An automatic settlement system for mineral products

Abstract

The application relates to the technical field of mineral product settlement, and discloses a mineral product automatic settlement system, which has the technical scheme as follows: a division unit, a checking unit, a changing unit, a calculation unit, an influence unit and a recalculation unit are cooperated, the division unit generates attribute positioning mark accurate management data, the checking unit filters abnormities in advance, the changing unit automatically marks and records tracks when data is changed, the calculation unit constructs a dependence graph, the influence unit divides correlation levels, and the recalculation unit is used for differential treatment, so that the data checking and recalculation time is reduced, the efficiency problem caused by rechecking data change is solved, and the settlement is efficient and accurate.

Description

Technical Field

[0001] This invention relates to the field of mineral product settlement technology, and specifically to an automatic mineral product settlement system. Background Technology

[0002] Currently, in the purchase and sale management of bulk minerals, business personnel usually use spreadsheets to collect data on the weight, moisture, grade, and price of mineral products, and calculate process data such as metal content, pricing coefficient, processing fees, grade fluctuations, impurity deductions, and the final settlement amount.

[0003] However, when staff enter data for each batch of mineral products into a spreadsheet for calculation, the data may change due to re-inspection of the mineral products. After re-entering the data, it is necessary to check whether all derived calculations of the data have changed. This will increase the data verification time and reduce the efficiency of automatic settlement. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an automated settlement system for mineral products, which solves the aforementioned problems.

[0005] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0006] An automated settlement system for mineral products, comprising:

[0007] The unit is used to obtain the electronic spreadsheet for mineral product settlement. It is divided into multiple functional data areas by rows and columns. By parsing the business attributes of each data area, the attribute positioning identifier of the data area is generated.

[0008] The verification unit is used to implicitly mark the data area based on the attribute positioning identifier after any functional data area is filled with data, and at the same time verify the format and range of the filled data, and generate the valid initial verification parameters for the data area.

[0009] The change unit is used to identify the data area as the changed data area when the data area of ​​the valid initial verification parameter changes, convert its implicit mark into an explicit mark, and record the complete values, operation time and operation subject of the changed data area before and after the change, and generate a data change trajectory.

[0010] The calculation unit is used to construct the calculation dependency relationship between each data area by traversing the calculation formula association chain of the spreadsheet based on the attribute location identifier and data change trajectory, and generate a data dependency map.

[0011] The influence unit is used to locate all data areas that have a computational relationship with the changing data area based on the data dependency map, and to generate a relationship influence level by treating them as derived data areas.

[0012] The recalculation unit is used to perform differentiated recalculation on derived data areas at different levels based on the correlation influence level, and generate recalculation result verification results.

[0013] Furthermore, by parsing the business attributes of each data zone, an attribute location identifier for that data zone is generated, including:

[0014] Extract the business attribute information of each data area, classify and process it, and generate a business attribute feature set;

[0015] The spatial location of the data area in the spreadsheet is analyzed, and spatial topology coding is generated by combining the row and column relationships of the table.

[0016] Associate the business attribute feature set with the spatial topology code to generate an attribute-space mapping matrix;

[0017] An importance analysis is performed on each attribute in the business attribute feature set to generate business weight coefficients;

[0018] Based on the business weight coefficient, the spatial topology coding and attribute-space mapping matrix are fused to generate the attribute location identifier for the data area.

[0019] Furthermore, importance analysis is performed on each attribute in the business attribute feature set to generate business weight coefficients, including:

[0020] Analyze the irreplaceability of each attribute in the business process and the frequency of its use as downstream calculation input to generate a link dependency index;

[0021] Analyze the impact of anomalies in each attribute on the settlement results and the correction costs, and generate an anomaly impact coefficient;

[0022] The link dependency index and the anomaly impact coefficient were analyzed, and the confidence scores of the link index and the anomaly coefficient were generated respectively.

[0023] The link dependency index and the anomaly impact coefficient are fused based on the link index confidence level and the anomaly coefficient confidence level to generate the business weight coefficient.

[0024] Furthermore, after any functional data area is filled with data, the data area is implicitly marked based on the attribute positioning identifier, and the format and range of the filled data are verified to generate valid initial verification parameters for the data area, including:

[0025] Based on the attribute location identifier, the business attribute characteristics of the data area are parsed to generate a unique corresponding implicit tag code;

[0026] Extract the format characteristics of the data area fill content, compare it with the normal format standard of this business attribute, and generate a format matching index;

[0027] Based on the business attribute feature set, determine the reasonable value range of the data to be filled in the data area, compare the filled data with the range range, and generate a range compliance index.

[0028] The format matching index and the range compliance index are integrated to generate valid initial verification parameters for the data area.

[0029] Furthermore, when the data area of ​​the valid initial verification parameters changes, this data area is identified as the changed data area, its implicit marker is converted into an explicit marker, and the complete values ​​before and after the change, the operation time, and the operation subject are recorded to generate a data change trajectory, including:

[0030] For the data area with valid initial verification parameters, the current data is compared with the baseline data in the valid initial verification parameters to generate a data variability index;

[0031] When the data variability index reaches the change judgment threshold, the data area is identified as the changed data area. At the same time, based on the implicit tag encoding, an explicit tag activation instruction is generated to trigger the implicit tag to be converted into an explicit tag.

[0032] Obtain the original data snapshot at the moment of change, the updated data snapshot, the digital signature of the operating entity, and the operation sequence code to generate a set of core elements of the change;

[0033] Based on the data variability index, the importance of the core variable element set is classified and element level identifiers are generated.

[0034] Analyze the element hierarchy identifiers and the core set of changes to generate a data change trajectory that includes the depth and scope of the change.

[0035] Furthermore, based on attribute location identifiers and data change trajectories, by traversing the calculation formula association chains of the spreadsheet, the calculation dependencies between different data areas are constructed, generating a data dependency graph, including:

[0036] Analyze the attribute location identifiers and data change trajectories of each data area to generate a set of formula-related links;

[0037] Based on the formula-related link set, the correlation frequency of each data region as a calculation input and output is analyzed to generate a dependency strength coefficient;

[0038] Based on the dependency strength coefficient, the set of formula-related links is sorted out, and a topology structure with data areas as nodes and dependency relationships as edges is constructed to generate a data dependency graph.

[0039] Furthermore, based on the dependency strength coefficient, the set of related links in the formula is sorted out, and a topology structure with data areas as nodes and dependencies as edges is constructed to generate a data dependency graph, including:

[0040] The dependency strength coefficient is divided into intervals, and the dependency relationships between data areas are divided into three levels according to the numerical value, and dependency level labels are generated.

[0041] Based on the formula-related link set, the indirect correlation degree between data areas under different dependency levels is calculated, and secondary correlation coefficients are generated.

[0042] Generate an attributed topology based on dependency level labels and secondary association coefficients;

[0043] Based on the attributed topology, a data dependency graph containing multi-level dependencies and attributes is generated.

[0044] Furthermore, based on the data dependency map, all data regions that are computationally correlated with the changing data regions are located and treated as derived data regions to generate a correlation influence hierarchy, including:

[0045] Starting with the variable data area, the topological structure in the data dependency graph is traversed to generate an initial associated dataset.

[0046] Based on the dependency level labels and secondary association coefficients in the data dependency graph, the transmission depth and intensity of the impact of changes on each derived data area in the initial associated dataset are calculated, and the impact transmission index is generated.

[0047] Based on the distribution range of the influence transmission index and the core position of the derived data area in the settlement logic, the initial associated dataset is hierarchically divided to generate hierarchical pre-classification results.

[0048] The hierarchical pre-classification results are verified by comparing the correlation degree of business attribute features of each level of derived data area, and a hierarchical verification coefficient is generated.

[0049] The hierarchical pre-classification results are re-divided based on the hierarchical verification coefficient, and finally a hierarchical relationship influence hierarchy containing the core layer, the related layer, and the radiation layer is generated.

[0050] Furthermore, based on the hierarchy of influence, differentiated recalculation is performed on derived data areas at different levels, generating recalculation result verification results, including:

[0051] Analyze the distribution of derived data areas at the level of correlation impact, and generate differentiated recalculation trigger thresholds by combining the business attribute characteristics of data at each level.

[0052] Based on the differentiated recalculation trigger threshold, the core layer, the related layer and the radiation layer are recalculated separately to generate recalculation performance indicators.

[0053] Obtain the recalculation output data of each level and integrate it into a hierarchical recalculation result set;

[0054] Extract the data difference before and after recalculation of each level of derived data area in the association influence hierarchy, and calculate the degree of deviation between the recalculation result and the theoretical value in combination with the business attribute feature set to generate the result deviation coefficient;

[0055] The recalculation performance index is correlated with the result deviation coefficient to generate a recalculation quality score;

[0056] Based on the recalculation quality score, the hierarchical recalculation result set is graded and verified, the derived data areas with abnormal deviations are marked, and the recalculation result verification result containing the verification conclusion and the abnormal annotation is generated.

[0057] Furthermore, based on the differentiated recalculation trigger threshold, the core layer, related layer, and radiation layer are recalculated separately to generate recalculation performance indicators, including:

[0058] The differential recalculation trigger thresholds are analyzed to generate core layer trigger critical points, associated layer batch thresholds, and radiation layer time window thresholds.

[0059] Based on the core layer trigger critical point, real-time recalculation of the core layer derived data area is initiated to generate real-time response coefficients.

[0060] Based on the batch threshold of the association layer, the derived data area of ​​the association layer is aggregated to generate a batch processing index;

[0061] Based on the radiation layer time window threshold, the radiation layer derived data area is divided into recalculation periods to generate timed execution coefficients;

[0062] The real-time response coefficient, batch processing index, and timed execution coefficient are integrated to generate a recalculation performance index.

[0063] In summary, the present invention has the following main beneficial effects:

[0064] By dividing the spreadsheet data area into units for precise parsing, attribute positioning identifiers containing business attributes and spatial location are generated. The verification unit verifies the data format and range in real time and filters abnormal data in advance. The change unit automatically captures data changes and generates a trajectory, eliminating the need for manual marking and recording. The linkage of each unit eliminates problems such as ambiguous data positioning, cumbersome verification, and difficulty in tracking changes in traditional spreadsheet operations. It transforms the settlement process from manual to system-processed, reduces human intervention, and improves the overall efficiency and standardization of mineral product settlement.

[0065] By constructing a data dependency graph based on attribute location markers and change trajectories through calculation units, the computational relationships of each data area are clearly presented. The influencing units locate the derived data areas and generate a hierarchy of related influences, distinguishing between the core layer, the related layer, and the radiation layer. This hierarchical processing mechanism eliminates the need for staff to manually check all derived calculations and allows them to focus on the core data areas for priority processing. This significantly shortens the verification and recalculation time after data changes, effectively solving the problem of low settlement efficiency caused by data changes in the traditional model, and ensuring the timeliness and accuracy of data change transmission.

[0066] By using recalculation units based on the hierarchy of related influences, the core layer is recalculated in real time, the related layer is processed in batches, and the radiation layer is recalculated on a regular basis. Combined with business weight coefficients, resources are allocated reasonably to avoid waste caused by indiscriminate calculations. At the same time, the quality score generated based on the recalculation efficiency index and the result deviation coefficient can accurately mark abnormal data areas. This hierarchical recalculation and intelligent verification mode not only ensures the high accuracy of core data, but also improves the overall recalculation efficiency, allowing staff to quickly locate and handle anomalies, reduce invalid verification work, and ultimately achieve the reliability of mineral product settlement results and the efficiency of the settlement process, providing strong technical support for the management of bulk mineral purchase and sale. Attached Figure Description

[0067] Figure 1 This is a schematic diagram of the automatic settlement system for mineral products of the present invention. Detailed Implementation

[0068] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0069] refer to Figure 1 An automated settlement system for mineral products, comprising:

[0070] The unit is used to obtain the electronic spreadsheet for mineral product settlement. It is divided into multiple functional data areas by rows and columns. By parsing the business attributes of each data area, the attribute positioning identifier of the data area is generated.

[0071] The verification unit is used to implicitly mark the data area based on the attribute positioning identifier after any functional data area is filled with data, and at the same time verify the format and range of the filled data, and generate the valid initial verification parameters for the data area.

[0072] The change unit is used to identify the data area as the changed data area when the data area of ​​the valid initial verification parameter changes, convert its implicit mark into an explicit mark, and record the complete values, operation time and operation subject of the changed data area before and after the change, and generate a data change trajectory.

[0073] The calculation unit is used to construct the calculation dependency relationship between each data area by traversing the calculation formula association chain of the spreadsheet based on the attribute location identifier and data change trajectory, and generate a data dependency map.

[0074] The influence unit is used to locate all data areas that have a computational relationship with the changing data area based on the data dependency map, and to generate a relationship influence level by treating them as derived data areas.

[0075] The recalculation unit is used to perform differentiated recalculation on derived data areas at different levels based on the correlation influence level, and generate recalculation result verification results.

[0076] By dividing the data into units to generate attribute location identifiers, and combining this with the verification unit to verify the validity of the data in real time, when the data changes, the change unit automatically records the trajectory and marks it, the calculation unit constructs a dependency graph, the impact unit quickly locates the associated data area, and the recalculation unit realizes differentiated recalculation, which greatly reduces the data verification time, significantly improves the efficiency of automatic settlement, and ensures the consistency and accuracy of the data.

[0077] In one embodiment, by parsing the business attributes of each data area, an attribute location identifier for that data area is generated, including:

[0078] Extract business attribute information from each data area, classify and process it, and generate a business attribute feature set. Specifically, this includes: using a pre-set list of business keywords, checking each cell in the spreadsheet one by one, comparing the cell content with the pre-set list of business keywords (including weight, moisture, grade, and processing fees, etc.) to find the cell location of the attribute name. After determining the attribute name location, using fixed coordinate offset rules (e.g., if the attribute name is in cell A1, then cells B1 to B10 to its right and cells A2 to A10 below it are considered the corresponding data area), the data range is defined, and the data is extracted: Numerical data retains the original number and decimal point (e.g., 50.23), which can directly participate in addition, subtraction, multiplication, and division operations, and the attribute type is numeric; Character data retains all symbols (e.g., 25.5%), which cannot be directly... The operation is performed, and the attribute type is character. Simultaneously, the data dimension is determined: data occupying only one cell is a single value, while data occupying five or more consecutive cells is a multiple sequence. The meaning of each keyword in the business keyword list is specified; for example, the keyword "processing fee" means: the fee calculated based on the ore weight and unit processing price. The identified attribute names are used as search keywords, and precise word matching is performed on the business keyword list to obtain a key-value pair structure of attribute name + detailed explanation. Then, the key-value pair structure is concatenated at the field level with the previously determined attribute type (e.g., numeric) and data dimension (e.g., single value) to form a structured data unit containing four elements: attribute name, business meaning, type, and dimension. Finally, all data units are arranged according to their spatial order in the spreadsheet, integrating them into a complete business attribute feature set.

[0079] The spatial location of the data area in the spreadsheet is analyzed, and a spatial topology code is generated based on the row and column relationships of the table. Specifically, this includes: locating the coordinates of the first cell (e.g., B2) and the last cell (e.g., B10) of the data area; converting the column letter to a number (B corresponds to 2) to obtain the starting coordinates (2, 2) and ending coordinates (2, 10); determining the header row (e.g., A1) to which the data area belongs based on the row and column relationships of the table; encoding it as the parent ID (A1); and finally generating a spatial topology code such as "Number Area-2-2-2-10-A1" according to the fixed format of "Area Type-Start Column Number-Start Row Number-End Column Number-End Row Number-Parent ID". The spatial topology code contains both the specific coordinates and reflects the hierarchical relationship with the header.

[0080] Associating business attribute feature sets and spatial topology codes to generate an attribute-space mapping matrix involves: using the parent ID in the spatial topology code as the association key, binding the structured data units in the business attribute feature set with the corresponding spatial topology codes, arranging the binding results in the order of rows and columns of a spreadsheet, generating a two-dimensional matrix where rows represent data areas and columns are attribute name, business meaning, type, dimension, and spatial topology code, forming an attribute-space mapping matrix.

[0081] An importance analysis is performed on each attribute in the business attribute feature set to generate business weight coefficients;

[0082] Based on the business weight coefficient, the spatial topology code and the attribute-spatial mapping matrix are fused to generate the attribute location identifier for the data area. Specifically, this includes extracting four elements from the rows of the attribute-spatial mapping matrix: attribute name, business meaning, type, and dimension. These elements and fields are then combined with the fields of the spatial topology code and sorted from high to low according to the business weight coefficient. The resulting string is the attribute location identifier for the data area.

[0083] By extracting business attributes to generate feature sets, constructing a mapping matrix by combining spatial topology coding, and then incorporating business weights to generate attribute positioning identifiers, the system can accurately locate each data area, avoiding the tediousness and errors of manual data searching and making data positioning more efficient. At the same time, the attribute positioning identifiers can clearly reflect the business attributes and spatial relationships of the data area. When data changes due to reasons such as re-inspection, the system can quickly locate the changed data area and its associated derivative data areas based on the identifiers, without the need for manual checking of dependencies. This significantly shortens the data verification and recalculation time, effectively solves the problem of low settlement efficiency caused by data changes, and ensures the accuracy of the settlement process.

[0084] In one embodiment, importance analysis is performed on each attribute in the business attribute feature set to generate business weight coefficients, including:

[0085] To generate a link dependency index, the analysis of the irreplaceability of each attribute in the business process and its frequency as downstream calculation input is performed. Specifically, for each attribute in the business attribute feature set, all settlement formulas in the spreadsheet are traversed, and each formula input item is scanned row by row. Each occurrence of the attribute is recorded as one, and the input frequency is accumulated. Then, the number of other attributes that can substitute for the attribute is counted (for example, if the settlement weight can be expressed as weighed weight or theoretical weight (volume multiplied by density)). Weighed weight and theoretical weight are substitutes for each other. If there is no substitute, the irreplaceability score is 1 point; if there is 1 substitute, the irreplaceability score is 0.5 points; if there are 2 substitutes, the irreplaceability score is 1 / 3 points, and so on. The irreplaceability score = (1 ÷ (number of substitutes + 1)). Finally, the input frequency is multiplied by 0.6 and the irreplaceability score is multiplied by 0.4 to obtain the link dependency index.

[0086] The analysis examines the impact of anomalies in various attributes on the settlement results and the cost of correction, generating an anomaly impact coefficient. Specifically, this includes: setting an allowable error range of 0.5% for each attribute; when an attribute value exceeds this range by 1%, calculating the deviation percentage of the settlement result as the impact factor (e.g., if the total amount deviates by 3%, the impact factor is 0.03); and then calculating the required manpower to correct the anomaly. For example, if the manpower time is 2 hours and the hourly labor cost is 50 yuan, then 2 multiplied by 50, divided by 1, and multiplied by 1000 (where 1000 is the total amount of a single settlement) yields the result. This is the correction cost. Multiplying the impact by 0.7 and then adding the correction cost multiplied by 0.3 gives the anomaly impact coefficient. In mineral transactions, the settlement result is directly related to the funds of both parties. Settlement deviations caused by attribute anomalies may lead to major economic disputes and even affect long-term cooperation. Therefore, the impact is more important and is given a high weight of 0.7. The correction cost reflects the efficiency loss of anomaly handling and is a secondary factor, so it is given a low weight of 0.3. This balances the core business objective (accurate results) with operational efficiency (controllable costs).

[0087] The link dependency index and the anomaly impact coefficient are analyzed, and the confidence scores of the link index and the anomaly coefficient are generated respectively. Specifically, the confidence scores of the link dependency index are obtained by dividing the number of settlement formulas used in the calculation of the link dependency index by the total number of settlement formulas in the spreadsheet; the confidence scores of the anomaly coefficient of this attribute are obtained by dividing the number of anomaly coefficient confidence scores of this attribute by the total number of anomaly records of all attributes.

[0088] The link dependency index and the anomaly impact coefficient are fused based on the link index confidence and the anomaly coefficient confidence to generate the business weight coefficient. Specifically, the link index confidence is used as the weight of the link dependency index, and the anomaly coefficient confidence is used as the weight of the anomaly impact coefficient. The link dependency index and the anomaly impact coefficient are multiplied by their corresponding weights and then added together. The sum is then divided by the sum of the two weights to obtain the business weight coefficient.

[0089] By generating business weight coefficients, the importance of each attribute in settlement is accurately quantified. The system comprehensively considers the attribute's dependency index and anomaly impact coefficient, and integrates these with confidence levels. This allows the system to prioritize key attributes and their associated calculations when data changes, reducing unnecessary verification steps, improving the targeting and efficiency of data processing, effectively solving the time-consuming problem of manual dependency checks, and ensuring the efficiency and accuracy of the settlement process.

[0090] In one embodiment, after any functional data area is filled with data, the data area is implicitly marked based on the attribute location identifier, and the format and range of the filled data are verified to generate valid initial verification parameters for the data area, including:

[0091] Based on the attribute location identifier, the business attribute characteristics of the data area are parsed to generate a unique corresponding implicit mark code. Specifically, this includes: extracting the field codes such as attribute name and business meaning from the attribute location identifier, concatenating them into a string in order, converting each character in the string into ASCII code and accumulating them, summing all the ASCII code values, and then dividing the sum by 10000 and taking the remainder. This remainder is the unique implicit mark code of the data area.

[0092] Extract the format features of the data area fill content and compare them with the normal format standards of this business attribute to generate a format matching index. Specifically, this includes: extracting the numerical type, unit symbol (such as whether it contains "ton" or "%"), and character length (such as 100.5 tons is 6 characters), setting the preset standards for this attribute (such as the grade must be a numerical value with 1 decimal place, the unit is %, and the character length is 5-8 digits). 1 point is awarded for each standard met, and 3 points are awarded for 3 standards. The total score is divided by 3 to obtain the format matching index.

[0093] Based on the business attribute feature set, determine the reasonable value range of the data to be filled in the data area, compare the filled data with the range range, and generate a range compliance index. Specifically, this includes: obtaining the historical reasonable value range of each attribute in the business attribute feature set, using the maximum and minimum values ​​of the historical reasonable value range as the upper and lower limits of the reasonable value range, comparing the filled data with the reasonable value range, and if it is within the range, the range compliance index is 1; if it is not within the range, the range compliance index is 0.

[0094] The format matching index and range compliance index are integrated to generate effective initial verification parameters for this data area. Specifically, the format matching index is weighted at 0.4, and the range compliance index is weighted at 0.6. The format matching index and range compliance index are multiplied by their respective weights and then summed. The result is the effective initial verification parameters. In mineral product settlement, the core of settlement is to ensure the substantive accuracy of the data, and the range compliance index has a more critical impact on the results, so it is given a high weight of 0.6. Although format matching is important, it falls under the category of operational standardization, so it is given a weight of 0.4, balancing the needs of result accuracy and process standardization.

[0095] By generating unique implicit tag codes through attribute positioning identifiers, data areas are accurately associated with business attributes, avoiding confusion and omissions caused by manual tagging. After data is populated, the system automatically completes the tagging, providing clear positioning for subsequent data change tracking and correlation calculation checks, reducing efficiency losses caused by unclear tagging, and making data management more orderly. At the same time, the system generates effective initial verification parameters from both the aspects of format matching and range compliance, and allocates weights according to the degree of impact, ensuring both data format standardization and reasonable value selection. This reduces the subjectivity and oversight of manual verification and lowers the recalculation costs caused by subsequent data errors.

[0096] In one embodiment, when the data area of ​​the valid initial verification parameters changes, this data area is identified as the changed data area, its implicit marker is converted into an explicit marker, and the complete values, operation time, and operation subject of the changed data area before and after the change are recorded to generate a data change trajectory, including:

[0097] For the data area of ​​valid initial verification parameters, the current data is compared with the benchmark data in the valid initial verification parameters to generate the data variability index. Specifically, this includes: extracting the benchmark data from the valid initial verification parameters, calculating the absolute difference between the current data and the benchmark data, and then dividing by the benchmark data to obtain the change ratio, which is the data variability index.

[0098] When the data variability index reaches the change judgment threshold, the data area is identified as a changed data area. At the same time, based on the implicit mark encoding, an explicit mark activation instruction is generated to trigger the implicit mark to become an explicit mark. Specifically, the change judgment threshold is set to 0.5. If the data variability index is ≥0.5, it is determined to be a changed data area. If the data variability index is <0.5, it is still a data area. The implicit mark encoding of the changed data area is extracted. According to the fixed format of "implicit mark encoding-instruction type-color", the explicit mark instruction field and the red mark are concatenated to form a string such as "8620-explicit mark-red". This string is the activation instruction, which is to convert the original implicit mark to a red cell explicit mark.

[0099] The system acquires the original data snapshot at the moment of change, the updated data snapshot, the digital signature of the operating entity, and the operation sequence code to generate a set of core elements of the change. Specifically, this includes: when a data change is triggered, immediately capturing the complete value before the change as the original snapshot and the value after the change as the updated snapshot; obtaining the operating account ID to generate a digital signature; extracting the system clock's timestamp accurate to the second as the sequence code; and combining the four in the order of "original-updated-signature-sequence" to form a set of core elements of the change.

[0100] The importance of the core variable element set is classified according to the data variability index, and element level identifiers are generated. Specifically, three threshold levels are set: when the data variability index is ≥1.0, it is level 1; when the data variability index is between 0.7 and 0.9, it is level 2; and when the data variability index is between 0.5 and 0.6, it is level 3. The elements of the core variable element set are matched with the corresponding levels, and level identifiers are added before the elements to generate element level identifiers.

[0101] Analyze the element hierarchy identifiers and the core element set of changes to generate a data change trajectory that includes the depth of change and the scope of impact. Specifically, this includes: assigning the first-level element hierarchy identifier to a change depth of 1, the second-level element hierarchy identifier to a depth of 2, and the third-level element hierarchy identifier to a depth of 3; using the total number of cells involved in the change as the scope of impact (e.g., if there are 5 cells, the scope of impact is 5); and concatenating the element hierarchy identifiers, the original data, the updated data, the digital signature, the operation sequence code, along with the change depth and scope of impact, into a string in the order of "hierarchy identifier - original data - updated data - signature - sequence code - depth - scope" to obtain the data change trajectory.

[0102] The system accurately determines whether a data area has changed by using a data variability index. When a threshold is reached, the implicit marker is automatically turned into a red explicit marker, eliminating the need for manual identification and marking. This reduces the time and error associated with manual operations and improves the efficiency of data change processing. At the same time, the system generates a data change trajectory that includes the data before and after the change, the operating entity, the time, the depth of the change, and the scope of impact. This comprehensively records the details of the data change, making it easy to trace the overall situation of the data change and allowing staff to clearly understand the impact of the change. This eliminates the need for manual investigation, further reducing data verification time and ensuring the efficient progress of settlement work.

[0103] In one embodiment, based on attribute location identifiers and data change trajectories, a data dependency graph is generated by traversing the calculation formula association chains of the spreadsheet to construct the calculation dependencies between data areas, including:

[0104] The process involves parsing the attribute location identifiers and data change trajectories of each data area to generate a set of formula association links. Specifically, this includes: extracting the functional categories and coordinates from the spatial topology codes from the attribute location identifiers; traversing all calculation formulas in the spreadsheet; and identifying the data area (source) referenced by each formula and the data area (result) where the calculation results are stored from the data change trajectory. The source data area includes both unchanged data areas referenced by formulas and source data areas affected by changes in the data change trajectory. Similarly, the result data area covers both unchanged calculation result data areas and result data areas affected by changes in the data change trajectory. These strings are then concatenated in the following order: source - functional category - source coordinates - result functional category - result coordinates - change identifier (if it is a changed data area, its feature level identifier is marked; if it is an unchanged data area, it is marked as none). All such strings are combined to generate a set of formula association links.

[0105] Based on the formula association link set, the frequency of association between each data area as input and output is analyzed to generate a dependency strength coefficient. Specifically, this involves: statistically analyzing the number of times each data area is referenced as input and output in the formula association link set, setting the input frequency weight to 0.7 and the output frequency weight to 0.3, multiplying the input and output frequencies by their respective weights, summing them, and then dividing by the maximum association frequency of all data areas (the maximum of the two frequencies for all data areas, i.e., the highest frequency), to obtain the dependency strength coefficient of that data area. In the mineral product settlement scenario, the input frequency reflects the basic supporting role of the data area as the source of calculation. Basic data such as weight and grade are frequently referenced by downstream formulas, directly determining the accuracy of subsequent settlement results, and have a more critical impact on the dependency relationship, hence it is given a high weight of 0.7. On the other hand, the output frequency reflects the derived value of the data area as the result of calculation. Its changes are mostly caused by changes in upstream inputs, and its proactive impact on the dependency relationship is weaker, hence it is given a low weight of 0.3. The entire weight allocation conforms to the logic that source data determines derived results in settlement business.

[0106] Based on the dependency strength coefficient, the set of formula-related links is sorted out, and a topology structure with data areas as nodes and dependency relationships as edges is constructed to generate a data dependency graph.

[0107] By analyzing attribute location markers and data change trajectories, and traversing the calculation formula association chains to generate a set of formula association links, the system accurately captures the source and result relationships of each data area. Combined with dependency strength coefficients to clarify the links, the constructed topology clearly presents the computational dependencies between data areas, eliminating the tedious manual process of sorting out dependencies and making the relationships between data readily apparent. Furthermore, the generated data dependency graph visually displays the dependency strength and association paths of each data area. When data changes due to re-inspection or other reasons, staff can quickly locate the affected derived data areas using the graph, eliminating the need for individual checks and significantly reducing the time spent troubleshooting association calculations, effectively improving the efficiency of automated settlement.

[0108] In one embodiment, the set of formula-related links is sorted out based on the dependency strength coefficient, and a topology structure with data areas as nodes and dependencies as edges is constructed to generate a data dependency graph, including:

[0109] The dependency strength coefficient is divided into intervals, and the dependency relationship between data areas is divided into three levels according to the value, and dependency level labels are generated. Specifically, the dependency strength coefficient above 0.7 is classified as strong dependency, the dependency strength coefficient between 0.3 and 0.7 is classified as medium dependency, and the dependency strength coefficient below 0.3 is classified as weak dependency. For each data area, a label with a level name is generated according to the interval in which its dependency strength coefficient falls. The label contains the level and the value of the dependency strength coefficient, thus generating the dependency level label of the data area.

[0110] Based on the formula-based association link set, the indirect association degree between data areas under different dependency levels is calculated to generate secondary association coefficients. Specifically, this involves finding the indirect association path between data areas in the formula-based association link set (for example, data area A is associated with data area B to data area C, the path is A→B→C), extracting the dependency strength coefficient of each direct association in the path (for example, A→B is 0.8, B→C is 0.6), and multiplying these coefficients (0.8×0.6=0.48) to obtain the secondary association coefficient between A and C. The secondary association coefficient is used to represent the strength of the indirect association.

[0111] Based on dependency level labels and secondary correlation coefficients, an attributed topology is generated. Specifically, the basic framework is built with dependency level labels. Directly related nodes are connected using corresponding level labels (e.g., strong dependency -0.8). Indirectly related nodes are connected by adding secondary correlation coefficients as connecting edges in the middle of the path. Each node is labeled with the functional category and coordinates in the attribute positioning identifier, and each edge is labeled with the dependency strength coefficient, forming an attributed topology containing direct and indirect dependencies.

[0112] Based on the attributed topology, a data dependency graph containing multi-level dependencies and attributes is generated. Specifically, it is divided into two layers according to the attributed topology: direct dependencies (strong, medium, and weak level labels) and indirect dependencies (secondary association coefficients). The nodes in the direct dependency layer are arranged in strong, medium, and weak order, while the nodes in the indirect dependency layer are placed below the corresponding direct dependency layer. The nodes are labeled with their functional categories and coordinates, and the edges are labeled with the values ​​of the dependency strength coefficients. The entire graph is sorted from high to low dependency strength, forming a data dependency graph containing multi-level relationships and attributes.

[0113] By dividing the dependency strength coefficient intervals to generate dependency level labels, and combining them with secondary correlation coefficients to construct an attribute-based topology, a data dependency graph containing multi-level dependencies and attributes is generated. This makes the direct and indirect dependencies between data areas clear and distinct, avoiding confusion during manual sorting. It enables staff to quickly grasp the degree of correlation between data. Moreover, the data dependency graph is sorted by dependency strength, clearly displaying multi-level relationships and attributes. When mineral product re-inspection leads to data changes, the graph can quickly locate the derived data areas with strong, medium, and indirect dependencies, significantly shortening the inspection time for correlation calculations, reducing data verification time, and effectively improving the efficiency of automatic settlement of mineral products.

[0114] In one embodiment, all data regions that are computationally correlated with the changing data region are located based on the data dependency map, and these regions are used as derived data regions to generate a correlation influence hierarchy, including:

[0115] Starting with the variable data area, the topological structure in the data dependency graph is traversed to generate an initial associated dataset. Specifically, starting with the variable data area, the nodes directly associated in the data dependency graph (nodes with strong / medium / weak dependency connections) are found. Then, the nodes are traced layer by layer along the indirect association path (edges containing secondary association coefficients) to collect all associated nodes. These associated nodes are the derived data areas. The data areas are then combined in the format of variable data area ID - associated node ID - association type to generate the initial associated dataset.

[0116] Based on the dependency level labels and secondary association coefficients in the data dependency graph, the transmission depth and strength of the impact of changes on each derived data region in the initial associated dataset are calculated to generate an impact transmission index. Specifically, the transmission depth is calculated based on the number of nodes in the association path: direct association is 1, indirect association through one intermediate node is 2, and so on; the transmission strength is 0.8 when there is a strong dependency between derived data regions, 0.5 when there is a medium dependency between derived data regions, and 0.3 when there is a weak dependency between derived data regions. For indirect associations, the secondary association coefficient is directly used as the transmission strength; dividing 1 by the transmission depth and then multiplying by the transmission strength yields the impact transmission index.

[0117] Based on the distribution range of the influence transmission index and the core importance of the derived data area in the settlement logic, the initial associated dataset is hierarchically divided to generate hierarchical pre-classification results. Specifically, this includes: dividing the base range according to the influence transmission index: the derived data area with an influence transmission index ≥ 0.6 is designated as the core layer, the derived data area with an influence transmission index between 0.3 and 0.5 is designated as the associated layer, and the derived data area with an influence transmission index < 0.3 is designated as the radiation layer. For the derived data area with core attributes such as weight and grade, if the difference between its influence transmission index and the influence transmission index of the upper-level range is ≤ 0.1, then the derived data area of ​​the core attribute is moved up to the upper level (e.g., from the associated layer to the core layer), and finally, hierarchical pre-classification results are generated.

[0118] The hierarchical pre-classification results are verified by comparing the correlation degree of business attribute features of each level of derived data area to generate hierarchical verification coefficients. Specifically, this includes: extracting business attribute features (attribute name, business meaning, type, dimension) of each level of derived data area; dividing the number of features that completely overlap within the same level by the total number of feature items (4 items) to obtain the correlation degree of a single level; and adding the correlation degrees of features of the core layer, related layer, and radiation layer and dividing by 3 to obtain the hierarchical verification coefficients.

[0119] The hierarchical pre-classification results are re-divided based on the hierarchical verification coefficient, ultimately generating a hierarchy of influence including a core layer, a related layer, and a radiating layer. Specifically, this involves: setting a threshold of 0.7 for the hierarchical verification coefficient; if the hierarchical verification coefficient is ≥0.7, maintaining the hierarchical pre-classification result; if the hierarchical verification coefficient is <0.7, identifying the derived data region with the fewest overlapping features (e.g., only one overlapping feature) in the hierarchy with the lowest feature correlation, and calculating the overlap between this derived data region and all other hierarchical layers (core layer, related layer, radiating layer, excluding the layer with the lowest feature correlation). The number of overlapping items in the core features of the layers outside the previous layer is used to select the layer with the most overlapping items as the target layer (if multiple layers have the same number of overlapping items and the most, the priority is core layer > related layer > radiation layer). The derived data area is moved to the target layer, and the layer verification coefficient of each layer is recalculated. This process continues until the layer verification coefficient of all layers is ≥0.7, and the layer with the final association influence is determined. If the layer verification coefficient is still <0.7 after repeated adjustments, the highest coefficient is used as the passing line, and the corresponding layer division is retained.

[0120] Starting with the changed data area, an initial related dataset is generated by traversing the data dependency graph. This dataset is then stratified based on the impact transmission index and the core nature of the settlement logic. After verification and adjustment, a related impact hierarchy is generated, accurately locating all affected derivative data areas. This avoids omissions and misjudgments inherent in manual checks, allowing staff to fully grasp the scope of data changes. The related impact hierarchy is presented in an orderly fashion, with core, related, and radiating layers clearly reflecting the degree and priority of impact on each derivative data area. Staff can prioritize processing core layer data before processing other layers sequentially, reducing unnecessary operations, shortening data verification and recalculation time, and effectively improving the efficiency and accuracy of automated mineral product settlement.

[0121] In one embodiment, based on the hierarchy of influence, differentiated recalculation is performed on derived data areas at different levels to generate recalculation result verification results, including:

[0122] The distribution of derived data areas in the associated impact layers is analyzed, and the business attribute characteristics of each layer of data are combined to generate differentiated recalculation trigger thresholds. Specifically, this includes: the proportion of the number of derived data areas in the core layer, associated layer, and radiation layer (e.g., 0.2), where the proportion of the average business weight coefficient of each layer in the core layer, associated layer, and radiation layer is used as the trigger threshold for that layer, thereby generating differentiated recalculation trigger thresholds.

[0123] Based on the differentiated recalculation trigger threshold, the core layer, the related layer and the radiation layer are recalculated separately to generate recalculation performance indicators.

[0124] Obtain the recalculation output data of each level and integrate it into a hierarchical recalculation result set;

[0125] Extract the data difference before and after recalculation of each level of derived data area in the correlation influence hierarchy, and calculate the degree of deviation between the recalculation result and the theoretical value in combination with the business attribute feature set to generate the result deviation coefficient. Specifically, this includes: calculating the absolute difference between the recalculated and recalculated data areas of each level; dividing the absolute difference by the theoretical value of the derived data area in the business attribute feature set (the theoretical value is the median of the historical reasonable value range) to obtain the deviation ratio of a single derived data area; and weighting and summing the deviation ratios of each derived data area according to the corresponding business weight coefficient to obtain the result deviation coefficient of that level.

[0126] The recalculation performance index and the result deviation coefficient are correlated to generate a recalculation quality score. Specifically, the performance index weight is set to 0.3 and the deviation coefficient weight is set to 0.7. For the core layer, the related layer, and the radiation layer, the recalculation performance index and the result deviation coefficient of each layer are multiplied by their corresponding weights and summed to obtain the recalculation quality score of that layer.

[0127] Based on the recalculation quality score, the hierarchical recalculation result set is graded and verified. Derivative data areas with abnormal deviations are marked, and a recalculation result verification result containing verification conclusions and anomaly labels is generated. Specifically, the following steps are taken: set recalculation quality score thresholds: the recalculation quality score of the core layer ≥ 0.85, the recalculation quality score of the related layer ≥ 0.75, and the recalculation quality score of the radiation layer ≥ 0.65 are considered qualified. For the layers below the recalculation quality score threshold, from the hierarchical recalculation result set, derivative data areas with a deviation ratio (absolute difference between the data before and after recalculation divided by the theoretical value) of more than 10% are selected and marked as abnormal with explicit marking (red border). The conclusions of whether each layer is qualified and the derivative data areas marked with anomalies are integrated to form a recalculation result verification result containing verification conclusions and anomaly labels.

[0128] By generating differentiated recalculation trigger thresholds based on the correlation influence hierarchy, targeted recalculation is performed on the core layer, correlation layer, and radiation layer, avoiding the waste of resources caused by indiscriminate recalculation. Furthermore, the core layer, as the key data area, is given priority for high-precision recalculation, while the radiation layer is processed according to the appropriate threshold. This ensures the accuracy of core data, improves overall recalculation efficiency, and reduces calculation time. At the same time, the recalculation results are graded and verified through result deviation coefficients and recalculation quality scores, accurately marking abnormal data areas. Staff do not need to check the recalculation results one by one, but can directly focus on the abnormal areas for processing, shortening the verification time of calculation results.

[0129] In one embodiment, based on a differentiated recalculation trigger threshold, the core layer, the associated layer, and the radiation layer are recalculated respectively to generate recalculation performance indicators, including:

[0130] The differential recalculation trigger threshold is analyzed to generate the core layer trigger critical point, the associated layer batch threshold, and the radiation layer time window threshold. Specifically, this includes: dividing the core layer differential recalculation trigger threshold by 2 to obtain the core layer trigger critical point; multiplying the associated layer differential recalculation trigger threshold by three-quarters to obtain the associated layer batch threshold; and multiplying the radiation layer differential recalculation trigger threshold by 240 minutes to obtain the radiation layer time window threshold (assuming the radiation layer differential recalculation trigger threshold is 0.3, 0.3 × 240 = 72 minutes, then a timed recalculation is performed every 72 minutes).

[0131] Based on the core layer trigger threshold, real-time recalculation of the core layer derived data area is initiated, recording the response time and accuracy compliance rate to generate a real-time response coefficient. Specifically, this includes: when the core layer derived data area changes to the core layer trigger threshold, real-time recalculation is immediately initiated, recording the time from triggering to completion; dividing this time by a preset standard time yields the time ratio; dividing the number of derived data areas meeting the accuracy standard in the recalculation results by the total number of data areas yields the compliance rate; setting the weight of the time ratio to 0.4 and the compliance rate to 0.6, multiplying the time ratio and compliance rate by their respective weights and summing them yields the real-time response coefficient; the accuracy compliance requirement is: the deviation ratio between the recalculated result of the core layer derived data area and the theoretical value is ≤5%; exceeding this is considered non-compliance.

[0132] Aggregating derived data areas of the association layer based on the batch threshold of the association layer, statistically analyzing processing speed and success rate, and generating a batch processing index. Specifically, this includes: when the changes in the derived data areas of the association layer reach the batch threshold, counting all derived data areas to be recalculated and recording the corresponding total processing time; dividing the total processing time by the total number of data areas to obtain the average processing time per area; then dividing the average processing time per area by the standard processing time per area to obtain the speed ratio; dividing the number of derived data areas that successfully completed recalculation by the total number of data areas to obtain the success rate; setting the weight of the speed ratio to 0.3 and the weight of the success rate to 0.7; and multiplying the speed ratio and success rate by their respective weights. The results are then summed to obtain the batch processing index. In the automatic settlement of mineral products, although the related layer derived data area is not the core, it involves the intermediate links of the settlement chain. The reliability of its recalculation results directly affects the downstream radiation layer data and the final settlement conclusion. Because mineral product settlement has strict requirements for data accuracy, the failure of the related layer recalculation will lead to the distortion of the batch derived data, which may cause subsequent settlement deviations. Therefore, a high weight of 0.7 is given to the success rate to ensure the correctness of the recalculation results. The related layer uses batch recalculation instead of real-time recalculation, which allows for a certain delay. Compared with the success rate, its importance is lower, so a low weight of 0.3 is given to the speed ratio.

[0133] Based on the radiation layer time window threshold, the radiation layer derived data area is divided into recalculation periods. The task completion rate and resource load matching degree are calculated to generate the timed execution coefficient. Specifically, the recalculation periods are divided according to the radiation layer time window threshold. The number of derived data areas that are actually recalculated in the period is divided by the planned number to obtain the task completion rate. The actual resource usage is divided by the standard load limit to obtain the resource load matching degree. The task completion rate weight is multiplied by 0.6 and the load matching degree is multiplied by 0.4. The two products are added together to obtain the timed execution coefficient.

[0134] The real-time response coefficient, batch processing index, and timed execution coefficient are integrated to generate a recalculation performance index. Specifically, the proportion of the number of derived data areas in the core layer, related layer, and radiation layer to the total number of data areas is used as the weights of the real-time response coefficient, batch processing index, and timed execution coefficient, respectively. The real-time response coefficient, batch processing index, and timed execution coefficient are multiplied by their respective weights and then summed to obtain the recalculation performance index.

[0135] By setting different recalculation thresholds and methods for the core layer, related layer, and radiation layer, the core layer performs real-time recalculation to ensure timely updates of key data, the related layer performs batch processing to balance efficiency and accuracy, and the radiation layer performs timed recalculation to reduce resource consumption. This differentiated recalculation mode avoids the blindness of uniform processing, ensuring the accuracy of core data while optimizing the processing rhythm of intermediate links and secondary data, significantly reducing invalid recalculation time. Recalculation performance indicators are generated through real-time response coefficients, batch processing indexes, and timed execution coefficients to comprehensively reflect the recalculation efficiency and quality of each level. Staff can optimize recalculation strategies based on the indicators to continuously improve system processing performance, further shorten the recalculation cycle after data changes, and effectively solve the problem of low efficiency in manual processing.

[0136] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An automated settlement system for mineral products, characterized in that, The method comprises the following steps: a division unit is configured to obtain an electronic form for settlement of mineral products, divide the form into multiple functional data areas according to rows and columns, analyze the business attributes of each data area, and generate attribute positioning identifiers for the data areas, including: extracting the business attribute information of each data area, classifying the information, and generating a business attribute feature set; analyzing the spatial positions of the data areas in the electronic form and generating spatial topology codes in combination with the row and column relationships of the form; associating the business attribute feature set and the spatial topology codes to generate an attribute-spatial mapping matrix; analyzing the importance of each attribute in the business attribute feature set to generate business weight coefficients, including: analyzing the irreplaceability of each attribute in the business process and the frequency of the attribute as an input for downstream calculation to generate a link dependency index; analyzing the influence degree of each attribute on the settlement result when the attribute is abnormal and the correction cost to generate an abnormal influence coefficient; analyzing the link dependency index and the abnormal influence coefficient to generate a link index confidence and an abnormal coefficient confidence, respectively; fusing the link dependency index and the abnormal influence coefficient based on the link index confidence and the abnormal coefficient confidence to generate the business weight coefficient; fusing the spatial topology codes and the attribute-spatial mapping matrix based on the business weight coefficient to generate the attribute positioning identifier for the data area; a verification unit is configured to implicitly mark any functional data area based on the attribute positioning identifier after the data area is filled with data, verify the format and range of the filled data, and generate valid initial verification parameters for the data area; a change unit is configured to determine a data area as a changed data area when the data area of the valid initial verification parameters changes, convert the implicit marking of the data area to explicit marking, record the complete values of the changed data area before and after the change, the operation time, and the operation subject, and generate a data change track; a calculation unit is configured to construct the calculation dependency relationship between data areas based on the attribute positioning identifier and the data change track by traversing the calculation formula association chain of the electronic form, and generate a data dependency graph; an influence unit is configured to locate all data areas that have a calculation association with the changed data area based on the data dependency graph, and generate an association influence hierarchy by taking the data areas as derivative data areas; a recalculation unit is configured to perform differential recalculation on the derivative data areas at different levels based on the association influence hierarchy, and generate a recalculation result verification result.

2. An automated mineral product settlement system according to claim 1, wherein When any functional data area is filled with data, the data area is implicitly marked based on the attribute positioning identifier, the format and range of the filled data are verified, and valid initial verification parameters for the data area are generated, including: analyzing the business attribute features of the data area based on the attribute positioning identifier to generate a unique implicit marking code; extracting the format features of the filled content of the data area, comparing the features with the normal format standard of the business attribute, and generating a format matching degree index; determining the reasonable value interval range of the filled data of the data area based on the business attribute feature set, comparing the filled data with the interval range, and generating a range compliance index; integrating the format matching degree index and the range compliance index to generate the valid initial verification parameters for the data area.

3. An automated mineral product settlement system according to claim 2, wherein When the data area of the effective initial check parameter changes, the data area is determined as a changed data area, the implicit label is converted into an explicit label, and the complete values before and after the change of the changed data area, the operation time and the operation subject are recorded to generate a data change track, including: For the data area of the effective initial check parameter, the current data is compared with the reference data in the effective initial check parameter to generate a data variation index; When the data variation index reaches a change judgment threshold, the data area is determined as a changed data area, and an explicit label activation instruction is generated based on the implicit label coding to trigger the conversion of the implicit label into the explicit label; The original data snapshot at the change moment, the updated data snapshot, the digital signature of the operation subject and the operation time sequence code are obtained to generate a change core element set; The change core element set is classified according to the importance of the data variation index to generate an element level identifier; The element level identifier and the change core element set are analyzed to generate a data change track containing the change depth and the influence range.

4. An automated mineral product settlement system according to claim 3, wherein According to the attribute positioning identifier and the data change track, the calculation formula association chain of the electronic form is traversed to construct the calculation dependence relationship between the data areas and generate a data dependence graph, including: The attribute positioning identifier and the data change track of each data area are analyzed to generate a formula association link set; Based on the formula association link set, the association frequency of each data area as a calculation input and output is analyzed to generate a dependence strength coefficient; Based on the dependence strength coefficient, the formula association link set is combed to construct a topological structure taking the data area as a node and the dependence relationship as an edge to generate a data dependence graph.

5. An automated mineral product settlement system according to claim 4, wherein Based on the dependence strength coefficient, the formula association link set is combed to construct a topological structure taking the data area as a node and the dependence relationship as an edge to generate a data dependence graph, including: The dependence strength coefficient is divided into intervals, the dependence relationship between the data areas is divided into three levels according to the numerical value, and a dependence level label is generated; Based on the formula association link set, the indirect association degree between the data areas under different dependence levels is calculated to generate a secondary association coefficient; Based on the dependence level label and the secondary association coefficient, an attribute topological structure is generated; Based on the attribute topological structure, a data dependence graph containing multi-level dependence relationship and attributes is generated.

6. An automated mineral product settlement system according to claim 5, wherein, According to the data dependence graph, all data areas that have calculation association with the changed data area are located and taken as derivative data areas to generate an association influence level, including: Starting from the changed data area, the topological structure in the data dependence graph is traversed to generate an initial associated data set; Based on the dependence level label and the secondary association coefficient in the data dependence graph, the conduction depth and strength of each derivative data area affected by the change in the initial associated data set are calculated to generate an influence conduction index; According to the distribution interval of the influence conduction index, the initial associated data set is hierarchically divided according to the core degree of the derivative data area in the settlement logic to generate a hierarchical pre-classification result; The hierarchical pre-classification result is verified, and a hierarchical verification coefficient is generated by comparing the business attribute feature association degree of each hierarchical derivative data area; The hierarchical pre-classification result is re-divided based on the hierarchical check coefficient, and finally the correlation influence hierarchy including the core layer, the correlation layer and the radiation layer is generated.

7. An automated mineral product settlement system according to claim 6, wherein, Based on the correlation influence hierarchy, differential recalculation is performed on the derivative data areas of different levels to generate a recalculation result check result, including: The distribution of the derivative data areas of the correlation influence hierarchy is analyzed, and combined with the business attribute characteristics of each level data, a differential recalculation trigger threshold is generated; Based on the differential recalculation trigger threshold, the core layer, the correlation layer and the radiation layer are respectively recalculated to generate a recalculation efficiency index; The recalculation output data of each level is obtained and integrated into a hierarchical recalculation result set; The data difference before and after the recalculation of each level derivative data area in the correlation influence hierarchy is extracted, and combined with the business attribute characteristic set, the deviation degree of the recalculation result from the theoretical value is calculated to generate a result deviation coefficient; The recalculation efficiency index and the result deviation coefficient are associated to generate a recalculation quality score; Based on the recalculation quality score, the hierarchical recalculation result set is checked by grade, and the derivative data area with abnormal deviation is marked to generate a recalculation result check result including the check conclusion and the abnormal marking.

8. An automated mineral product settlement system according to claim 7, characterised in that, Based on the differential recalculation trigger threshold, the core layer, the correlation layer and the radiation layer are respectively recalculated to generate a recalculation efficiency index, including: The differential recalculation trigger threshold is analyzed to generate a core layer trigger critical point, a correlation layer batch threshold and a radiation layer time window threshold; Based on the core layer trigger critical point, real-time recalculation of the core layer derivative data area is started to generate a real-time response coefficient; Based on the correlation layer batch threshold, the correlation layer derivative data area is aggregated to generate a batch processing index; Based on the radiation layer time window threshold, the radiation layer derivative data area is divided into recalculation time periods to generate a timing execution coefficient; The real-time response coefficient, the batch processing index and the timing execution coefficient are integrated to generate a recalculation efficiency index.

Images