain the final category label of coal gangue spontaneous combustion for each feature data in H feature data.

10. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call instructions stored in the memory to execute the method according to any one of claims 1 to 7.

11. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images