A method and system for evaluating metallogenic potential based on multi-element combination correlation analysis

Abstract

This invention discloses a method and system for evaluating mineralization potential based on multi-element combination correlation analysis. In one specific embodiment, the method includes: collecting and preprocessing data from multiple sampling points in a target area to obtain standardized concentration data, normalized contrast values, and anomaly Boolean matrices for each element; identifying elemental combinations for multiple single mineral types based on the standardized concentration data and anomaly Boolean matrices of all sampling points, and determining the optimal element combination and weight coefficient for each single mineral type; calculating the mineralization potential value for each single mineral type corresponding to each sampling point based on the optimal element combination and weight coefficients for each single mineral type at each sampling point and the normalized contrast values ​​for each element at each sampling point; calculating the total mineralization potential value for each sampling point based on the total mineralization potential value for each single mineral type at each sampling point; and selecting the area formed by sampling points whose total mineralization potential value is greater than a first preset value as the target area.

Description

Technical Field

[0001] This invention relates to the field of geological exploration and mineral resource evaluation technology. More specifically, it relates to a method and system for evaluating mineralization potential based on multi-element combination correlation analysis. Background Technology

[0002] In related technologies, methods based on single-element anomalies or fixed element ratios have revealed shortcomings such as poor adaptability, insufficient information utilization, and lack of comprehensive evaluation when facing the problem of mineralization prediction in the context of multiple mineral types and complex geochemical backgrounds. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for evaluating mineralization potential based on multi-element combination correlation analysis, so as to solve at least one of the problems existing in the prior art.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of this invention provides a method for evaluating mineralization potential based on multi-element combination correlation analysis, comprising: Data collection and preprocessing were performed on multiple sampling points in the target area to obtain standardized concentration data, normalized contrast values, and outlier Boolean matrices for each element. Based on the standardized concentration data and abnormal Boolean matrix of all sampling points, element combination identification is performed for multiple single mineral types to determine the optimal element combination and weight coefficient for each single mineral type. The mineralization potential value of each sampling point is calculated based on the optimal element combination and weight coefficient of each single mineral type at each sampling point, as well as the normalized contrast value of each element at each sampling point. The total mineralization potential value of each sampling point is calculated based on the mineralization potential value of each single mineral at each sampling point. The area formed by each sampling point whose total mineralization potential value is greater than the first preset value is selected as the target area.

[0005] Optionally, the step of collecting and preprocessing data from multiple sampling points in the target area to obtain standardized concentration data, normalized contrast values, and outlier Boolean matrices for each element includes: Data was collected from multiple sampling points in the target area to obtain the concentration data of each element at each sampling point; Determine whether there are missing element concentration data at each sampling point. If so, use the median filling method to fill in the missing element concentration values ​​to obtain the completed element concentration data. The completed element concentration data is transformed to obtain the transformed element concentration data. The median absolute deviation method was used to calculate the regional background value and the lower limit of anomalies for each element on the transformed data scale of the element concentration data. The abnormal Boolean matrix is ​​obtained by binarizing the transformed concentration data of each element based on the abnormal lower limit value of each element. The contrast value of each element is calculated based on the transformed concentration data of each element and the regional background value of each element. The contrast values ​​of each element are normalized to obtain the normalized contrast values ​​of each element; The transformed element concentration data are standardized to obtain the standardized concentration data of each element.

[0006] Optionally, the step of performing data transformation on the completed element concentration data to obtain the transformed element concentration data includes: Logarithmic transformation or Box-Cox transformation is performed on the completed element concentration data to obtain the transformed element concentration data.

[0007] Optionally, the step of identifying elemental combinations of multiple single minerals based on standardized concentration data and outlier Boolean matrices from all sampling points, and determining the optimal elemental combination and weighting coefficients for each single mineral, includes: Correlation analysis was performed on the standardized concentration data of each element at each sampling point to obtain the first set of elements with the first correlation to each single mineral type. Principal component analysis is used to select elements with first loadings on principal components whose cumulative variance contribution rate is greater than a second preset value for each element's standardized concentration data at each sampling point, thus obtaining a second set of elements. The second set of elements is then merged with the first set of elements and deduplicated to obtain a candidate set of elements. The machine learning algorithm is used to train the candidate element set with the element concentration as the feature and the target column of the single mineral type as the label to obtain the importance score of each element in the candidate element set. The importance score of each element in the candidate element set is then normalized to obtain the initial weight coefficient of each element in the candidate element set. Based on the anomaly Boolean matrix, the association rule mining algorithm is used to select frequently co-occurring element combinations from the candidate element set to obtain the third element combination set, and the support of each element combination in the third element combination set is calculated; wherein, the minimum support of the frequently co-occurring element combination is greater than or equal to 0.03 and less than or equal to 0.5. The sum of the initial weight coefficients of each element in each frequently co-occurring element combination in the third element combination set is multiplied by the support of that element combination to obtain the comprehensive score of that element combination. Select the element combination with the highest comprehensive score from the third element combination set as the initial optimal element combination for a single mineral type; Select the third preset number of elements from the initial optimal element combination as the optimal element combination, and normalize the initial weight coefficients of each element in the optimal element combination to obtain the weight coefficients.

[0008] Optionally, the first set of elements with a first correlation to each single mineral type obtained by performing correlation analysis on the standardized concentration data of each element at each sampling point includes: Correlation analysis using Pearson correlation coefficient or Spearman rank correlation coefficient was performed on the standardized concentration data of each element at the sampling points to obtain the first set of elements with the first correlation to each individual mineral.

[0009] Optionally, the step of selecting frequently co-occurring element combinations from the candidate element set using an association rule mining algorithm based on the anomaly Boolean matrix to obtain a third element combination set and calculating the support of each element combination in the third element combination set includes: Based on the anomalous Boolean matrix, the Apriori algorithm or FP-Growth algorithm is used to select frequently co-occurring element combinations from the candidate element set to obtain the third element combination set, and the support of each element combination in the third element combination set is calculated.

[0010] Optionally, the step of calculating the mineralization potential value of each sampling point corresponding to each single mineral type based on the optimal element combination and weighting coefficient of each single mineral type at each sampling point and the normalized contrast value of each element at each sampling point includes:

[0011] In the formula, For a single mineral The mineralization potential value; This represents the number of elements in the optimal combination. For a single mineral In the optimal combination of elements, the first The weight coefficient of each element; For a single mineral In the optimal combination of elements, the first The normalized contrast values ​​of each element; where... By analyzing a single mineral In the optimal combination of elements, the first Contrast value of each element Normalization is performed to obtain the result.

[0012] Optionally, the single mineral type In the optimal combination of elements, the first The formula for calculating the contrast value of an element is:

[0013] In the formula, For a single mineral In the optimal combination of elements, the first The transformed concentration data of each element; For the first The background value of the area for each element.

[0014] Optionally, the step of calculating the total mineralization potential value of each sampling point based on the mineralization potential value of each single mineral at each sampling point includes:

[0015] In the formula, This represents the total mineralization potential value. This is the first adjustment coefficient; This is the second adjustment coefficient; This is the third adjustment coefficient, where, ; It is a function for maximizing the value; For a single mineral The mineralization potential value; For a single mineral The mineralization potential value; For a single mineral The mineralization potential value; For a single mineral The mineralization potential value; The quantity of a single mineral.

[0016] A second aspect of the present invention provides a mineralization potential evaluation system based on multi-element combination correlation analysis, comprising: The data acquisition and preprocessing module is used to acquire and preprocess data from multiple sampling points in the target area to obtain standardized concentration data, normalized contrast values, and outlier Boolean matrices for each element. The element combination identification module is used to identify the element combination of multiple single minerals based on the standardized concentration data and abnormal Boolean matrix of all sampling points, and to determine the optimal element combination and weight coefficient for each single mineral. The single mineral potential calculation module is used to calculate the mineralization potential value of each sampling point corresponding to each single mineral based on the optimal element combination and weight coefficient of each single mineral and the normalized contrast value of each element at each sampling point. The total potential synthesis module is used to calculate the total mineralization potential value of each sampling point based on the mineralization potential value of each individual mineral at each sampling point. The target area delineation module is used to select the area formed by each sampling point whose total mineral potential value is greater than the first preset value as the target area.

[0017] The beneficial effects of this invention are as follows: The technical solution described in this invention effectively solves the problem of related technologies failing to effectively utilize multi-element synergistic information, overcoming their limitations of relying on a single element or fixed element ratios, and enabling the automatic identification of the element combinations most relevant to specific mineral types (such as copper and molybdenum) from multi-element geochemical data; it effectively solves the problem of related technologies lacking a systematic quantitative evaluation system, overcoming the subjective and singular defects of their prediction results, and establishing a quantitative mineralization potential evaluation model for each specific mineral type; it effectively solves the problem of related technologies' insufficient comprehensive prediction capabilities for multiple mineral types, overcoming their limitations of only targeting a single mineral type, and achieving the integration of potential information from multiple mineral types to generate a comprehensive total mineralization potential; and it effectively overcomes the defects of low prediction accuracy and universality of related technologies. Through the technical solution described in this invention, the accuracy and reliability of mineralization prediction under different geological backgrounds are significantly improved. Attached Figure Description

[0018] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0019] Figure 1 This diagram shows the overall flowchart of the mineralization potential evaluation method based on multi-element combination correlation analysis provided by an embodiment of the present invention.

[0020] Figure 2 This paper presents a detailed flowchart of the mineralization potential evaluation method based on multi-element combination correlation analysis provided by an embodiment of the present invention.

[0021] Figure 3 This diagram illustrates the data processing and preprocessing flow of the mineralization potential evaluation method based on multi-element combination correlation analysis provided in an embodiment of the present invention.

[0022] Figure 4 This diagram illustrates a characteristic element combination identification method for a mineralization potential evaluation method based on multi-element combination correlation analysis, as provided in an embodiment of the present invention.

[0023] Figure 5 This diagram illustrates a multi-mineral potential synthesis model based on a multi-element combination correlation analysis-based mineralization potential evaluation method provided in an embodiment of the present invention.

[0024] Figure 6 This paper presents a heat map showing the distribution of copper ore mineralization potential using a method for evaluating ore mineralization potential based on multi-element combination correlation analysis, as provided in an embodiment of the present invention.

[0025] Figure 7 This paper presents a heat map showing the distribution of tungsten ore mineralization potential using a method for evaluating mineralization potential based on multi-element combination correlation analysis, as provided in an embodiment of the present invention.

[0026] Figure 8This paper presents a heat map showing the distribution of cobalt mineralization potential using a method for evaluating mineralization potential based on multi-element combination correlation analysis, as provided in an embodiment of the present invention.

[0027] Figure 9 This paper presents a heat map showing the total mineralization potential distribution of the mineralization potential evaluation method based on multi-element combination correlation analysis provided in an embodiment of the present invention. Detailed Implementation

[0028] To more clearly illustrate the present invention, the following description, in conjunction with embodiments and accompanying drawings, further explains the invention. Similar components in the drawings are indicated by the same reference numerals. Those skilled in the art should understand that the specific description below is illustrative rather than restrictive and should not be construed as limiting the scope of protection of the present invention.

[0029] Currently, most existing geochemical prospecting methods rely on single-element anomalies or simple elemental ratios for mineralization prediction. For example, a method and system for predicting pegmatite-type lithium deposits is disclosed in related technologies by calculating the ratio of lithium (Li) to lanthanum (La) (Li / La) and using the anomaly regions of this ratio. The core of this method is to use a fixed element pair (Li / La) ratio as a prospecting indicator. However, this related technology has the following shortcomings: 1. Inability to effectively identify complex synergistic relationships among multiple elements: The methods of this related technology rely only on Li and La, failing to fully utilize the comprehensive geochemical data obtained during exploration. In actual mineralization processes, mineralization is often closely related to complex geochemical behaviors such as the coexistence and differentiation of multiple elements, and relying solely on a fixed pair of element ratios is insufficient to comprehensively and profoundly reflect the mineralization regularity.

[0030] 2. Limited ability to identify characteristic element combinations of different mineral types: The method of this related technology is designed for the specific mineral type of "pegsite lithium deposit". Its fixed Li / La ratio index is difficult to be directly transferred and effectively applied to the prediction of other mineral types (such as copper deposits, molybdenum deposits, etc.), and lacks universality and adaptability.

[0031] 3. Lack of a systematic quantitative comprehensive evaluation system: The methods of this related technology are essentially still based on a single indicator (albeit a ratio indicator) for anomaly delineation. It has failed to construct a comprehensive evaluation model that can integrate information from multiple elements and combinations and output quantitative "mineralization potential", resulting in relatively simple prediction results and a lack of hierarchy.

[0032] 4. Prediction results are significantly influenced by the subjective nature of the preset model: The prediction effect of this related technology is highly dependent on the empirically-based model of the "Li / La ratio". For mining areas with different geological backgrounds, the optimal combination of indicator elements may change, and the fixed ratio model is difficult to adaptively adjust, resulting in low accuracy or poor applicability when predicting new areas.

[0033] In summary, existing methods based on single-element anomalies or fixed element ratios exhibit shortcomings such as poor adaptability, insufficient information utilization, and incomplete evaluation when facing the problem of mineralization prediction in complex geochemical contexts with multiple mineral types.

[0034] In view of this, such as Figure 1 As shown, one embodiment of the present invention provides a method for evaluating mineralization potential based on multi-element combination correlation analysis, including: collecting and preprocessing data from multiple sampling points in a target area to obtain standardized concentration data, normalized contrast values, and anomaly Boolean matrices for each element; identifying elemental combinations for multiple single mineral types based on the standardized concentration data and anomaly Boolean matrices of all sampling points, and determining the optimal elemental combination and weight coefficient for each single mineral type; calculating the mineralization potential value for each single mineral type corresponding to each sampling point based on the optimal elemental combination and weight coefficient for each single mineral type at each sampling point and the normalized contrast values ​​for each element at each sampling point; calculating the total mineralization potential value for each sampling point based on the total mineralization potential value for each single mineral type at each sampling point; and selecting the area formed by sampling points whose total mineralization potential value is greater than a first preset value as the target area.

[0035] In a specific example, a comprehensive evaluation method for mineralization potential based on multi-element combination correlation analysis is provided. The method's flowchart can be found here. Figure 2 Its core lies in transforming raw geochemical data into single mineral types and comprehensive total mineralization potential through a systematic process.

[0036] The technical solution described in this invention effectively solves the problem of related technologies failing to effectively utilize multi-element synergistic information, overcoming their limitations of relying on a single element or fixed element ratios, and enabling the automatic identification of the element combinations most relevant to specific mineral types (such as copper and molybdenum) from multi-element geochemical data; it effectively solves the problem of related technologies lacking a systematic quantitative evaluation system, overcoming the subjective and singular defects of their prediction results, and establishing a quantitative mineralization potential evaluation model for each specific mineral type; it effectively solves the problem of related technologies' insufficient comprehensive prediction capabilities for multiple mineral types, overcoming their limitations of only targeting a single mineral type, and achieving the integration of potential information from multiple mineral types to generate a comprehensive total mineralization potential; and it effectively overcomes the defects of low prediction accuracy and universality of related technologies. Through the technical solution described in this invention, the accuracy and reliability of mineralization prediction under different geological backgrounds are significantly improved.

[0037] This invention is also used for environmental geochemical assessment: identifying combinations of polluting elements and calculating regional pollution risk potential; geological disaster risk assessment: analyzing combinations of disaster-related geochemical indicators to achieve quantitative prediction of risk areas; and agricultural geological assessment: identifying combinations of soil fertility-related elements to provide a basis for precision agriculture.

[0038] In one possible implementation, the step of collecting and preprocessing data from multiple sampling points in the target area to obtain standardized concentration data, normalized contrast values, and anomaly Boolean matrices for each element includes: collecting data from multiple sampling points in the target area to obtain element concentration data for each sampling point; determining whether there are missing values ​​in the element concentration data for each sampling point, and if so, using the median imputation method to fill in the missing element concentration values ​​to obtain the filled element concentration data; performing data transformation on the filled element concentration data to obtain transformed element concentration data; calculating the regional background value and the lower limit of anomalies for each element on the data scale of the transformed element concentration data using the median absolute deviation method; performing binarization transformation on the transformed element concentration data based on the lower limit of anomalies for each element to obtain the anomaly Boolean matrix; calculating the contrast value for each element based on the transformed element concentration data and the regional background value for each element; normalizing the contrast value for each element to obtain the normalized contrast value for each element; and standardizing the transformed element concentration data to obtain standardized element concentration data.

[0039] In a specific example Figure 3 This demonstrates the entire data process from acquisition to standardization. The purpose of this step is to obtain high-quality, standardized basic geochemical data, providing reliable input for subsequent analysis.

[0040] In a specific example, data acquisition includes collecting geochemical data (concentration values ​​of multiple elements) for the target area. The sampling point density for the target area is 5 to 10 points per square kilometer, a range that strikes the optimal balance between exploration costs and data representativeness.

[0041] Specifically, the data preparation phase includes: firstly, collecting geochemical data of the target area. The dataset contains 9020 sampling points, each containing East and North coordinates, as well as concentration data (in ppm) of 39 elements including Cu, Mo, W, Co, Ag, Pb, and Zn.

[0042] This embodiment forms a fully automated pipeline that greatly improves work efficiency. From data reading, preprocessing, combined recognition to result visualization, it can be completed with one click. It only takes 1 minute to process a dataset containing more than 9,000 sampling points and 35 elements. Compared with the 1-3 working days required for traditional manual interpretation, the efficiency is improved by more than 95%, while completely avoiding the random errors that may occur in manual operation.

[0043] This embodiment enhances the adaptability of the method under different geological backgrounds. It uses robust statistical methods such as median and MAD to calculate background values ​​and anomaly lower limits, effectively suppressing the interference of extreme high values. In applications in multiple different landscape areas, the automatically calculated element anomaly lower limits have an error of less than 15% compared with manually determined thresholds.

[0044] In one possible implementation, the step of performing data transformation on the completed element concentration data to obtain transformed element concentration data includes: performing a logarithmic transformation or a Box-Cox transformation on the completed element concentration data to obtain transformed element concentration data.

[0045] In a specific example, data preprocessing includes: first, checking and imputing missing values ​​in the raw data; then, to eliminate the influence of dimensional differences and skewed distributions in element concentration values, data transformation is performed on the element concentration data. Feasible transformation methods include logarithmic transformation or Box-Cox transformation, where the parameters of the Box-Cox transformation are... The feasible range is [-2, 2], and its optimal value is 0 (i.e., logarithmic transformation). Finally, the robust statistical method of median absolute deviation (MAD) is used to calculate the regional background value and the lower limit of anomalies for each element.

[0046] Specifically, data preprocessing includes: first, checking the sampled data for missing values ​​and using median imputation to fill in the missing element concentration values; then, performing a logarithmic transformation on the imputed concentration data to make its distribution approach a normal distribution; next, using the Median Absolute Deviation (MAD) robust statistical method, calculating the regional background value and lower limit of anomalies for each element on the transformed data scale, for example, Cu has a background value of 2.38 and an anomaly lower limit of 5.03, W has a background value of 0.80 and an anomaly lower limit of 2.32, and Co has a background value of 1.83 and an anomaly lower limit of 4.40. The resulting data is the transformed data; the transformed element concentration values ​​are converted to Boolean values ​​(e.g., determining whether they are greater than the anomaly lower limit, True if greater, False if less), obtaining an anomaly Boolean matrix for subsequent association rule mining modules. Additionally, the logarithmically transformed data needs to be standardized (i.e., subtracting the mean and dividing by the standard deviation) for input into the PCA analysis module. The element contrast value for each sampling point is calculated, which is the ratio of element concentration to the corresponding background value. This is used to quantitatively characterize the enrichment degree of elements relative to the background. The element contrast value is normalized using the Min-Max normalization method for calculating mineralization favorability.

[0047] In one possible implementation, the step of identifying element combinations for multiple single mineral types based on standardized concentration data and anomaly Boolean matrices from all sampling points, and determining the optimal element combination and weight coefficients for each single mineral type, includes: performing correlation analysis on the standardized concentration data of each element at each sampling point to obtain a first set of elements with a first correlation to each single mineral type; using principal component analysis to select elements with a first loading on the principal components whose cumulative variance contribution rate is greater than a second preset value from the standardized concentration data of each element at each sampling point to obtain a second set of elements, and merging the second set of elements with the first set of elements and removing duplicates to obtain a candidate set of elements; using a machine learning algorithm with the element concentrations in the candidate set as features and the target column of the single mineral type as labels for training, obtaining the importance scores of each element in the candidate set of elements, and assigning importance scores to each element in the candidate set of elements. The initial weight coefficients of each element in the candidate element set are obtained by score normalization. Based on the anomaly Boolean matrix, an association rule mining algorithm is used to select frequently co-occurring element combinations from the candidate element set to obtain a third element combination set, and the support of each element combination in the third element combination set is calculated. The minimum support of frequently co-occurring element combinations ranges from greater than or equal to 0.03 to less than or equal to 0.5. The product of the sum of the initial weight coefficients of each element in each frequently co-occurring element combination in the third element combination set and the support of that element combination is calculated as the comprehensive score of that element combination. The element combination with the highest comprehensive score is selected from the third element combination set as the initial optimal element combination for a single mineral type. From the initial optimal element combination, elements with a third preset number of elements are selected as the optimal element combination, and the initial weight coefficients of each element in the optimal element combination are normalized to obtain the weight coefficients.

[0048] In a specific example, append Figure 4 This invention demonstrates a multi-algorithm fusion process for identifying multiple mineral feature element combinations. The purpose of multi-mineral feature element combination identification is to automatically and objectively identify the key element combinations and their weights most relevant to each specific target mineral, thereby overcoming the limitations of existing technologies that rely on fixed element pairs.

[0049] This embodiment significantly improves the objectivity and effectiveness of element combination identification. By integrating correlation analysis, principal component analysis, association rule mining, and random forest feature importance assessment, it automatically filters key element combinations related to specific mineral types from multiple dimensions, effectively overcoming the subjective bias problem caused by the reliance on human experience in traditional methods.

[0050] In one possible implementation, the step of performing correlation analysis on the standardized concentration data of each element at each sampling point to obtain a first set of elements with a first correlation to each single mineral type includes: performing correlation analysis on the standardized concentration data of each element at the sampling points using Pearson correlation coefficient or Spearman rank correlation coefficient to obtain a first set of elements with a first correlation to each single mineral type.

[0051] In a specific example, candidate element screening includes: for each target mineral type, firstly, performing correlation analysis using Pearson correlation coefficient or Spearman rank correlation coefficient to screen for elements with the highest spatial correlation to known mineralization points; simultaneously, principal component analysis (PCA) can be used to select elements with high loadings on the first few principal components (typically with a cumulative variance contribution rate >70%). This process will form a candidate element set.

[0052] In one possible implementation, the step of selecting frequently co-occurring element combinations from the candidate element set using an association rule mining algorithm based on the anomalous Boolean matrix to obtain a third element combination set and calculating the support of each element combination in the third element combination set includes: selecting frequently co-occurring element combinations from the candidate element set based on the anomalous Boolean matrix and using the Apriori algorithm or FP-Growth algorithm to obtain a third element combination set and calculating the support of each element combination in the third element combination set.

[0053] In a specific example, element combination mining and evaluation includes: applying the Apriori association rule mining algorithm to discover frequently co-occurring element combinations from the candidate element set. The criterion for frequent co-occurrence is defined by a minimum support threshold, with a feasible range of [0.05, 0.3] and an optimal value of 0.1. Subsequently, a random forest machine learning model is used, trained with all element concentrations as features and the mineral target column as labels; the overall importance score of each element combination is quantitatively evaluated through the feature importance output by the model.

[0054] In a specific example, the FP-Growth algorithm can be used instead of the Apriori algorithm for association rule mining to improve computational efficiency.

[0055] In a specific example, feature evaluation can use gradient boosting trees or deep learning autoencoders instead of random forests.

[0056] In a specific example, determining the optimal combination and weights involves selecting an optimal combination of elements (typically containing 2 to 5 elements) for each target mineral type based on the importance scores evaluated by the machine learning model. The weight coefficients of each element within the combination are determined by normalizing its feature importance score in the machine learning model.

[0057] In a specific example, element combination identification involves multiple analytical steps, including correlation analysis, PCA analysis, association rule mining, and model importance assessment, ultimately determining the optimal element combination for copper ore as {Cu, Mo, Ag, S, Bi}, with weighting coefficients of {0.77, 0.08, 0.07, 0.05, 0.03}. Similarly, the optimal element combination for tungsten ore is determined to be {W, Se, Mo, Tl, Rb}, with weighting coefficients of {0.69, 0.14, 0.1, 0.04, 0.03}. The optimal element combination for cobalt ore is {S, Cu, Co, Ag, Sb}, with weighting coefficients of {0.33, 0.24, 0.21, 0.11, 0.11}.

[0058] This embodiment employs a multi-algorithm fusion process, sequentially using correlation analysis to screen candidate elements, principal component analysis to extract major element combinations, association rule mining to discover frequently co-occurring itemsets, and finally combining a random forest model to evaluate the importance of each combination, thereby achieving automatic identification of element combinations for different mineral types. Furthermore, it dynamically determines the optimal element combination and weight coefficients for each mineral type by weighted fusion of correlation scores, PCA loadings, association rule support, and model feature importance, forming a complete mineral-specific combination optimization selection scheme.

[0059] In one possible implementation, the calculation of the mineralization potential value of each sampling point corresponding to each single mineral type based on the optimal element combination and weighting coefficient of each single mineral type at each sampling point and the normalized contrast value of each element at each sampling point includes:

[0060] In the formula, For a single mineral The mineralization potential value; This represents the number of elements in the optimal combination. For a single mineral In the optimal combination of elements, the first The weight coefficients of each element satisfy the following condition: ; For a single mineral In the optimal combination of elements, the first The normalized contrast values ​​of each element; where... By analyzing a single mineral In the optimal combination of elements, the first Contrast value of each element Normalization is performed to obtain the result.

[0061] In a specific example, the role of calculating the mineralization potential of a single mineral type is to establish a quantitative evaluation model for each target mineral type, transforming elemental concentration information into continuous mineralization potential values, thus overcoming the shortcomings of existing technologies that only perform binary (abnormal or normal) determinations.

[0062] In a specific example, the mineralization potential of any target mineral is calculated using the above formula. This is a transformation function for elemental concentrations, used to map concentration values ​​to potential contributions. This function can be a linear function, a logarithmic function, or a cumulative distribution function. This example uses its contrast value relative to the background value.

[0063] In one possible implementation, the single mineral type In the optimal combination of elements, the first The formula for calculating the contrast value of an element is:

[0064] In the formula, For a single mineral In the optimal combination of elements, the first The transformed concentration data of each element; For the first The background value of the area for each element.

[0065] In a specific example, the copper ore potential F_Cu is calculated as: F_Cu = 0.77 × F(Cu) + 0.08 × F(Mo) + 0.07 × F(Ag) + 0.05 × F(S) + 0.03 × F(Bi). The contrast values ​​for each element are calculated as follows: f(Cu) = Cu concentration / 2.38, f(Mo) = Mo concentration / 0.30, f(Ag) = Ag concentration / 0.06, f(S) = S concentration / 5.24, f(Bi) = Bi concentration / 0.16. The contrast values ​​for each element are then normalized, and the final mineralization potential distribution is as follows: Figure 6 As shown. Among them, Figures 6 to 9 The units for the horizontal and vertical axes are meters, Favorability is the mineralization potential value, scatter is a scatter plot, East is the eastward coordinate, and North is the northward coordinate.

[0066] In a specific example, the tungsten ore potential F_W is calculated as: F_W = 0.69 × F(W) + 0.14 × F(Se) + 0.1 × F(Mo) + 0.04 × F(Tl) + 0.03 × F(Rb). The contrast values ​​for each element are calculated as follows: f(W) = W concentration / 0.80, f(Se) = Se concentration / 0.06, f(Mo) = Mo concentration / 0.3, f(Tl) ​​= Tl concentration / 0.49, f(Rb) = Rb concentration / 4.23. The contrast values ​​for each element are then normalized, resulting in the final mineralization potential distribution as shown below. Figure 7 As shown.

[0067] In a specific example, the cobalt ore potential F_Co is calculated as: F_Co = 0.33 × F(S) + 0.24 × F(Cu) + 0.21 × F(Co) + 0.11 × F(Ag) + 0.11 × F(Sb). The contrast values ​​for each element are calculated as follows: f(S) = S concentration / 5.24, f(Cu) = Cu concentration / 2.38, f(Co) = Co concentration / 1.83, f(Ag) = Ag concentration / 0.06, f(Sb) = Sb concentration / 0.56. The contrast values ​​for each element are then normalized, resulting in the final mineralization potential distribution as shown below. Figure 8 As shown.

[0068] This embodiment constructs a quantitative calculation formula based on element contrast value and combined weights. The element concentration conversion function can be the contrast value, logarithmic transformation or cumulative distribution function (CDF). The weights are derived from the optimization selection strategy. The results are normalized to the range [0,1] using the Min-Max normalization method based on robust quantiles (1%-99%), thus realizing a continuous and quantitative evaluation of mineralization potential.

[0069] In one possible implementation, calculating the total mineralization potential value of each sampling point based on the mineralization potential value of each individual mineral at each sampling point includes:

[0070] In the formula, This represents the total mineralization potential value. This is the first adjustment coefficient; This is the second adjustment coefficient; This is the third adjustment coefficient, where, ; It is a function for maximizing the value; For a single mineral The mineralization potential value; For a single mineral The mineralization potential value; For a single mineral The mineralization potential value; For a single mineral The mineralization potential value; The quantity of a single mineral.

[0071] In a specific example, append Figure 5 This is a schematic diagram of a multi-mineral potential synthesis model. The purpose of this step is to integrate the potential information of multiple individual minerals to generate a comprehensive evaluation index, in order to solve the problem that existing technologies are unable to perform multi-mineral collaborative prediction.

[0072] In a specific example, the total mineral potential is generated by combining the potential of all multiple target mineral types, and the calculation formula is as shown above. α, β, γThese are adjustment coefficients used to balance the contributions of the main mineralization process (max term), the average mineralization potential (mean term), and the mineralization intensity (root mean square term). The feasible values ​​for these coefficients are all within the range [0,1] and satisfy the constraints. α + β + γ =1. Set in the embodiment. α =0.5, β =0.3, γ =0.2, to highlight the main mineralization role while taking into account the comprehensive potential.

[0073] In a specific example, multi-mineral synthesis includes: employing =0.5, =0.3, With a weighting coefficient of 0.2, the total mineralization potential is calculated as F_total = 0.5 × max(F_Cu, F_W, F_Co) + 0.3 × mean(F_Cu, F_W, F_Co) + 0.2 × RMS(F_Cu, F_W, F_Co). The final heatmap of the total mineralization potential distribution is attached. Figure 9 As shown.

[0074] This embodiment designs a multi-parameter synthesis formula that integrates the maximum value term, the average value term, and the root mean square term, by adjusting the coefficients. α, β, γ (For example, satisfying) α + β + γ =1) to balance the contributions of major mineral types, average mineralization potential and overall mineralization intensity, and realize the effective integration of multi-mineral potential information and the quantitative characterization of comprehensive mineralization potential.

[0075] This embodiment possesses efficient multi-mineral collaborative prediction capabilities. By designing a comprehensive synthesis formula that includes maximum, average, and root mean square terms, it can balance the contributions of major and associated minerals, and simultaneously complete the potential evaluation of multiple minerals in a single analysis. This improves the efficiency of comprehensive analysis and avoids the information fragmentation problem caused by traditional mineral-specific prediction.

[0076] In a specific example, by comparing the calculation results with known geological data, two anomalous concentration areas with continuous area and regular shape were identified in the high potential area with a total mineralization potential greater than 0.6. These areas can be used as preferred target areas for the next exploration work.

[0077] Specifically, the calculation results are compared with known geological data, such as... Figure 9 As shown, in the high-potential area with a total mineralization potential greater than 0.6, two main areas of concentrated anomalies were identified, located in... Figure 9 The eastern and western locations shown can serve as preferred target areas for the next stage of exploration.

[0078] This embodiment achieves accurate quantitative evaluation of mineralization potential. By establishing a quantitative calculation model based on element contrast values ​​and combined weights, the mineralization potential is transformed into a continuous potential value between 0 and 1. In the verification experiment of known mining areas, the high potential area (Favorability>0.6) delineated by this invention has a correlation of more than 80% with the known mineralization area, which improves the prediction accuracy compared with the traditional single-element anomaly threshold method.

[0079] Another embodiment of the present invention provides a mineralization potential evaluation system based on multi-element combination correlation analysis, comprising: a data acquisition and preprocessing module for acquiring and preprocessing data from multiple sampling points in a target area to obtain standardized concentration data, normalized contrast values, and anomaly Boolean matrices for each element; an element combination identification module for identifying element combinations of multiple single mineral types based on the standardized concentration data and anomaly Boolean matrices of all sampling points, and determining the optimal element combination and weight coefficient for each single mineral type; a single mineral type potential calculation module for calculating the mineralization potential value of each sampling point corresponding to each single mineral type based on the optimal element combination and weight coefficient for each single mineral type and the normalized contrast values ​​of each element at each sampling point; a total potential synthesis module for calculating the total mineralization potential value of each sampling point based on the mineralization potential value of each single mineral type at each sampling point; and a target area delineation module for selecting the area formed by sampling points with a total mineralization potential value greater than a first preset value as the target target area.

[0080] The technical solution described in this invention effectively solves the problem of related technologies failing to effectively utilize multi-element synergistic information, overcoming their limitations of relying on a single element or fixed element ratios, and enabling the automatic identification of the element combinations most relevant to specific mineral types (such as copper and molybdenum) from multi-element geochemical data; it effectively solves the problem of related technologies lacking a systematic quantitative evaluation system, overcoming the subjective and singular defects of their prediction results, and establishing a quantitative mineralization potential evaluation model for each specific mineral type; it effectively solves the problem of related technologies' insufficient comprehensive prediction capabilities for multiple mineral types, overcoming their limitations of only targeting a single mineral type, and achieving the integration of potential information from multiple mineral types to generate a comprehensive total mineralization potential; and it effectively overcomes the defects of low prediction accuracy and universality of related technologies. Through the technical solution described in this invention, the accuracy and reliability of mineralization prediction under different geological backgrounds are significantly improved.

[0081] This invention is also used for environmental geochemical assessment: identifying combinations of polluting elements and calculating regional pollution risk potential; geological disaster risk assessment: analyzing combinations of disaster-related geochemical indicators to achieve quantitative prediction of risk areas; and agricultural geological assessment: identifying combinations of soil fertility-related elements to provide a basis for precision agriculture.

[0082] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. For those skilled in the art, other variations or modifications can be made based on the above description. It is impossible to exhaustively list all the implementation methods here. All obvious variations or modifications derived from the technical solutions of the present invention are still within the protection scope of the present invention.

Claims

1. A method for evaluating mineralization potential based on multi-element combination correlation analysis, characterized in that, include: Data collection and preprocessing were performed on multiple sampling points in the target area to obtain standardized concentration data, normalized contrast values, and outlier Boolean matrices for each element. Based on the standardized concentration data and abnormal Boolean matrix of all sampling points, element combination identification is performed for multiple single mineral types to determine the optimal element combination and weight coefficient for each single mineral type. The mineralization potential value of each sampling point is calculated based on the optimal element combination and weight coefficient of each single mineral type at each sampling point, as well as the normalized contrast value of each element at each sampling point. The total mineralization potential value of each sampling point is calculated based on the mineralization potential value of each single mineral at each sampling point. The area formed by each sampling point whose total mineralization potential value is greater than the first preset value is selected as the target area.

2. The method for evaluating mineralization potential according to claim 1, characterized in that, The process of collecting and preprocessing data from multiple sampling points in the target area to obtain standardized concentration data, normalized contrast values, and outlier Boolean matrices for each element includes: Data was collected from multiple sampling points in the target area to obtain the concentration data of each element at each sampling point; Determine whether there are missing element concentration data at each sampling point. If so, use the median filling method to fill in the missing element concentration values ​​to obtain the completed element concentration data. The completed element concentration data is transformed to obtain the transformed element concentration data. The median absolute deviation method was used to calculate the regional background value and the lower limit of anomalies for each element on the transformed data scale of the element concentration data. The abnormal Boolean matrix is ​​obtained by binarizing the transformed concentration data of each element based on the abnormal lower limit value of each element. The contrast value of each element is calculated based on the transformed concentration data of each element and the regional background value of each element. The contrast values ​​of each element are normalized to obtain the normalized contrast values ​​of each element; The transformed element concentration data are standardized to obtain the standardized concentration data of each element.

3. The method for evaluating mineralization potential according to claim 1, characterized in that, The process of transforming the completed element concentration data to obtain the transformed element concentration data includes: Logarithmic transformation or Box-Cox transformation is performed on the completed element concentration data to obtain the transformed element concentration data.

4. The method for evaluating mineralization potential according to claim 1, characterized in that, The method of identifying elemental combinations for multiple single mineral types based on standardized concentration data and outlier Boolean matrices from all sampling points, and determining the optimal elemental combination and weighting coefficients for each single mineral type, includes: Correlation analysis was performed on the standardized concentration data of each element at each sampling point to obtain the first set of elements with the first correlation to each single mineral type. Principal component analysis is used to select elements with first loadings on principal components whose cumulative variance contribution rate is greater than a second preset value for each element's standardized concentration data at each sampling point, thus obtaining a second set of elements. The second set of elements is then merged with the first set of elements and deduplicated to obtain a candidate set of elements. The machine learning algorithm is used to train the candidate element set with the element concentration as the feature and the target column of the single mineral type as the label to obtain the importance score of each element in the candidate element set. The importance score of each element in the candidate element set is then normalized to obtain the initial weight coefficient of each element in the candidate element set. Based on the anomaly Boolean matrix, the association rule mining algorithm is used to select frequently co-occurring element combinations from the candidate element set to obtain the third element combination set, and the support of each element combination in the third element combination set is calculated; wherein, the minimum support of the frequently co-occurring element combination is greater than or equal to 0.03 and less than or equal to 0.

5. The sum of the initial weight coefficients of each element in each frequently co-occurring element combination in the third element combination set is multiplied by the support of that element combination to obtain the comprehensive score of that element combination. Select the element combination with the highest comprehensive score from the third element combination set as the initial optimal element combination for a single mineral type; Select the third preset number of elements from the initial optimal element combination as the optimal element combination, and normalize the initial weight coefficients of each element in the optimal element combination to obtain the weight coefficients.

5. The method for evaluating mineralization potential according to claim 4, characterized in that, The first set of elements with primary correlation to each individual mineral type obtained by performing correlation analysis on the standardized concentration data of each element at each sampling point includes: Correlation analysis using Pearson correlation coefficient or Spearman rank correlation coefficient was performed on the standardized concentration data of each element at the sampling points to obtain the first set of elements with the first correlation to each individual mineral.

6. The method for evaluating mineralization potential according to claim 4, characterized in that, The step of selecting frequently co-occurring element combinations from the candidate element set using an association rule mining algorithm based on the anomaly Boolean matrix to obtain a third element combination set and calculating the support of each element combination in the third element combination set includes: Based on the anomalous Boolean matrix, the Apriori algorithm or FP-Growth algorithm is used to select frequently co-occurring element combinations from the candidate element set to obtain the third element combination set, and the support of each element combination in the third element combination set is calculated.

7. The method for evaluating mineralization potential according to claim 1, characterized in that, The calculation of the mineralization potential value of each sampling point corresponding to each single mineral type based on the optimal element combination and weighting coefficient of each single mineral type at each sampling point, as well as the normalized contrast value of each element at each sampling point, includes: In the formula, For a single mineral The mineralization potential value; This represents the number of elements in the optimal combination. For a single mineral In the optimal combination of elements, the first The weight coefficient of each element; For a single mineral In the optimal combination of elements, the first The normalized contrast values ​​of each element; where... By analyzing a single mineral In the optimal combination of elements, the first Contrast value of each element Normalization is performed to obtain the result.

8. The method for evaluating mineralization potential according to claim 7, characterized in that, The single mineral In the optimal combination of elements, the first The formula for calculating the contrast value of an element is: In the formula, For a single mineral In the optimal combination of elements, the first The transformed concentration data of each element; For the first The background value of the area for each element.

9. The method for evaluating mineralization potential according to claim 8, characterized in that, The calculation of the total mineralization potential value of each sampling point based on the mineralization potential value of each single mineral at each sampling point includes: In the formula, This represents the total mineralization potential value. This is the first adjustment coefficient; This is the second adjustment coefficient; This is the third adjustment coefficient, where, ; It is a function for maximizing the value; For a single mineral The mineralization potential value; For a single mineral The mineralization potential value; For a single mineral The mineralization potential value; For a single mineral The mineralization potential value; The quantity of a single mineral.

10. A mineralization potential evaluation system based on multi-element combination correlation analysis, characterized in that, include: The data acquisition and preprocessing module is used to acquire and preprocess data from multiple sampling points in the target area to obtain standardized concentration data, normalized contrast values, and outlier Boolean matrices for each element. The element combination identification module is used to identify the element combination of multiple single minerals based on the standardized concentration data and abnormal Boolean matrix of all sampling points, and to determine the optimal element combination and weight coefficient for each single mineral. The single mineral potential calculation module is used to calculate the mineralization potential value of each sampling point corresponding to each single mineral based on the optimal element combination and weight coefficient of each single mineral and the normalized contrast value of each element at each sampling point. The total potential synthesis module is used to calculate the total mineralization potential value of each sampling point based on the mineralization potential value of each individual mineral at each sampling point. The target area delineation module is used to select the area formed by each sampling point whose total mineral potential value is greater than the first preset value as the target area.

Images