Intelligent prospecting method and system based on association rule algorithm and knowledge graph technology

Abstract

The invention discloses an intelligent prospecting method and system based on an association rule algorithm and a knowledge graph technology. The method comprises the following steps: collecting ore deposit related data of each sampling point in a mineral exploration area; symbolizing and fusing each geological element of the ore deposit related data of each sampling point to obtain an association rule data table; using an association rule algorithm to obtain all strong association rules causing the occurrence of the to-be-searched element from the association rule data table; constructing a knowledge graph according to all the strong association rules; and obtaining a rule group of the to-be-searched element from the knowledge graph, wherein the rule group is used as a prospecting mark for searching the to-be-searched element in the mineral exploration area. According to the method, the association rule algorithm and the knowledge graph are combined, the strong association rules related to the ore deposit are mined through the association rules, the knowledge graph is created according to all the strong association rules, and the clustering group between the entities can be found by using the clustering algorithm of the knowledge graph.

Description

Technical Field

[0001] The present invention relates to the field of mineral exploration technology, and more specifically, to an intelligent mineral prospecting method and system based on association rule algorithm and knowledge graph technology. Background Art

[0002] In the field of mineral exploration, there are three kinds of ore-finding ideas: one is "seeking similarities", one is "seeking differences", and the last one is a combination of these two ideas. "Seeking similarities" means that the same ore deposit has similar metallogenic background and metallogenic environment. The metallogenic possibility of the predicted area is judged by analogy between the similarity of the predicted area and the known area, emphasizing the internal factors of the formation of the ore deposit; "seeking differences" is to determine the ore-finding signs through the abnormal information of geophysics, geochemistry and other aspects after the formation of the ore deposit, and then delineate the target area; the third idea is to "seek similarities" and "seek differences", considering both the ore-causing factors and the ore-causing factors, and emphasizing "internal and external cultivation".

[0003] The traditional thinking mode of prospecting mostly relies on the reasoning of strong causal relationships, which is as follows:

[0004]

[0005] In the above thinking paradigm, there is a strict causal relationship from A to B to C.

[0006] In the traditional prospecting method, after all the preliminary data preparation is completed, geologists need to process the data and use commercial software to draw various geological maps. They also need to consider which interpolation methods to use. The manual participation is too high and the workload is huge. Even if all the maps are drawn, highly qualified scientific researchers are still needed to integrate all the data to determine the prospecting signs of the area, which will then serve the prospecting work. The overall professional level of the team is too high and it is highly subjective, which may result in incomplete and inaccurate prospecting signs.

[0007] However, with the advent of the big data era, this thinking paradigm has gradually changed to a correlation relationship, as shown below:

[0008]

[0009] In this thinking paradigm, D is deduced to C, not by causality, but by correlation, combined with the knowledge constraints of mineralization theory to confirm the accuracy of C.

[0010] Currently, in the field of mineral exploration, association rule algorithms are mostly used for geochemical data analysis, and knowledge graph technology is used for the construction of mineral deposit knowledge bases and intelligent question and answer based on knowledge bases.

[0011] However, there are many problems with the current association rules in the process of prospecting. First of all, the geological data involved in prospecting often have the characteristics of multi-source heterogeneity and difficulty in integration, such as: various geological maps, text descriptions, data tables. It is difficult to reflect all the data on one map, and it is difficult to use the association rule algorithm to mine frequent item sets. At the same time, there are often thousands of strong association rules mined by the association rule algorithm. Which ones are useful or which ones to use are generally determined by humans. This selection is full of subjectivity. The abandoned strong association rules are a kind of data loss. Moreover, the formation process of mineral deposits is extremely complex. The few rules selected separately may only be local events or special events in the formation process of mineral deposits. Whether these rules can be directly applied to regional prospecting as prospecting signs remains to be discussed. And the research on knowledge graphs in the field of mineral deposits has mostly stayed at the establishment of knowledge bases. Summary of the invention

[0012] One of the purposes of the present invention is to provide an intelligent prospecting method based on association rule algorithm and knowledge graph technology to solve the problem of misleading regional prospecting due to the subjectivity of personal screening of strong association rules in the prior art; the second purpose of the present invention is an intelligent prospecting system based on association rule algorithm and knowledge graph technology.

[0013] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0014] The first aspect of the present invention provides an intelligent prospecting method based on association rule algorithm and knowledge graph technology, comprising the following steps:

[0015] Collect relevant data on mineral deposits at various sampling points in the mineral exploration area;

[0016] After symbolizing and fusing each geological element of the ore deposit related data of each sampling point, an association rule data table is obtained;

[0017] Using an association rule algorithm, obtaining all strong association rules that lead to the appearance of the element to be found in the association rule data table;

[0018] According to all the strong association rules, a knowledge graph is constructed;

[0019] A rule group related to the element to be found is obtained from the knowledge graph, and the rule group is used as a prospecting mark for finding the element to be found in the mineral exploration area.

[0020] In the above-mentioned technical means, the association rule algorithm and the knowledge graph are combined, and the strong association rules related to the mineral deposits are mined through the association rules. From the data perspective, each strong association rule has a high value, so all strong association rules should be applied to mineral exploration. Therefore, according to all the strong association rules, a knowledge graph is created, and then the clustering algorithm of the knowledge graph is used to find the clustering groups between entities. This clustering group is further clustered from the strong association rules that are highly correlated with the mineralization events of the mineral deposits. Therefore, this clustering group is a prospecting sign that is highly correlated with the mineralization events in the region.

[0021] Furthermore, the ore deposit related data includes multiple geological elements, and the geological elements include geological structure data, geophysical data and geochemical data.

[0022] Furthermore, each geological element of the ore deposit related data of each sampling point is symbolized, including:

[0023] Each geological element of the ore deposit related data of each sampling point is represented by a different symbol, and the geological elements that need to be classified are represented by a method of subscripting numbers with symbols.

[0024] Furthermore, the association rule data table includes:

[0025] The symbolized data of each geological element of the ore deposit related data of each sample point are formed into rows of data, and the rows of data formed by all the sample points are combined to form the association rule data table.

[0026] Furthermore, the association rule algorithm includes:

[0027] Let A and B be two events under the same rule, where A is the element to be found, then the support of event A is:

[0028] support(B→A)=P(B∩A)

[0029] The confidence level of event A is:

[0030] confidence(A→B)=P(B|A)

[0031] Each row in the association rule data table is a rule, and the support and confidence of event A occurring under each rule are calculated;

[0032] The rule whose support of event A is greater than the preset support is called a frequent item set;

[0033] A rule in which the confidence level of the occurrence of event A in the frequent item set is greater than a preset confidence level is called a strong association rule.

[0034] Furthermore, the association rule algorithm includes an Apriori algorithm, an Eclat algorithm or an Fpgrowth algorithm.

[0035] Furthermore, according to all the strong association rules, a knowledge graph is constructed, including:

[0036] The antecedent and consequent of the strong association rule are taken as entities, the corresponding support and confidence are taken as relationships, each of the strong association rules is saved as a triple, and then a knowledge graph is constructed.

[0037] Further, obtaining a rule group related to the element to be found from the knowledge graph includes:

[0038] Processing the knowledge graph using a clustering algorithm to obtain a clustered knowledge graph;

[0039] According to the clustered knowledge graph, a cluster with high correlation is obtained, and the rule corresponding to the cluster is the rule group of the element to be found.

[0040] Furthermore, the clustering algorithm includes a community discovery algorithm, a graph clustering algorithm or a random walk algorithm combined with a deep learning neural network model.

[0041] The second aspect of the present invention provides an intelligent prospecting system based on association rule algorithm and knowledge graph technology, comprising:

[0042] A data collection module, which collects mineral deposit related data at each sampling point in the mineral exploration area;

[0043] A data symbolization and fusion module, wherein the data symbolization and fusion module symbolizes and fuses each geological element of the mineral deposit-related data of each sampling point to obtain an association rule data table;

[0044] A strong association rule acquisition module, wherein the strong association rule acquisition module uses an association rule algorithm to acquire all strong association rules that cause the element to be found to appear in the association rule data table;

[0045] A graph construction module, wherein the graph construction module constructs a knowledge graph according to all the strong association rules;

[0046] A prospecting mark module, wherein the prospecting mark module obtains a rule group related to the element to be found from the knowledge graph, and the rule group serves as a prospecting mark for finding the element to be found in the mineral exploration area.

[0047] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0048] The present invention realizes the fusion of multi-source heterogeneous data of mineral deposits, so that all strong association rules are involved in mineral prospecting, which not only solves the problem of data loss, but also overcomes the problem of misleading regional mineral prospecting due to the subjectivity of personal screening of strong association rules, and solves the limitation that a single strong association rule cannot be generalized to regional mineral prospecting, so that the original mineral prospecting using a single or multiple association rules is transformed into finding a regional strong association rule group to guide mineral prospecting. In addition, a single strong association rule and a strong association rule group can be seen at the same time from the present invention, showing the change from local to regional mineral prospecting, and the change of other elements in the process of graded element level change can be seen, and at the same time, multiple functions such as correlation analysis and cluster analysis are possessed. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A schematic diagram of a process flow of an intelligent prospecting method based on association rule algorithm and knowledge graph technology provided by an embodiment of the present invention;

[0050] Figure 2 A schematic diagram of a knowledge graph after clustering provided by an embodiment of the present invention;

[0051] Figure 3 A module schematic diagram of an intelligent mineral prospecting system based on association rule algorithm and knowledge graph technology provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0053] In order to better illustrate the present embodiment, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product;

[0054] It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0055] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0056] Example

[0057] The embodiment of the present invention provides an intelligent prospecting method based on association rule algorithm and knowledge graph technology, such as Figure 1 As shown, the following steps are included:

[0058] Collect relevant data on mineral deposits at various sampling points in the mineral exploration area;

[0059] After symbolizing and fusing each geological element of the ore deposit related data of each sampling point, an association rule data table is obtained;

[0060] Using an association rule algorithm, obtaining all strong association rules that lead to the appearance of the element to be found in the association rule data table;

[0061] According to all the strong association rules, a knowledge graph is constructed;

[0062] A rule group related to the element to be found is obtained from the knowledge graph, and the rule group is used as a prospecting mark for finding the element to be found in the mineral exploration area.

[0063] In a further embodiment, the ore deposit related data includes a plurality of geological elements, wherein the geological elements include geological structure data, geophysical data and geochemical data.

[0064] In a specific embodiment, the geological element may be a certain stratum, a certain fracture or a certain chemical element.

[0065] In a further embodiment, symbolizing each geological element of the ore deposit related data of each sampling point includes:

[0066] Each geological element of the ore deposit related data of each sampling point is represented by a different symbol, and the geological elements that need to be classified are represented by a method of subscripting numbers with symbols.

[0067] In a specific embodiment, there are multiple sampling points in the mineral exploration area. According to the coordinate information of the sampling points, there are three different elements A, B, and C, and their concentration levels are represented by 1, 2, 3, and 4. Then A1 represents the concentration level 1 of element A, and the specific concentration level design can be designed according to actual needs; let D represent a fault in the mineral exploration area, and use 1, 2, 3, and 4 to represent the distance between the sampling point and the fault. D1 represents that the sampling point is at a level 1 distance from the D fault, and the specific distance level design can be designed according to actual needs; E can also be used to represent the exposed stratum at the sampling point. All elements, structures, strata, etc. belonging to the ore-related data have no quantitative restrictions when using standardized representation, and as many as possible are counted.

[0068] In a further embodiment, the association rule data table includes:

[0069] The symbolized data of each geological element of the ore deposit related data of each sample point are formed into rows of data, and the row data formed by all the sample points are combined to form the association rule data table, as shown in Table 1.

[0070] Table 1

[0071] Sample 1 A1 B2 C2 D4 E Sample 2 B C1 D4 E Sample 3 A3 B1 C2 D3 E Sample 4 C2 D4 E Sample... ... ... ... ... ...

[0072] In a further embodiment, the basic principle of the association rule algorithm is to assume that A and B are two events under the same rule, where A is the element to be found, and the support of event A is:

[0073] support(B→A)=P(B∩A)

[0074] The confidence level of event A is:

[0075] confidence(A→B)=P(B|A)

[0076] Each row in the association rule data table is a rule, and the support and confidence of event A occurring under each rule are calculated;

[0077] The rule whose support of event A is greater than the preset support is called a frequent item set;

[0078] A rule in which the confidence level of the occurrence of event A in the frequent item set is greater than a preset confidence level is called a strong association rule.

[0079] In the prospecting work, A, B, C, D, and E in Table 1 are all factors related to mineralization. Assuming that the mine of element A is to be found, the association rule algorithm is used to run out all the strong association rules that lead to the appearance of A, such as: antecedent (D1, C3, E) → consequent (A3), which means that when D1, C3, and E occur, A3 can be deduced to occur, which is a rule. However, especially when the amount of data is extremely large, there are tens of thousands of such strong association rules generated. Other methods of association rule analysis generally manually select rules that individuals think are meaningful, that is, rules that can be directly interpreted geologically for analysis. This processing solution not only has the characteristics of serious data loss and strong subjectivity, but also deviates from the big data thinking paradigm based on correlation, and returns to the traditional causal analysis. In addition, the singled-out rules are likely to be the rules of a small area or a small point. Whether it is reasonable to use it as a regional prospecting mark remains to be considered. This embodiment combines a variety of commonly used association rule algorithms, and all strong association rules can be seen. It also uses them as data sources as input data for the knowledge graph. This fundamentally solves the problem of all strong association rules participating in big data prospecting, and also solves the problem that the correlation between traditional knowledge graph input data and the mineral deposit itself cannot be determined. All strong association rules obtained by using the association rule algorithm are related combinations of various elements in regional mineral exploration, which is more scientific and accurate as the input of the knowledge graph.

[0080] In a further embodiment, the association rule algorithm includes an Apriori algorithm, an Eclat algorithm or an Fpgrowth algorithm.

[0081] In a further embodiment, a knowledge graph is constructed according to all the strong association rules, including:

[0082] The antecedent and consequent of the strong association rule are taken as entities, the corresponding support and confidence are taken as relationships, each of the strong association rules is saved as a triple, and then a knowledge graph is constructed.

[0083] In a further embodiment, obtaining a rule group related to the element to be found from the knowledge graph includes:

[0084] Processing the knowledge graph using a clustering algorithm to obtain a clustered knowledge graph;

[0085] According to the clustered knowledge graph, a cluster with high correlation is obtained, and the rule corresponding to the cluster is the rule group of the element to be found.

[0086] In a further embodiment, the clustering algorithm can be a community discovery algorithm or a graph clustering algorithm, for example, clustering analysis can be performed through the Laplacian matrix of a graph, or a density-based clustering algorithm can be used, such as DBSCAN (Density-Based Spatial Clustering of Applications with Noise), or a random walk algorithm combined with a deep learning neural network model can be used.

[0087] In a specific embodiment, a community discovery algorithm is used as an illustration. The community discovery algorithm (Louvain algorithm) is an algorithm based on modularity, which allows a node to try to traverse the community labels of all neighbors and select the community label that maximizes the modularity increment. After maximizing the modularity, each community is regarded as a new node, and the process is repeated until the modularity no longer increases.

[0088] Modularity calculation formula:

[0089] Among them, Σ i n represents the weight within community c, Σt o t represents the weight of the edge connected to the points inside community c, including the edges inside the community and the edges outside the community. m is the sum of the weights of all links in the network.

[0090] After clustering the knowledge graph using the above-mentioned community discovery algorithm, we get a knowledge graph cluster. From the graph cluster, we can see the regular groups that are closely related to mineralization in the area. These regular groups are the signs of mineral exploration in this area.

[0091] In a specific embodiment, taking a gold mine survey area as an example, the geochemical prospecting signs of the gold deposits in the area, predecessors obtained the following understanding through mapping and comprehensive analysis:

[0092] The elements with high regional anomaly intensity, large area and obvious concentration center are mainly Au, Sb, Hg, As, Ag, Cd, etc. The abnormal association between elements is shown in Table 2.

[0093] Table 2

[0094] Main abnormal element Associated Abnormal Elements Sb Au, As, Hg Au Sb, As, Hg Cd, Au, Sb, Hg As, Ag, Pb, Zn, Y As Au, Sb, Hg, Ag

[0095] The entire process from obtaining raw data to drawing, detailed comprehensive analysis, and finally drawing conclusions would take at least half a month or even longer for one person, and would require a high level of professionalism.

[0096] If this system is used to find mineral exploration signs, the cluster diagram obtained through analysis is as follows Figure 2 shown.

[0097] From the cluster analysis results, the highly abnormal elements (level 2 or above, such as Au3) are: Au, As, Ag, Sb, Cd, Hg, Bi, Sn, Pb, etc. According to community discovery, the associated communities obtained are shown in Table 3.

[0098] Table 3

[0099] Abnormal Community Group (Au3, As3, Mo3, Cu3), (Au2, As2, Mo2), (Au1, As1, Sb1, Ag1, Hg1) Pb3, Pb2, Ni2, Cr2, Cu2, Ag2, Co2, Sn2 Cd3, Cd2, Hg3, Hg2, Zn3, Zn2, Ag3, Sn3, W3, W2, Sb2, Cr3, Bi3 Pb1, Mo1, Bi1, Ba2 V1, Cu1, Co1, Ag0, Hg0

[0100] Through community discovery, it is found that the prospecting signs related to Au anomalies discovered by predecessors are anomalies of As, Hg, and Sb. However, this conclusion has a more detailed explanation after systematic clustering. The Au anomaly is most closely related to As, and has a positive correlation in content. The extremely high Au anomaly is also closely correlated with Mo, while Au is only well correlated with the extremely high anomalies of Hg, Sb, and Ag at low contents; the high anomaly of Cd is highly correlated with the high anomalies of Hg, Zn, W, Ag, Cr, Bi, and Sb, and the concentrations of Cd and Hg, Zn, and W are positively correlated; the high anomaly of Pb is highly correlated with the medium and high anomalies of Ni, Cr, Cu, Ag, Co, and Sn.

[0101] With this understanding, it only takes about 30 minutes for one person to construct the association rule data table, generate strong association rules, and then comprehensively, objectively and accurately mine the regional prospecting signs to reach these conclusions, while it only takes a few seconds to obtain a visual clustering diagram through the algorithm.

[0102] The embodiment of the present invention also provides an intelligent prospecting system based on association rule algorithm and knowledge graph technology, such as Figure 3 As shown, including:

[0103] A data collection module, which collects mineral deposit related data at each sampling point in the mineral exploration area;

[0104] A data symbolization and fusion module, wherein the data symbolization and fusion module symbolizes and fuses each geological element of the mineral deposit-related data of each sampling point to obtain an association rule data table;

[0105] A strong association rule acquisition module, wherein the strong association rule acquisition module uses an association rule algorithm to acquire all strong association rules that cause the element to be found to appear in the association rule data table;

[0106] A graph construction module, wherein the graph construction module constructs a knowledge graph according to all the strong association rules;

[0107] A prospecting mark module, wherein the prospecting mark module obtains a rule group related to the element to be found from the knowledge graph, and the rule group serves as a prospecting mark for finding the element to be found in the mineral exploration area.

[0108] The same or similar reference numerals correspond to the same or similar components;

[0109] The terms used in the drawings to describe positional relationships are only used for illustrative purposes and should not be construed as limiting this patent;

[0110] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. An intelligent prospecting method based on association rule algorithm and knowledge graph technology, It is characterized in that The following steps are involved: Collect relevant data on mineral deposits at various sampling points in the mineral exploration area; After symbolizing and integrating each geological element of the ore deposit related data of each sampling point, an association rule data table is obtained; Using an association rule algorithm, obtaining all strong association rules that lead to the appearance of the element to be found in the association rule data table; According to all the strong association rules, a knowledge graph is constructed; A rule group related to the element to be found is obtained from the knowledge graph, and the rule group is used as a prospecting mark for finding the element to be found in the mineral exploration area.

2. The intelligent prospecting method based on association rule algorithm and knowledge graph technology according to claim 1, It is characterized in that The ore deposit related data includes multiple geological elements, including geological structure data, geophysical data and geochemical data.

3. The intelligent prospecting method based on association rule algorithm and knowledge graph technology according to claim 2, It is characterized in that Symbolize each geological element of the ore deposit related data of each sampling point, including: Each geological element of the ore deposit related data of each sampling point is represented by a different symbol, and the geological elements that need to be classified are represented by a method of subscripting numbers with symbols.

4. The intelligent prospecting method based on association rule algorithm and knowledge graph technology according to claim 3, It is characterized in that The association rule data table includes: The symbolized data of each geological element of the ore deposit related data of each sample point are formed into rows of data, and the rows of data formed by all the sample points are combined to form the association rule data table.

5. The intelligent prospecting method based on association rule algorithm and knowledge graph technology according to claim 4, It is characterized in that The association rule algorithm comprises: Let A and B be two events under the same rule, where A is the element to be found, then the support of event A is: support(B→A)=P(B∩A) The confidence level of event A is: confidence(A→B)=P(B|A) Each row in the association rule data table is a rule, and the support and confidence of event A occurring under each rule are calculated; The rule whose support of event A is greater than the preset support is called a frequent item set; A rule in which the confidence level of the occurrence of event A in the frequent item set is greater than a preset confidence level is called a strong association rule.

6. The intelligent prospecting method based on association rule algorithm and knowledge graph technology according to claim 5, It is characterized in that The association rule algorithm includes Apriori algorithm, Eclat algorithm or Fpgrowth algorithm.

7. The intelligent prospecting method based on association rule algorithm and knowledge graph technology according to claim 5, It is characterized in that According to all the strong association rules, a knowledge graph is constructed, including: The antecedent and consequent of the strong association rule are taken as entities, the corresponding support and confidence are taken as relationships, each of the strong association rules is saved as a triple, and then a knowledge graph is constructed.

8. The intelligent prospecting method based on association rule algorithm and knowledge graph technology according to claim 7, It is characterized in that Obtaining a rule group related to the element to be found from the knowledge graph includes: Processing the knowledge graph using a clustering algorithm to obtain a clustered knowledge graph; According to the clustered knowledge graph, a cluster with high correlation is obtained, and the rule corresponding to the cluster is the rule group of the element to be found.

9. The intelligent prospecting method based on association rule algorithm and knowledge graph technology according to claim 7, It is characterized in that The clustering algorithm includes a community discovery algorithm, a graph clustering algorithm or a random walk algorithm combined with a deep learning neural network model.

10. An intelligent mineral prospecting system based on association rule algorithm and knowledge graph technology. It is characterized in that include: A data collection module, which collects mineral deposit related data at each sampling point in the mineral exploration area; A data symbolization and fusion module, wherein the data symbolization and fusion module symbolizes and fuses each geological element of the mineral deposit-related data of each sampling point to obtain an association rule data table; A strong association rule acquisition module, wherein the strong association rule acquisition module uses an association rule algorithm to acquire all strong association rules that cause the element to be found to appear in the association rule data table; A graph construction module, wherein the graph construction module constructs a knowledge graph according to all the strong association rules; A prospecting mark module, wherein the prospecting mark module obtains a rule group related to the element to be found from the knowledge graph, and the rule group serves as a prospecting mark for finding the element to be found in the mineral exploration area.

Images