A knowledge graph-based mine target area delineation method and system

Abstract

The application discloses a kind of based on knowledge graph's mine target area delineation method and system, the method includes: obtaining geochemical data and constructing geochemical knowledge graph;The lower limit value of geochemical element anomaly in knowledge graph is calculated, and the abnormal value combination is obtained;Abnormal value combination is carried out thermal matrix visualization analysis, and the abnormal value combination relationship is obtained;According to abnormal value combination relationship, delineate mine target area.Through using the application, learning and query reasoning can be carried out without needing a large number of training samples, delineate mine target area, guide mineral resources prospect prediction.The application is a kind of based on knowledge graph's mine target area delineation method and system, and can be widely applied in geological science and technology field.

Description

Technical Field

[0001] This invention relates to the field of geological science and technology, and in particular to a method and system for delineating mineral target areas based on knowledge graphs. Background Technology

[0002] Geochemical data, involving features formed during complex mineralization processes, is crucial for mineral exploration and prediction. Extracting multi-scale ore-forming element features and identifying geochemical anomalies is essential. Many frequency-based and frequency-space-based statistical methods have been used to identify geochemical anomalies. Frequency-based methods, such as univariate statistical methods, probability maps, multivariate statistical methods, fractal and multifractal models, and geostatistics, often exhibit high dimensionality, spatial correlation, and nonlinearity due to the complexity of geological mineralization processes across different spatial and temporal scales. This hinders the direct identification and delineation of geochemical anomalies using these classical methods. Existing machine learning and deep learning technologies, as data-driven methods, provide powerful tools for mineral prospect modeling. However, machine learning algorithms require a sufficient number of mineral point training samples. Insufficient training samples lead to poor extraction of ore-related features, resulting in ineffective and unreliable results. Over-reliance on training data also affects the generalization ability of machine learning models. Deep learning methods lack transparency, essentially acting as a "black box," with effectiveness largely dependent on the parameters set by researchers. Therefore, it is difficult to interpret the input and output results of deep learning models from a geological perspective, and the underlying mechanisms cannot be inferred. Summary of the Invention

[0003] To address the aforementioned technical problems, the present invention aims to provide a knowledge graph-based method and system for delineating mineral target areas, which can perform learning and query reasoning without requiring a large number of training samples to delineate mineral target areas and guide mineral prospect prediction.

[0004] The first technical solution adopted in this invention is: a method for delineating mineral target areas based on knowledge graphs, comprising the following steps:

[0005] Acquire geochemical data and construct a geochemical knowledge graph;

[0006] Calculate the lower limit of geochemical element anomalies in the knowledge graph to obtain combinations of anomalies;

[0007] A thermodynamic matrix visualization analysis was performed on outlier combinations to obtain the relationships between them.

[0008] Based on the relationships between outlier combinations, the target mining area is delineated.

[0009] Furthermore, the step of acquiring geochemical data and constructing a geochemical knowledge graph specifically includes:

[0010] Acquire geochemical data;

[0011] Geochemical data are analyzed to obtain the abundance values ​​and attributes of the geochemical data;

[0012] Geochemical ternary datasets were constructed based on abundance values ​​and spatial attributes of geochemical data.

[0013] Build a database using Neo4j software;

[0014] Geochemical ternary datasets were imported into a database to obtain a geochemical knowledge graph.

[0015] Furthermore, the step of calculating the lower limit of geochemical element anomalies in the knowledge graph to obtain the combination of outliers specifically includes:

[0016] Calculate the mean and standard deviation of geochemical elements in the knowledge graph to obtain the original data values;

[0017] The original data values ​​are analyzed and processed to remove data values ​​that are greater than a preset value, resulting in the data values ​​after removal.

[0018] The mean and standard deviation of the removed data values ​​are calculated, and the sum of the mean and twice the standard deviation is taken as the lower limit of geochemical element anomalies.

[0019] Based on the lower limit of geochemical element anomalies, the nodes of the lower limit of anomalies are obtained by querying using the Cypher statement in the Neo4j software;

[0020] The number of nodes is calculated based on the lower limit of geochemical element anomalies, and then the proportion of nodes for different lower limit geochemical element anomalies is calculated using a proportional formula to construct anomaly combination.

[0021] Furthermore, the formula for calculating the root mean square error of the geochemical elements is as follows:

[0022]

[0023] In the above formula, S1 represents the original mean squared error of the geochemical elements, n represents the original quantity of the geochemical elements, and X... i Represents the original sample values ​​of geochemical elements. This represents the average value of the original sample of geochemical elements.

[0024] Furthermore, the step of performing thermal matrix visualization analysis on outlier combinations to obtain the relationships between outlier combinations specifically includes:

[0025] The coefficients of geochemical elements are obtained by calculating the combination of outliers using a thermodynamic matrix.

[0026] The correlation coefficients of geochemical elements are obtained by visualizing their coefficient values.

[0027] Determine the correlation coefficients of geochemical elements;

[0028] If the correlation coefficient of a geochemical element is found to be greater than a preset value, the geochemical elements are combined to obtain an outlier combination relationship.

[0029] Furthermore, the formula for calculating the thermal matrix is ​​as follows:

[0030]

[0031] In the above formula, ρ represents the correlation coefficient value of geochemical elements, and X i X j σX represents a geochemical element. i ,X j This represents the covariance value of two geochemical elements. It represents the standard deviation between two geochemical elements.

[0032] Furthermore, the step of delineating the target mining area based on the outlier combination relationship specifically includes:

[0033] Based on the combination relationships of outliers, visualization interpolation is performed in GIS to obtain an outlier map;

[0034] Anomaly graphs are overlaid to construct anomaly graph combinations;

[0035] Based on the combination of anomaly maps, the abnormal data are delineated to obtain the delineated mineral target area.

[0036] The second technical solution adopted in this invention is: a mineral target area delineation system based on knowledge graphs, comprising:

[0037] The module is used to acquire geochemical data and construct a geochemical knowledge graph;

[0038] The calculation module is used to calculate the lower limit of geochemical element anomalies in the knowledge graph and obtain combinations of anomalies;

[0039] The analysis module is used to perform thermal matrix visualization analysis on outlier combinations to obtain the relationships between outlier combinations;

[0040] The delineation module is used to delineate the target mining area based on the combination of outliers.

[0041] The beneficial effects of the method and system of this invention are as follows: This invention constructs a geochemical knowledge graph by storing semi-structured geochemical data in a graph database, which has the ability to process high-dimensional nonlinear geochemical data quickly and efficiently. Then, the geochemical knowledge graph is visualized and analyzed, which can perform learning and multiple query reasoning without a large number of training samples, and explore the connections between geochemical data themselves. By delineating outliers in geochemical composite data through GIS, it can guide mineral prospect prediction, greatly reducing data analysis and field geological work. Attached Figure Description

[0042] Figure 1 This is a flowchart of the steps of a knowledge graph-based method for delineating mineral target areas according to the present invention.

[0043] Figure 2 This is a structural block diagram of a mineral target area delineation system based on knowledge graphs according to the present invention;

[0044] Figure 3 This is a data graph showing the results of the geochemical element thermodynamic matrix analysis of this invention;

[0045] Figure 4 This is a schematic diagram of the prospecting target area for the anomalous combination of geochemical elements Mo-W-Bi-Sn-Cu according to the present invention;

[0046] Figure 5 This is a schematic diagram of the prospecting target area for the anomalous combination of geochemical elements Au-Ag-Pb-Zn-Sn-Cu-Bi according to the present invention;

[0047] Figure 6 This is a schematic diagram of the prospecting target area for the anomalous combination of geochemical elements Au-Ag-As-Sb according to the present invention;

[0048] Figure 7 This is a schematic diagram showing the query results of the proportion of the lower limit of geochemical element anomalies in this invention. Detailed Implementation

[0049] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only for ease of explanation and do not limit the order of the steps. The execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art.

[0050] Reference Figure 1 This invention provides a method for delineating mineral target areas based on knowledge graphs, the method comprising the following steps:

[0051] S1. Acquire geochemical data and construct a geochemical knowledge graph;

[0052] S11. Obtain geochemical data;

[0053] Specifically, the data used in this invention are 1:50,000 stream sediment geochemical data from Pangxidong, Guangdong Province, from the "Mineral Prospect Survey in Pangxidong Area, Guangdong Province" project. A total of 7,234 stream sediment samples were collected in the field. The sample analysis was undertaken by the Geophysical and Geochemical Experimental Testing Institute of Jiangxi Provincial Bureau of Geology and Mineral Resources Exploration and Development. The 1:50,000 stream analysis was conducted between September 24 and December 28, 2010. The analysis items for this 1:50,000 stream sediment measurement sample included 16 elements: Cu, Pb, Zn, Mn, Sn, W, Ag, Au, As, Sb, Bi, Mo, Hg, F, Ba, and B. Among them, Au was chemical spectrum, B, Sn, and Ag were spectral quantitative, Cu was atomic absorption, Ba, Mn, Pb, and Zn were full-spectrum spectra, As, Sb, Bi, and Hg were atomic fluorescence, F was ion electrode, and Mo and W were catalytic polarography. The obtained geochemical data are shown in the table below.

[0054] Table 1 Geochemical element data

[0055] X Y Au B Sn Cu Ag Ba Mn Pb 422240 2418800 0.9 3 8.7 4 0.03 33 147 27 X Y Zn As Sb Bi Hg Mo W F 422240 2418800 26 1.17 0.31 0.23 0.04 2.67 0.79 212

[0056] S12. Analyze the geochemical data to obtain the abundance value and attribute of the geochemical data;

[0057] S13. Construct geochemical ternary datasets based on abundance values ​​and spatial attributes of geochemical data;

[0058] Specifically, since geochemical data is semi-structured data, it does not require knowledge extraction as is done for unstructured data. It only needs to convert the original geochemical data format into machine-readable triples. The original geochemical data consists of the geographic coordinates of the sampling points plus the abundance values ​​of 16 elements. The triple data format is <entity, relation, attribute>. This patent displays the original geochemical data in an Excel spreadsheet in the format <geographic coordinates, spatial fitting relation, element abundance>. The table below shows the converted geochemical triple data.

[0059] Table 2 Geochemical Triad Data

[0060] entity Entity type relation property Attribute type 397479,2424850 Geographical location Ag space fitting relationship 0.11 Ag enrichment type

[0061] S14. Build a database using Neo4j software;

[0062] Specifically, to build a geochemical map database, first create a local database in the Neo4j software, set the database name and password, and then run the database to complete the map database construction.

[0063] S15. Import the geochemical ternary set data into the database to obtain the geochemical knowledge graph.

[0064] Specifically, geochemical ternary data is imported in two ways. The first is to modify the data to the CREATE (entity)-[:relationship]->(attribute) format and enter this statement in the graph database constructed in the first step to import the data. The second is to write the application code of the <entity, relation, attribute> format data into the graph database. This patent uses the second method to construct the database and import geochemical data, which can construct a geochemical knowledge graph.

[0065] S2. Calculate the lower limit of geochemical element anomalies in the knowledge graph to obtain the anomaly combination;

[0066] S21. Calculate the mean and standard deviation of geochemical elements in the knowledge graph to obtain the original data values;

[0067] Specifically, the mean and standard deviation are initially calculated using all original samples. Then, the mean plus three times the standard deviation is calculated. The calculation process for the mean of the original samples is as follows:

[0068]

[0069] In the above formula, X represents the average value of the original geochemical element sample. i This represents the original geochemical element sample value, and n represents the number of original geochemical elements in the sample.

[0070] The calculation process for the mean squared error of the original sample is as follows:

[0071]

[0072] In the above formula, S1 represents the original mean squared error of the geochemical elements, n represents the original quantity of the geochemical elements, and X... i Represents the original sample values ​​of geochemical elements. This represents the average value of the original sample of geochemical elements.

[0073] S22. Analyze and process the original data values, remove data values ​​that are greater than a preset value from the original data values, and obtain the data values ​​after removal.

[0074] Specifically, it is often found that some original samples have values ​​higher than the mean plus three times the standard deviation, which does not meet the hypothesis, so these original samples with high values ​​are removed.

[0075] S23. Calculate the mean and standard deviation of the data values ​​after removal to obtain the lower limit of geochemical element anomalies.

[0076] Specifically, the mean and standard deviation are recalculated using the remaining samples, and then the new mean plus three times the standard deviation is calculated to obtain the lower limit of the anomaly for the elements in the main geological units. The calculation process for the mean is as follows:

[0077]

[0078] In the above formula, X represents the mean of the sample after removal, p represents the number of samples after removal, and X represents the mean of the sample after removal. k This represents the sample values ​​after removal;

[0079] The calculation process for its mean squared error is as follows:

[0080]

[0081] In the above formula, S2 represents the mean square error of the sample after removal;

[0082] in,

[0083] p = nm

[0084] X k =X i -X j

[0085] In the above formula, m represents the number of high outliers;

[0086] The calculated lower limit of the abnormality of the element is

[0087] S24. Use the Cypher statement in Neo4j software to query and process the lower limit values ​​of geochemical element anomalies to obtain the combination of anomalies.

[0088] Specifically, based on the lower limit of outliers, the Cypher statement is used in the geochemical knowledge graph to query nodes where each element's outlier exceeds its lower limit, and the number of nodes A is calculated. x ; Find nodes where one element is abnormal and the other element is also abnormal, and calculate the number of such nodes B. xy Then according to formula B xy / A x This yields the proportion of combined anomalies to single-element anomalies, where x and y represent element names, such as B. Pb-Zn / A Pb In Neo4j, the built-in Cypher query statement can be used to perform outlier queries, which can be used to query outliers of single elements and outliers of pairs of elements.

[0089] S3. Perform thermal matrix visualization analysis on outlier combinations to obtain the relationships between outlier combinations;

[0090] S31. Calculate the coefficient values ​​of geochemical elements by combining outliers using a thermodynamic matrix;

[0091] S32. Visualize the coefficient values ​​of geochemical elements to obtain the correlation coefficients of geochemical elements;

[0092] Specifically, refer to Figure 3 Heatmaps use color intensity to represent the distribution and correlation of data. Color intensity indicates the magnitude of the heat value, i.e., through heat matrix analysis, a correlation map of geochemical element heat matrices is obtained. Higher heat values ​​indicate stronger correlations between elements. The correlation between elements can be determined from the heat matrix map. This invention selects the top 5 correlated elements with heat values ​​greater than 0.2 for each element, and combines them in pairs to obtain 3 element combinations. This invention uses Python code for heat matrix analysis to read... Figure 3 The data is used to obtain a correlation map of the geochemical element thermodynamic matrix. A higher correlation between two elements indicates a higher thermodynamic value. The thermodynamic matrix map is obtained by calculating the correlation coefficient between each pair of elements and visualizing the levels of these coefficients, thus providing a direct understanding of the correlation between elements. The correlation coefficient is defined as the product of the covariance of two variables and their standard deviations. The calculation process is shown below:

[0093]

[0094] In the above formula, ρ represents the correlation coefficient value of geochemical elements, and X i X j σX represents a geochemical element. i ,X j This represents the covariance value of two geochemical elements. It represents the standard deviation between two geochemical elements.

[0095] S33. Determine the correlation coefficients of geochemical elements;

[0096] S34. If the correlation coefficient of the geochemical element is found to be greater than the preset coefficient value, combine the geochemical elements to obtain the outlier combination relationship.

[0097] Specifically, the results of the heatmap visualization were then statistically analyzed. The top 5 elements with the highest correlation (greater than 0.2) were selected as anomalous element combinations. Finally, based on the correlation of each element, it was determined which elements were correlated in pairs, resulting in three anomalous element combinations: Mo-W-Bi-Sn-Cu, Au-Ag-Pb-Zn-Sn-Cu-Bi, and Au-Ag-As-Sb. The data of the top 5 most correlated anomalous element combinations are shown in the table below.

[0098] Table 3 shows the abnormal combination data of the first 5 elements.

[0099]

[0100] S4. Delineate the target mining area based on the combination relationship of outliers.

[0101] S41. Visualize the interpolation of outlier combinations to obtain an outlier graph;

[0102] S42. Overlay the anomaly graphs to construct an anomaly graph combination;

[0103] S43. Based on the combination of anomaly maps, delineate the abnormal data to obtain the delineated mineral target area.

[0104] Specifically, refer to Figures 4 to 7 In the figure, anomalous data are selected in the anomalous combination map. Based on the elemental correlations mentioned above, three sets of anomalous elemental combinations were obtained: Mo-W-Bi-Sn-Cu, Au-Ag-Pb-Zn-Sn-Cu-Bi, and Au-Ag-As-Sb. Based on these anomalous elemental combinations, Cypher was used to query and export the anomalous data for each element in the graph database. The exported anomalous values ​​of each element were then visualized and interpolated in GIS. The anomalous maps of these combinations were then overlaid to form a prospecting target area. Specifically, the anomalous elements and the number of pairs of elements sharing a common anomalous value were first obtained from the map based on the lower limit of the anomalous value. After statistical analysis, the elemental correlations were obtained using a thermal coefficient matrix. Then, the elemental combinations were derived from these correlations, and prospecting areas were delineated based on these elemental combinations. Taking the first combination, Mo-W-Bi-Sn-Cu, as an example, data with the content of these five elements exceeding the lower limit of the anomalous value were selected from the map. The data for each element exceeding the lower limit were imported into ArcGIS for individual interpolation to obtain the values ​​of Mo, W, Bi, Sn, and Cu. These five anomalous area maps, because these elements are combined in anomalies, are overlaid with the above five layers. The overlapping anomalous areas are the prospecting target areas for this anomaly combination, which is... Figures 4 to 6 The boxed part in the middle.

[0105] Reference Figure 2 A knowledge graph-based mineral target area delineation system includes:

[0106] The module is used to acquire geochemical data and construct a geochemical knowledge graph;

[0107] The calculation module is used to calculate the lower limit of geochemical element anomalies in the knowledge graph and obtain combinations of anomalies;

[0108] The analysis module is used to perform thermal matrix visualization analysis on outlier combinations to obtain the relationships between outlier combinations;

[0109] The delineation module is used to delineate the target mining area based on the combination of outliers.

[0110] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0111] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for delineating mineral target areas based on knowledge graphs, characterized in that, Includes the following steps: The process of acquiring geochemical data and constructing a geochemical knowledge graph includes: acquiring geochemical data; analyzing the geochemical data to obtain its abundance values ​​and spatial attributes; constructing geochemical triplet data based on the abundance values ​​and spatial attributes of the geochemical data; building a database using Neo4j software; and importing the geochemical triplet data into the database to obtain the geochemical knowledge graph. The calculation of the lower limit values ​​of geochemical element anomalies in the knowledge graph to obtain anomaly combinations specifically includes: calculating the mean and standard deviation of geochemical elements in the knowledge graph to obtain raw data values; analyzing and processing the raw data values, removing data values ​​greater than a preset value to obtain the removed data values; calculating the mean and standard deviation of the removed data values, and using the sum of the mean and twice the standard deviation as the lower limit value of the geochemical element anomaly; querying the nodes of the geochemical element anomaly lower limit values ​​using the Cypher statement in Neo4j software; calculating the number of nodes of the geochemical element anomaly lower limit values, and then calculating the proportion of nodes of different geochemical element anomaly lower limit values ​​using a proportional formula to construct anomaly combinations; the formula for calculating the standard deviation of the geochemical elements is shown below: In the above formula, This represents the original mean squared error of geochemical elements. Indicates the original quantity of geochemical elements. Represents the original sample values ​​of geochemical elements. Represents the average value of the original sample of geochemical elements; A thermodynamic matrix visualization analysis is performed on outlier combinations to obtain outlier combination relationships. Specifically, this includes: calculating the coefficient values ​​of geochemical elements through a thermodynamic matrix; visualizing the coefficient values ​​of geochemical elements to obtain their correlation coefficients; judging the correlation coefficients of geochemical elements; and if the correlation coefficient of a geochemical element is determined to be greater than a preset coefficient value, combining the geochemical elements to obtain outlier combination relationships. Based on the relationships between outlier combinations, the target mining area is delineated.

2. The method for delineating mineral target areas based on knowledge graphs according to claim 1, characterized in that, The formula for calculating the thermodynamic matrix is ​​as follows: In the above formula, This represents the correlation coefficient values ​​of geochemical elements. , Represents geochemical elements, This represents the covariance value of two geochemical elements. It represents the standard deviation between two geochemical elements.

3. The method for delineating mineral target areas based on knowledge graphs according to claim 2, characterized in that, The step of delineating the target mining area based on outlier combinations specifically includes: Based on the combination relationships of outliers, visualization interpolation is performed in GIS to obtain an outlier map; Anomaly graphs are overlaid to construct anomaly graph assemblies; Based on the combination of anomaly maps, the abnormal data are delineated to obtain the delineated mineral target area.

4. A mineral target area delineation system based on knowledge graphs, characterized in that, Includes the following modules: The module is used to acquire geochemical data and construct a geochemical knowledge graph. Specifically, it includes: acquiring geochemical data; analyzing geochemical data to obtain its abundance values ​​and attributes; constructing geochemical triplet data based on the abundance values ​​and spatial attributes of the geochemical data; building a database using Neo4j software; and importing the geochemical triplet data into the database to obtain the geochemical knowledge graph. The calculation module is used to calculate the lower limit values ​​of geochemical element anomalies in the knowledge graph, obtaining anomaly combinations. Specifically, it includes: calculating the mean and standard deviation of geochemical elements in the knowledge graph to obtain raw data values; analyzing and processing the raw data values, removing data values ​​greater than a preset value to obtain the removed data values; calculating the mean and standard deviation of the removed data values, and using the sum of the mean and twice the standard deviation as the lower limit value of the geochemical element anomalies; querying the nodes of the geochemical element anomalies using the Cypher statement in Neo4j software based on the lower limit values; calculating the number of nodes based on the lower limit values ​​of the geochemical element anomalies, and then calculating the proportion of nodes for different geochemical element anomaly lower limit values ​​using a proportional formula to construct anomaly combinations; the formula for calculating the standard deviation of the geochemical elements is shown below: In the above formula, This represents the original mean squared error of geochemical elements. Indicates the original quantity of geochemical elements. Represents the original sample values ​​of geochemical elements. Represents the average value of the original sample of geochemical elements; The analysis module is used to perform thermodynamic matrix visualization analysis on outlier combinations to obtain outlier combination relationships. Specifically, it includes: calculating the coefficient values ​​of geochemical elements through a thermodynamic matrix; visualizing the coefficient values ​​of geochemical elements to obtain their correlation coefficients; judging the correlation coefficients of geochemical elements; and if the correlation coefficient of a geochemical element is greater than a preset coefficient value, combining the geochemical elements to obtain outlier combination relationships. The delineation module is used to delineate the target mining area based on the combination of outliers.

Images