A multi-source data parameter evaluation method based on data deep learning calculation

Abstract

The application discloses a kind of multi-source data parameter evaluation methods based on data deep learning calculation, belong to data parameter evaluation technical field.The application of a kind of multi-source data parameter evaluation methods based on data deep learning calculation, including the following steps: real-time acquisition to multi-source data, the multi-source data information of acquisition is handled, and the accuracy calculation of multi-source data parameter is based on the mode of data deep learning calculation.The application solves the transmission of the existing multi-source data, cannot be based on multi-source data parameter evaluation and then transmission, resulting in multi-source data transmission risk, and multi-source data transmission is unstable and the problem of low data transmission accuracy, the transmission of the application for multi-source data, can be based on multi-source data parameter evaluation and then transmission, reduce multi-source data transmission risk, and multi-source data transmission can be stable and improve data transmission accuracy.

Description

Technical Field

[0001] This invention relates to the field of data parameter evaluation technology, specifically to a multi-source data parameter evaluation method based on deep learning computation. Background Technology

[0002] Deep learning is a new research direction in the field of machine learning. It was introduced into machine learning to bring it closer to its original goal—artificial intelligence. Deep learning learns the inherent patterns and hierarchical representations of sample data. The information gained during this learning process greatly helps in interpreting data such as text, images, and sound. The ultimate goal is to enable machines to have analytical and learning capabilities like humans, capable of recognizing data such as text, images, and sound. Deep learning is a complex machine learning algorithm that has achieved results in speech and image recognition far exceeding previous related technologies. Deep learning has also made significant progress in search technology, data mining, machine learning, machine translation, natural language processing, multimedia learning, speech recognition, recommendation and personalization technologies, and other related fields. Deep learning enables machines to mimic human activities such as sight, hearing, and thinking, solving many complex pattern recognition problems and leading to significant advancements in artificial intelligence-related technologies.

[0003] Chinese patent CN111858346A discloses a multi-dimensional data quality evaluation technology based on deep learning test datasets. This technology uses deep neural networks for test set metric evaluation, providing two levels of metrics at the input and output levels, and ultimately generating a quality report for the test dataset. The input to this patent is a deep learning test dataset. Based on the specific structure and information of the dataset, and combined with the neural network used, it performs a multi-dimensional quality evaluation of the dataset. It introduces topological data analysis, connecting the input and output layer results, and linking them to the basic framework and properties of deep learning to help generate a quality evaluation report that guides the selection and adjustment of deep learning test datasets. However, the aforementioned patent has the following drawbacks:

[0004] For the transmission of multi-source data, it is not possible to evaluate the parameters of the multi-source data before transmission, which leads to high risks, instability and low accuracy of multi-source data transmission. Summary of the Invention

[0005] The purpose of this invention is to provide a method for evaluating multi-source data parameters based on deep learning computation. For the transmission of multi-source data, the method can evaluate the multi-source data parameters before transmission, thereby reducing the risk of multi-source data transmission, stabilizing multi-source data transmission, and improving the accuracy of data transmission, thus solving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for evaluating multi-source data parameters based on deep learning computation includes the following steps:

[0008] S1: Real-time acquisition of multi-source data, acquisition of multi-source data information, processing of the acquired multi-source data information, accuracy calculation of multi-source data parameters based on data deep learning calculation method, determination of multi-source data parameter calculation results, and execution of different multi-source data storage and transmission strategies based on the multi-source data parameter calculation results;

[0009] S2: If the calculation results of the multi-source data parameters are within the acceptable accuracy range, then the multi-source data parameters will be transmitted and saved. If the calculation results of the multi-source data parameters are not within the acceptable accuracy range, then the multi-source data parameters will be saved separately and will not be transmitted.

[0010] S3: Obtain multi-source data parameters with unsatisfactory accuracy, analyze the multi-source data parameters based on the data nodes, evaluate the multi-source data parameters through data node analysis, and present the evaluation results in tabular form.

[0011] Preferably, the collected multi-source data is processed by performing the following operations:

[0012] Completely extract information from the acquired multi-source data;

[0013] Data retrieval is performed on multi-source data information. Based on the multi-source data parameter evaluation requirements, multi-source data information that is not useful for multi-source data parameter evaluation is filtered out, while multi-source data information that is useful for multi-source data parameter evaluation is retained.

[0014] The retained multi-source data information is classified and divided into multiple categories according to different keywords. Each category stores different multi-source data information.

[0015] The system performs calculations on the multi-source data information for classification, calculates the accuracy of the multi-source data information based on deep learning, and determines the calculation results of the multi-source data parameters based on the accuracy calculation of the multi-source data information.

[0016] Preferably, different multi-source data storage and transmission strategies are executed based on the calculation results of multi-source data parameters, and the following operations are performed:

[0017] Obtain parameter calculation results from multi-source data information;

[0018] By referring to the acceptable accuracy range of the stored multi-source data parameters, the parameter calculation results of the multi-source data information are compared and analyzed.

[0019] If the parameter calculation results of the multi-source data information are within the acceptable range of the accuracy of the multi-source data parameters, then the multi-source data parameters will be transmitted and saved.

[0020] If the calculated result of the parameters of the multi-source data information is not within the acceptable range of the accuracy of the multi-source data parameters, then the multi-source data parameters will be saved separately and will not be transmitted.

[0021] Preferably, the multi-source data parameters with unacceptable accuracy are obtained, and the data nodes of the multi-source data parameters are found. The data nodes of the multi-source data parameters include data node name, data node creator, data node creation type, data node modification time, data node modifier, number of associated verification points, data source, fields involved in verification, sorting field, hash field, hash bit length and description.

[0022] Preferably, the multi-source data parameters are analyzed based on the data nodes of the multi-source data parameters, and the following operations are performed:

[0023] Data nodes that acquire parameters from multiple data sources;

[0024] Extract the multi-source information of data nodes for multi-source data parameters, and analyze the multi-source information of data nodes for multi-source data parameters with the stored standard information of data nodes;

[0025] If a single piece of information in the multi-source information of a data node with multi-source data parameters is consistent with the standard information of the data node, then the accuracy of that single piece of information in the data node is qualified.

[0026] If a single piece of information in the multi-source information of a data node is inconsistent with the standard information of the data node, then the accuracy of that single piece of information in the data node is unqualified.

[0027] Based on the analysis results of the data nodes, the data parameters of the multi-source data are evaluated.

[0028] Preferably, the multi-source information of data nodes for multi-source data parameters is analyzed together with the stored standard information of data nodes, and the following operations are performed:

[0029] Extract individual information of data nodes from the multi-source information of multi-source data parameters one by one;

[0030] Each piece of information extracted from multiple data nodes is compared and analyzed with the standard information of the data nodes.

[0031] For cases where the accuracy of a single piece of information in a data node is unqualified, the unqualified individual pieces of information in the data node are extracted, statistically analyzed, and uniformly stored in a list of unqualified accuracy.

[0032] Preferably, based on data node analysis, the multi-source data parameters are evaluated, and the following operations are performed:

[0033] Obtain individual information of data nodes in the list of data nodes with unacceptable accuracy, and calculate the number S of data parameters with unacceptable accuracy within the data node;

[0034] By combining the number S of data parameters with substandard accuracy, the multi-source data parameters are evaluated by integrating individual information from multiple data nodes with substandard accuracy.

[0035] Different data parameter evaluation results are determined based on the data node conditions with different data parameters;

[0036] Obtain evaluation results for multiple sets of data parameters and present the evaluation results in tabular form.

[0037] Preferably, the multi-source data parameters are evaluated based on the number S of data parameters with unacceptable accuracy, and the following operations are performed:

[0038] If 0 ≤ number of data parameter items S < 3, the determined data parameter evaluation result indicates that the current multi-source data parameter risk is relatively low;

[0039] If 3 ≤ number of data parameter items S < 5, the determined data parameter evaluation result is that the current multi-source data parameter risk is generally considered to be moderate.

[0040] If 5 ≤ the number of data parameter items S, the determined data parameter evaluation result is of high risk for the current multi-source data parameters;

[0041] Based on the evaluation results of the determined data parameters, the evaluation results are presented in tabular form, and the corresponding data parameter evaluation risk situation is shown.

[0042] Preferably, the multi-source data parameter evaluation method further includes:

[0043] Real-time monitoring of multi-source data information whose separately stored parameter calculation results are outside the acceptable accuracy range of multi-source data parameters, and selective deletion of the multi-source data information based on storage space requirements, including:

[0044] Real-time detection and monitoring of storage units used for storing multi-source data information whose individual parameter calculation results are not within the accuracy range of multi-source data parameters, and obtaining the remaining storage capacity of the storage unit;

[0045] Extract the amount of multi-source data whose parameter calculation results are outside the accuracy range of the multi-source data parameters for each unit of time, and use it as a reference factor;

[0046] The remaining storage capacity of the storage unit and a reference factor are used to obtain the remaining space resource evaluation parameters of the storage unit; wherein, the remaining space resource evaluation parameters are obtained by the following formula:

[0047] R = λ1·R0 + λ2·△R

[0048]

[0049] Where R represents the remaining space resource evaluation parameter; R0 represents the remaining storage capacity of the current storage unit; ΔR represents the parameter compensation amount; n represents the total number of times that the amount of multi-source data information within the accuracy range of the multi-source data parameters generated in the next unit time exceeds the amount of multi-source data information within the accuracy range of the multi-source data parameters generated in the previous unit time; ΔC i This represents the data volume fluctuation difference when the amount of multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the next unit of time exceeds the amount of multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the previous unit of time; S maxi This represents the sum of the number of data parameters S corresponding to the risk level type with the largest data proportion among the multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the next unit time when the amount of multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the previous unit time exceeds the amount of multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the next unit time; f t This represents the weight value corresponding to each risk level type. When the risk level is low due to multi-source data parameters, f... t =0.26, when the risk level is general for multi-source data parameters, f t =0.31, when the risk level is high for multi-source data parameters, f t =0.43; m represents the total number of units of time that have been elapsed; C j represents the amount of multi-source data information generated within the acceptable accuracy range of the multi-source data parameters in the j-th unit of time; m2 represents the number of data items with a moderate risk level among the existing data; m3 represents the number of data items with a relatively high risk level among the existing data; S 2i S represents the number of data parameters in the i-th data set with a risk level of "general"; 3i This represents the number of data parameter items corresponding to the i-th data set with a relatively high risk level;

[0050] When the remaining space resource evaluation parameter is lower than the preset evaluation parameter threshold, the multi-source data information is selectively deleted.

[0051] Preferably, selective deletion of the multi-source data information includes:

[0052] Extract each multi-source data information from the storage unit;

[0053] The storage value parameter for each of the multi-source data information is obtained using a comprehensive value evaluation model, wherein the comprehensive value evaluation model is as follows:

[0054]

[0055] Among them, R f C1 represents the storage value parameter; C1 represents the amount of data corresponding to the multi-source data information within the accuracy range of each multi-source data parameter; C represents the total amount of data stored in the storage unit; P represents the total number of data retrievals in the storage unit; P 0i This indicates the number of times the multi-source data information is invoked within the acceptable accuracy range of the i-th multi-source data parameter;

[0056] When the stored value parameter is lower than a preset value parameter threshold, the multi-source data information whose stored value parameter is lower than the preset value parameter threshold is deleted.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] 1. This invention acquires multi-source data in real time, processes the acquired multi-source data, and calculates the accuracy of multi-source data parameters based on deep learning computation. The calculation results are then used to determine the accuracy of the multi-source data parameters. Different multi-source data storage and transmission strategies are implemented based on these results. If the calculation results are within the acceptable accuracy range, the multi-source data parameters are transmitted and stored. If the calculation results are outside the acceptable accuracy range, the multi-source data parameters are stored separately and not transmitted. The invention then analyzes the multi-source data parameters based on their data nodes, evaluates the data parameters through this analysis, and presents the evaluation results in tabular form. Transmission of multi-source data can be based on parameter evaluation before transmission, reducing the risk of multi-source data transmission and improving the stability and accuracy of data transmission.

[0059] 2. This invention, when processing collected multi-source data, first extracts the complete multi-source data, performs data retrieval, filters out useless multi-source data for parameter evaluation according to the evaluation requirements, and retains the useful multi-source data. Then, the retained useful multi-source data is categorized according to different keywords, into multiple categories, each storing different multi-source data. Finally, the categorized multi-source data is calculated using deep learning to determine the accuracy of the multi-source data calculations. Based on this accuracy calculation, the multi-source data parameter calculation results are determined, providing sufficient assurance for subsequent multi-source data transmission and reducing the risk of multi-source data transmission. Attached Figure Description

[0060] Figure 1 This is a flowchart of the multi-source data parameter evaluation method of the present invention;

[0061] Figure 2 This is an algorithm diagram of the present invention that executes different multi-source data storage and transmission strategies based on the calculation results of multi-source data parameters;

[0062] Figure 3 This is a diagram of the algorithm for analyzing multi-source data parameters based on data nodes of multi-source data parameters according to the present invention.

[0063] Figure 4 This is a diagram of the algorithm for evaluating multi-source data parameters based on the number S of data parameters with unacceptable accuracy, as described in this invention. Detailed Implementation

[0064] 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.

[0065] To address the issues of high risk, instability, and low accuracy in multi-source data transmission caused by the current method of transmitting multi-source data without prior evaluation of multi-source data parameters, please refer to [link to relevant documentation]. Figures 1-3 This embodiment provides the following technical solution:

[0066] A method for evaluating multi-source data parameters based on deep learning computation includes the following steps:

[0067] S1: Real-time acquisition of multi-source data, acquisition of multi-source data information, processing of the acquired multi-source data information, accuracy calculation of multi-source data parameters based on data deep learning calculation method, determination of multi-source data parameter calculation results, and execution of different multi-source data storage and transmission strategies based on the multi-source data parameter calculation results;

[0068] S2: If the calculation results of the multi-source data parameters are within the acceptable accuracy range, then the multi-source data parameters will be transmitted and saved. If the calculation results of the multi-source data parameters are not within the acceptable accuracy range, then the multi-source data parameters will be saved separately and will not be transmitted.

[0069] S3: Obtain multi-source data parameters with unsatisfactory accuracy, analyze the multi-source data parameters based on the data nodes, evaluate the multi-source data parameters through data node analysis, and present the evaluation results in tabular form.

[0070] It should be noted that multi-source data is collected in real time to acquire multi-source data information. This information is then processed, and accuracy calculations are performed on the multi-source data parameters using deep learning. The calculation results are then used to determine the accuracy of the multi-source data parameters. Different multi-source data storage and transmission strategies are implemented based on these results. If the calculation results are within the acceptable accuracy range, the multi-source data parameters are transmitted and stored. If the results are outside the acceptable accuracy range, the parameters are stored separately and not transmitted. The multi-source data parameters with unacceptable accuracy are then identified. The data nodes of these parameters are analyzed to evaluate the data parameters, and the evaluation results are presented in tabular form. Transmission of multi-source data can be based on parameter evaluation before transmission, reducing the risk of multi-source data transmission and improving its stability and accuracy.

[0071] The collected multi-source data is processed by performing the following operations:

[0072] Completely extract information from the acquired multi-source data;

[0073] Data retrieval is performed on multi-source data information. Based on the multi-source data parameter evaluation requirements, multi-source data information that is not useful for multi-source data parameter evaluation is filtered out, while multi-source data information that is useful for multi-source data parameter evaluation is retained.

[0074] The retained multi-source data information is classified and divided into multiple categories according to different keywords. Each category stores different multi-source data information.

[0075] The system performs calculations on the multi-source data information for classification, calculates the accuracy of the multi-source data information based on deep learning, and determines the calculation results of the multi-source data parameters based on the accuracy calculation of the multi-source data information.

[0076] It should be noted that when processing the collected multi-source data, the process begins by fully extracting the acquired data, performing data retrieval, and filtering out useless data based on the multi-source parameter evaluation requirements. Only the useful data is retained. This useful data is then categorized using different keywords, forming multiple classes, each containing different data. Finally, the categorized data is calculated using deep learning to determine its accuracy. This accuracy calculation determines the multi-source parameter calculation results, providing a solid foundation for subsequent multi-source data transmission and reducing transmission risks.

[0077] Based on the calculation results of multi-source data parameters, different multi-source data storage and transmission strategies are executed, and the following operations are performed:

[0078] Obtain parameter calculation results from multi-source data information;

[0079] By referring to the acceptable accuracy range of the stored multi-source data parameters, the parameter calculation results of the multi-source data information are compared and analyzed.

[0080] If the parameter calculation results of the multi-source data information are within the acceptable range of the accuracy of the multi-source data parameters, then the multi-source data parameters will be transmitted and saved.

[0081] If the calculated result of the parameters of the multi-source data information is not within the acceptable range of the accuracy of the multi-source data parameters, then the multi-source data parameters will be saved separately and will not be transmitted.

[0082] To obtain multi-source data parameters with unsatisfactory accuracy, locate the data nodes of the multi-source data parameters. The data nodes of the multi-source data parameters include the data node name, data node creator, data node creation type, data node modification time, data node modifier, number of associated verification points, data source, fields involved in verification, sorting field, hash field, hash bit length, and description.

[0083] It should be noted that if the calculation results of multi-source data parameters are outside the acceptable accuracy range, these multi-source data parameters should be saved separately and not transmitted. The multi-source data parameters with unacceptable accuracy should be retrieved, and the data nodes of these parameters should be analyzed to evaluate them. The evaluation results should be presented in tabular form. Transmission should only proceed after the multi-source data parameters have been evaluated, reducing the risk of multi-source data transmission and ensuring stable and accurate data transmission.

[0084] Analyze the multi-source data parameters based on the data nodes of the multi-source data parameters, and perform the following operations:

[0085] Data nodes that acquire parameters from multiple data sources;

[0086] Extract the multi-source information of data nodes for multi-source data parameters, and analyze the multi-source information of data nodes for multi-source data parameters with the stored standard information of data nodes;

[0087] If a single piece of information in the multi-source information of a data node with multi-source data parameters is consistent with the standard information of the data node, then the accuracy of that single piece of information in the data node is qualified.

[0088] If a single piece of information in the multi-source information of a data node is inconsistent with the standard information of the data node, then the accuracy of that single piece of information in the data node is unqualified.

[0089] Based on the analysis results of the data nodes, the data parameters of the multi-source data are evaluated.

[0090] Analyze the multi-source information of data nodes and the stored standard information of data nodes for multi-source data parameters, and perform the following operations:

[0091] Extract individual information of data nodes from the multi-source information of multi-source data parameters one by one;

[0092] Each piece of information extracted from multiple data nodes is compared and analyzed with the standard information of the data nodes.

[0093] For cases where the accuracy of a single piece of information in a data node is unqualified, the unqualified individual pieces of information in the data node are extracted, statistically analyzed, and uniformly stored in a list of unqualified accuracy.

[0094] It should be noted that the analysis of multi-source data parameters is based on the data nodes of the multi-source data parameters. The data nodes of the multi-source data parameters contain multiple sets of data information. The individual information of the data nodes in the multi-source information of the multi-source data parameters is extracted one by one. The extracted individual information of multiple data nodes is compared and analyzed with the standard information of the data nodes. For the cases where the accuracy of the individual information of the data nodes is not up to standard, the unqualified individual information of the data nodes is extracted, statistically analyzed, and uniformly stored in the list of unqualified accuracy.

[0095] Please see Figure 4 Based on data node analysis, evaluate the data parameters of multi-source data and perform the following operations:

[0096] Obtain individual information of data nodes in the list of data nodes with unacceptable accuracy, and calculate the number S of data parameters with unacceptable accuracy within the data node;

[0097] By combining the number S of data parameters with substandard accuracy, the multi-source data parameters are evaluated by integrating individual information from multiple data nodes with substandard accuracy.

[0098] Different data parameter evaluation results are determined based on the data node conditions with different data parameters;

[0099] Obtain evaluation results for multiple sets of data parameters and present the evaluation results in tabular form.

[0100] To evaluate the multi-source data parameters based on the number S of data parameters with unacceptable accuracy, perform the following operations:

[0101] If 0 ≤ number of data parameter items S < 3, the determined data parameter evaluation result indicates that the current multi-source data parameter risk is relatively low;

[0102] If 3 ≤ number of data parameter items S < 5, the determined data parameter evaluation result is that the current multi-source data parameter risk is generally considered to be moderate.

[0103] If 5 ≤ the number of data parameter items S, the determined data parameter evaluation result is of high risk for the current multi-source data parameters;

[0104] Based on the evaluation results of the determined data parameters, the evaluation results are presented in tabular form, and the corresponding data parameter evaluation risk situation is shown.

[0105] In one embodiment of the present invention, the multi-source data parameter evaluation method further includes:

[0106] Real-time monitoring of multi-source data information whose separately stored parameter calculation results are outside the acceptable accuracy range of multi-source data parameters, and selective deletion of the multi-source data information based on storage space requirements, including:

[0107] Real-time detection and monitoring of storage units used for storing multi-source data information whose individual parameter calculation results are not within the accuracy range of multi-source data parameters, and obtaining the remaining storage capacity of the storage unit;

[0108] Extract the amount of multi-source data whose parameter calculation results are outside the accuracy range of the multi-source data parameters for each unit of time, and use it as a reference factor;

[0109] The remaining storage capacity of the storage unit and a reference factor are used to obtain the remaining space resource evaluation parameters of the storage unit; wherein, the remaining space resource evaluation parameters are obtained by the following formula:

[0110] R = λ1·R0 + λ2·△R

[0111]

[0112] Where R represents the remaining space resource evaluation parameter; R0 represents the remaining storage capacity of the current storage unit; ΔR represents the parameter compensation amount; n represents the total number of times that the amount of multi-source data information within the accuracy range of the multi-source data parameters generated in the next unit time exceeds the amount of multi-source data information within the accuracy range of the multi-source data parameters generated in the previous unit time; ΔC i This represents the data volume fluctuation difference when the amount of multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the next unit of time exceeds the amount of multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the previous unit of time; S maxi This represents the sum of the number of data parameters S corresponding to the risk level type with the largest data proportion among the multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the next unit time when the amount of multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the previous unit time exceeds the amount of multi-source data information within the accuracy acceptable range of the multi-source data parameters generated in the next unit time; f t This represents the weight value corresponding to each risk level type. When the risk level is low due to multi-source data parameters, f... t =0.26, when the risk level is general for multi-source data parameters, f t =0.31, when the risk level is high for multi-source data parameters, f t =0.43; m represents the total number of units of time that have been elapsed; C jrepresents the amount of multi-source data information generated within the acceptable accuracy range of the multi-source data parameters in the j-th unit of time; m2 represents the number of data items with a moderate risk level among the existing data; m3 represents the number of data items with a relatively high risk level among the existing data; S 2i S represents the number of data parameters in the i-th data set with a risk level of "general"; 3i This represents the number of data parameter items corresponding to the i-th data set with a relatively high risk level;

[0113] When the remaining space resource evaluation parameter is lower than the preset evaluation parameter threshold, the multi-source data information is selectively deleted.

[0114] The above technical solution achieves the following effects: It effectively improves the accuracy and rationality of assessing the remaining space resources of storage units. By combining the different data occupancy levels and remaining space within the already occupied storage space as factors in the remaining space resource evaluation parameters, it effectively improves the matching between the evaluation parameters and the actual data storage and risk situation. Furthermore, using the number of data parameters to obtain the remaining space resource evaluation parameters allows for a better understanding of the overall risk of the data stored within the current storage unit. Although the risk level of multi-source data information is generally moderate, increasing the overall number of data parameters increases the overall data risk of the storage unit. Therefore, using the number of data parameters as one of the reference factors to obtain the remaining space resource evaluation parameters enhances the overall responsiveness and accuracy of the remaining space resource evaluation parameters to the risk of the storage unit.

[0115] One embodiment of the present invention includes selectively deleting the multi-source data information, comprising:

[0116] Extract each multi-source data information from the storage unit;

[0117] The storage value parameter for each of the multi-source data information is obtained using a comprehensive value evaluation model, wherein the comprehensive value evaluation model is as follows:

[0118]

[0119] Among them, R f C1 represents the storage value parameter; C1 represents the amount of data corresponding to the multi-source data information within the accuracy range of each multi-source data parameter; C represents the total amount of data stored in the storage unit; P represents the total number of data retrievals in the storage unit; P 0i This indicates the number of times the multi-source data information is invoked within the acceptable accuracy range of the i-th multi-source data parameter;

[0120] When the stored value parameter is lower than a preset value parameter threshold, the multi-source data information whose stored value parameter is lower than the preset value parameter threshold is deleted.

[0121] The above technical solution achieves the following results: it effectively improves the risk assessment accuracy of multi-source data information within the acceptable accuracy range of each multi-source data parameter in the storage unit. Furthermore, by combining data retrieval with the acquisition of storage value parameters, it effectively improves the value evaluation accuracy of multi-source data information within the acceptable accuracy range of multi-source data parameters. Simultaneously, it simplifies the evaluation model structure, improves computational response speed, and enhances parameter evaluation efficiency.

[0122] In summary, the multi-source data parameter evaluation method based on deep learning computation of the present invention collects multi-source data in real time, acquires multi-source data information, processes the acquired multi-source data information, calculates the accuracy of multi-source data parameters based on deep learning computation, determines the calculation results of multi-source data parameters, and executes different multi-source data storage and transmission strategies according to the calculation results. If the calculation results of multi-source data parameters are within the acceptable accuracy range, the multi-source data parameters are transmitted and stored. If the calculation results of multi-source data parameters are not within the acceptable accuracy range, the multi-source data parameters are stored separately and not transmitted, and the multi-source data parameters with unacceptable accuracy are acquired. The multi-source data parameters are analyzed based on the data nodes of the multi-source data parameters, and the data parameters are evaluated through data node analysis. The evaluation results are presented in tabular form. For the transmission of multi-source data, it is possible to conduct multi-source data parameter evaluation before transmission, which reduces the risk of multi-source data transmission and can make multi-source data transmission stable and improve the accuracy of data transmission.

[0123] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0124] 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. A method for evaluating multi-source data parameters based on deep learning computation, characterized in that: Includes the following steps: S1: Real-time acquisition of multi-source data, acquisition of multi-source data information, processing of the acquired multi-source data information, accuracy calculation of multi-source data information based on data deep learning computation, determination of multi-source data parameter calculation results, and execution of different multi-source data storage and transmission strategies based on the multi-source data parameter calculation results; S2: If the calculation results of the multi-source data parameters are within the acceptable accuracy range, then the multi-source data parameters will be transmitted and saved. If the calculation results of the multi-source data parameters are not within the acceptable accuracy range, then the multi-source data parameters will be saved separately and will not be transmitted. S3: Obtain multi-source data parameters with unsatisfactory accuracy, analyze the multi-source data parameters based on the data nodes, evaluate the multi-source data parameters through data node analysis, and present the evaluation results in tabular form.

2. The method for evaluating multi-source data parameters based on deep learning computation according to claim 1, characterized in that: The collected multi-source data is processed by performing the following operations: Completely extract information from the acquired multi-source data; Data retrieval is performed on multi-source data information. Based on the multi-source data parameter evaluation requirements, multi-source data information that is not useful for multi-source data parameter evaluation is filtered out, while multi-source data information that is useful for multi-source data parameter evaluation is retained. The retained multi-source data information is classified and divided into multiple categories according to different keywords. Each category stores different multi-source data information. The system performs calculations on the multi-source data information for classification, calculates the accuracy of the multi-source data information based on deep learning, and determines the calculation results of the multi-source data parameters based on the accuracy calculation of the multi-source data information.

3. The method for evaluating multi-source data parameters based on deep learning computation according to claim 2, characterized in that: Based on the calculation results of multi-source data parameters, different multi-source data storage and transmission strategies are executed, and the following operations are performed: Obtain the calculation results of parameters from multi-source data; The accuracy range of the stored multi-source data information is used as a reference to compare and analyze the calculation results of the multi-source data parameters. If the calculation results of the multi-source data parameters are within the acceptable range of the accuracy of the multi-source data information, then the multi-source data parameters will be transmitted and saved. If the calculation result of the multi-source data parameter is not within the acceptable range of the accuracy of the multi-source data information, then the multi-source data parameter will be saved separately and will not be transmitted.

4. The method for evaluating multi-source data parameters based on deep learning computation according to claim 3, characterized in that: To obtain multi-source data parameters with unsatisfactory accuracy, locate the data nodes of the multi-source data parameters. The data nodes of the multi-source data parameters include the data node name, data node creator, data node creation type, data node modification time, data node modifier, number of associated verification points, data source, fields involved in verification, sorting field, hash field, hash bit length, and description.

5. The method for evaluating multi-source data parameters based on deep learning computation according to claim 4, characterized in that: Analyze the multi-source data parameters based on the data nodes of the multi-source data parameters, and perform the following operations: Data nodes that acquire parameters from multiple data sources; Extract the multi-source information of data nodes for multi-source data parameters, and analyze the multi-source information of data nodes for multi-source data parameters with the stored standard information of data nodes; If a single piece of information in the multi-source information of a data node with multi-source data parameters is consistent with the standard information of the data node, then the accuracy of that single piece of information in the data node is qualified. If a single piece of information in the multi-source information of a data node is inconsistent with the standard information of the data node, then the accuracy of that single piece of information in the data node is unqualified. Based on the analysis results of the data nodes, the data parameters of the multi-source data are evaluated.

6. The method for evaluating multi-source data parameters based on deep learning computation according to claim 5, characterized in that: Analyze the multi-source information of data nodes and the stored standard information of data nodes for multi-source data parameters, and perform the following operations: Extract individual information of data nodes from the multi-source information of multi-source data parameters one by one; Each piece of information extracted from multiple data nodes is compared and analyzed with the standard information of the data nodes. For cases where the accuracy of a single piece of information in a data node is unqualified, the unqualified individual pieces of information in the data node are extracted, statistically analyzed, and uniformly stored in a list of unqualified accuracy.

7. The method for evaluating multi-source data parameters based on deep learning computation according to claim 6, characterized in that: Based on data node analysis, evaluate the parameters of multi-source data and perform the following operations: Obtain individual information of data nodes in the list of data nodes with unacceptable accuracy, and calculate the number S of data parameters with unacceptable accuracy within the data node; By combining the number S of data parameters with substandard accuracy, the multi-source data parameters are evaluated by integrating individual information from multiple data nodes with substandard accuracy. Different data parameter evaluation results are determined based on the data node conditions with different data parameters; Obtain evaluation results for multiple sets of data parameters and present the evaluation results in tabular form.

8. The method for evaluating multi-source data parameters based on deep learning computation according to claim 7, characterized in that: To evaluate the multi-source data parameters based on the number S of data parameters with unacceptable accuracy, perform the following operations: If 0 ≤ number of data parameter items S < 3, the determined data parameter evaluation result indicates that the current multi-source data parameter risk is relatively low; If 3 ≤ number of data parameter items S < 5, the determined data parameter evaluation result is that the current multi-source data parameter risk is generally considered to be moderate. If 5 ≤ the number of data parameter items S, the determined data parameter evaluation result is of high risk for the current multi-source data parameters; Based on the evaluation results of the determined data parameters, the evaluation results are presented in tabular form, and the corresponding data parameter evaluation risk situation is shown.

9. The method for evaluating multi-source data parameters based on deep learning computation according to claim 1, characterized in that: The multi-source data parameter evaluation method also includes: Real-time monitoring of multi-source data whose parameter calculation results are outside the acceptable accuracy range of multi-source data information, and selective deletion of such multi-source data information based on storage space requirements, including: Real-time detection and monitoring of storage units used to separately store multi-source data information whose multi-source data parameter calculation results are not within the accuracy acceptable range of multi-source data information, and obtaining the remaining storage capacity of the storage unit; Extract the amount of multi-source data whose calculation results for the multi-source data parameters determined in each unit of time are not within the acceptable range of the accuracy of the multi-source data information, and use it as a reference factor; The remaining storage capacity of the storage unit and a reference factor are used to obtain the remaining space resource evaluation parameters of the storage unit; wherein, the remaining space resource evaluation parameters are obtained by the following formula: ； Where R represents the remaining space resource evaluation parameter; R0 represents the remaining storage capacity of the current storage unit; ΔR represents the parameter compensation amount; n represents the total number of times the amount of multi-source data within the accuracy range of the multi-source data information generated in the next unit of time exceeds the amount of multi-source data within the accuracy range of the multi-source data information generated in the previous unit of time; ΔC i S represents the data volume fluctuation when the amount of multi-source data generated in the next unit of time with an accuracy range exceeding the amount of multi-source data generated in the previous unit of time; maxi This represents the sum of the number of data parameters S corresponding to the risk level type with the largest data proportion among the multi-source data information within the accuracy acceptable range generated in the next unit of time when the amount of multi-source data information within the accuracy acceptable range of the multi-source data information generated in the previous unit of time exceeds the amount of multi-source data information within the accuracy acceptable range of the multi-source data information generated in the next unit of time; f t This represents the weight value corresponding to each risk level type. When the risk level is low due to multi-source data parameters, f... t =0.26, when the risk level is general for multi-source data parameters, f t =0.31, when the risk level is that the multi-source data parameters have a relatively high risk, f t =0.43; m represents the total number of units of time that have been elapsed; C j represents the amount of multi-source data generated within the accuracy acceptable range in the j-th unit of time; m2 represents the number of data items with a moderate risk level among the existing data; m3 represents the number of data items with a relatively high risk level among the existing data; S 2i S represents the number of data parameters in the i-th data set with a risk level of "general"; 3i This represents the number of data parameter items corresponding to the i-th data set with a relatively high risk level; When the remaining space resource evaluation parameter is lower than the preset evaluation parameter threshold, the multi-source data information is selectively deleted.

10. The method for evaluating multi-source data parameters based on deep learning computation according to claim 9, characterized in that: Selective deletion of the multi-source data information includes: Extract each multi-source data information from the storage unit; The storage value parameter for each of the multi-source data information is obtained using a comprehensive value evaluation model, wherein the comprehensive value evaluation model is as follows: ； Among them, R f The storage value parameter is represented by C1; the data volume corresponding to the multi-source data information within the accuracy range of each multi-source data information is represented by C; the total data volume stored in the storage unit is represented by P; and the total number of data retrievals in the storage unit is represented by P. 0i This represents the number of times the multi-source data information within the accuracy range of the i-th multi-source data information is invoked; When the storage value parameter is lower than the preset value parameter threshold, the multi-source data information whose storage value parameter is lower than the preset value parameter threshold is deleted.

Images